ae732e01-04b2-4665-b570-a77210e73e28.txt

====================================================================================================
import os
import sys
with open(sys.argv[0]) as f:
    code = f.read() # read the code of this file ASAP, for logging
import uuid
import glob
import time
from dataclasses import dataclass

import numpy as np
import torch
from torch import nn
import torch.nn.functional as F
import torch.distributed as dist
import torch._inductor.config as config
from torch.nn.parallel import DistributedDataParallel as DDP
# Use of FlexAttention contributed by @KoszarskyB
from torch.nn.attention.flex_attention import flex_attention, create_block_mask
flex_attention = torch.compile(flex_attention, dynamic=False)
create_block_mask = torch.compile(create_block_mask, dynamic=False)

# -----------------------------------------------------------------------------
# Muon optimizer

def zeropower_via_svd(G, steps=None):
    U, S, V = G.svd()
    return U @ V.T

@torch.compile
def zeropower_via_newtonschulz5(G, steps=10, eps=1e-7):
    """
    Newton-Schulz iteration to compute the zeroth power / orthogonalization of G. We opt to use a
    quintic iteration whose coefficients are selected to maximize the slope at zero. For the purpose
    of minimizing steps, it turns out to be empirically effective to keep increasing the slope at
    zero even beyond the point where the iteration no longer converges all the way to one everywhere
    on the interval. This iteration therefore does not produce UV^T but rather something like US'V^T
    where S' is diagonal with S_{ii}' ~ Uniform(0.5, 1.5), which turns out not to hurt model
    performance at all relative to UV^T, where USV^T = G is the SVD.
    """
    assert len(G.shape) == 2
    a, b, c = (3.4445, -4.7750,  2.0315)
    X = G.bfloat16()
    X /= (X.norm() + eps) # ensure top singular value <= 1
    if G.size(0) > G.size(1):
        X = X.T
    for _ in range(steps):
        A = X @ X.T
        B = b * A + c * A @ A # adapted from suggestion by @jxbz, @leloykun, and @YouJiacheng
        X = a * X + B @ X
    if G.size(0) > G.size(1):
        X = X.T
    return X

zeropower_backends = dict(svd=zeropower_via_svd, newtonschulz5=zeropower_via_newtonschulz5)

class Muon(torch.optim.Optimizer):
    """
    Muon - MomentUm Orthogonalized by Newton-schulz

    Muon internally runs standard SGD-momentum, and then performs an orthogonalization post-
    processing step, in which each 2D parameter's update is replaced with the nearest orthogonal
    matrix. To efficiently orthogonalize each update, we use a Newton-Schulz iteration, which has
    the advantage that it can be stably run in bfloat16 on the GPU.

    Some warnings:
    - This optimizer assumes that all parameters passed in are 2D.
    - It should not be used for the embedding layer, the final fully connected layer, or any {0,1}-D
    parameters; those should all be optimized by a standard method (e.g., AdamW).
    - To use it with 4D convolutional filters, it works well to just flatten their last 3 dimensions.
    - We believe it is unlikely to work well for training with small batch size.
    - We believe it may not work well for finetuning pretrained models, but we haven't tested this.
    - We have not yet tried this optimizer for training scenarios larger than NanoGPT (124M).

    Arguments:
        lr: The learning rate used by the internal SGD.
        momentum: The momentum used by the internal SGD.
        nesterov: Whether to use Nesterov-style momentum in the internal SGD. (recommended)
        backend: The chosen backend for the orthogonalization step. (recommended: 'newtonschulz5')
        backend_steps: The number of iteration steps to use in the backend, if it is iterative.
    """
    def __init__(self, params, lr=0.02, momentum=0.95, nesterov=True,
                 backend='newtonschulz5', backend_steps=5):
        defaults = dict(lr=lr, momentum=momentum, nesterov=nesterov, backend=backend, backend_steps=backend_steps)
        super().__init__(params, defaults)

    def step(self):

        for group in self.param_groups:

            lr = group['lr']
            momentum = group['momentum']
            zeropower_backend = zeropower_backends[group['backend']]

            # generate weight updates in distributed fashion
            total_params = sum(p.numel() for p in group['params'])
            updates_flat = torch.zeros(total_params, device='cuda', dtype=torch.bfloat16)
            curr_idx = 0
            for i, p in enumerate(group['params']):
                # luckily this will perfectly distribute a transformer with multiple of 4 layers to 8 GPUs
                if i % int(os.environ['WORLD_SIZE']) == int(os.environ['RANK']):
                    g = p.grad
                    assert g is not None
                    state = self.state[p]
                    if 'momentum_buffer' not in state:
                        state['momentum_buffer'] = torch.zeros_like(g)
                    buf = state['momentum_buffer']
                    buf.mul_(momentum).add_(g)
                    if group['nesterov']:
                        g = g.add(buf, alpha=momentum)
                    g = zeropower_backend(g, steps=group['backend_steps'])
                    g *= max(1, g.size(0)/g.size(1))**0.5
                    updates_flat[curr_idx:curr_idx+p.numel()] = g.flatten()
                curr_idx += p.numel()

            # sync updates across devices. we are not memory-constrained so can do this simple deserialization
            dist.all_reduce(updates_flat, op=dist.ReduceOp.SUM)

            # deserialize and apply updates
            curr_idx = 0
            for p in group['params']:
                g = updates_flat[curr_idx:curr_idx+p.numel()].view_as(p.data).type_as(p.data)
                p.data.add_(g, alpha=-lr)
                curr_idx += p.numel()

# -----------------------------------------------------------------------------
# PyTorch nn.Module definitions for the GPT-2 model

class Rotary(torch.nn.Module):

    def __init__(self, dim, base=10000):
        super().__init__()
        self.dim = dim
        self.base = base
        self.inv_freq = None
        self.seq_len_cached = None
        self.cos_cached = None
        self.sin_cached = None

    def forward(self, x):
        seq_len = x.shape[1]
        if seq_len != self.seq_len_cached:
            self.inv_freq = 1.0 / (self.base ** (torch.arange(0, self.dim, 2, device=x.device).float() / self.dim))
            self.seq_len_cached = seq_len
            t = torch.arange(seq_len, device=x.device).type_as(self.inv_freq)
            freqs = torch.outer(t, self.inv_freq)
            self.cos_cached = freqs.cos().bfloat16()
            self.sin_cached = freqs.sin().bfloat16()
        return self.cos_cached[None, :, None, :], self.sin_cached[None, :, None, :]

def apply_rotary_emb(x, cos, sin):
    assert x.ndim == 4 # multihead attention
    d = x.shape[3]//2
    x1 = x[..., :d]
    x2 = x[..., d:]
    y1 = x1 * cos + x2 * sin
    y2 = x1 * (-sin) + x2 * cos
    return torch.cat([y1, y2], 3).type_as(x)

class CastedLinear(nn.Linear):
    def forward(self, x):
        return F.linear(x, self.weight.to(x.dtype))

class CausalSelfAttention(nn.Module):

    def __init__(self, config):
        super().__init__()
        self.n_head = config.n_head
        self.n_embd = config.n_embd
        self.head_dim = self.n_embd // self.n_head
        assert self.n_embd % self.n_head == 0
        self.c_q = CastedLinear(self.n_embd, self.n_embd, bias=False)
        self.c_k = CastedLinear(self.n_embd, self.n_embd, bias=False)
        self.c_v = CastedLinear(self.n_embd, self.n_embd, bias=False)
        # output projection
        self.c_proj = CastedLinear(self.n_embd, self.n_embd, bias=False)
        self.c_proj.weight.data.zero_() # zero init suggested by @Grad62304977
        self.rotary = Rotary(self.head_dim)
        self.lamb = nn.Parameter(torch.tensor(0.5)) # @Grad62304977

    def forward(self, x, v1, block_mask):
        B, T, C = x.size() # batch size, sequence length, embedding dimensionality (n_embd)
        q = self.c_q(x).view(B, T, self.n_head, self.head_dim)
        k = self.c_k(x).view(B, T, self.n_head, self.head_dim)
        v = self.c_v(x).view(B, T, self.n_head, self.head_dim)
        if v1 is None:
            v1 = v # This happens if we are in the first block. v needs to be accessed by subsequent blocks
        v = (1 - self.lamb) * v + self.lamb * v1.view_as(v) # @Grad62304977
        cos, sin = self.rotary(q)
        q, k = F.rms_norm(q, (q.size(-1),)), F.rms_norm(k, (k.size(-1),)) # QK norm suggested by @Grad62304977
        q, k = apply_rotary_emb(q, cos, sin), apply_rotary_emb(k, cos, sin)
        y = flex_attention(q.transpose(1, 2), k.transpose(1, 2), v.transpose(1, 2), block_mask=block_mask)
        y = y.transpose(1, 2).contiguous().view_as(x) # re-assemble all head outputs side by side
        y = self.c_proj(y)
        return y, v1

class MLP(nn.Module):

    def __init__(self, config):
        super().__init__()
        self.c_fc    = CastedLinear(config.n_embd, 4 * config.n_embd, bias=False)
        self.c_proj  = CastedLinear(4 * config.n_embd, config.n_embd, bias=False)
        self.c_proj.weight.data.zero_() # zero init suggested by @Grad62304977

    def forward(self, x):
        x = self.c_fc(x)
        x = F.relu(x).square() # https://arxiv.org/abs/2109.08668v2; ~1-2% better than GELU; suggested by @SKYLINEZ007 and @Grad62304977
        x = self.c_proj(x)
        return x

class Block(nn.Module):

    def __init__(self, config):
        super().__init__()
        self.attn = CausalSelfAttention(config)
        self.mlp = MLP(config)
        self.lambdas = nn.Parameter(torch.tensor([1., 0.]))

    def forward(self, x, v1, x0, block_mask):
        x = self.lambdas[0] * x + self.lambdas[1] * x0
        x1, v1 = self.attn(F.rms_norm(x, (x.size(-1),)), v1, block_mask)
        x = x + x1
        x = x + self.mlp(F.rms_norm(x, (x.size(-1),)))
        return x, v1

# -----------------------------------------------------------------------------
# The main GPT-2 model

@dataclass
class GPTConfig:
    vocab_size : int = 50304
    n_layer : int = 12
    n_head : int = 6 # head dim 128 suggested by @Grad62304977
    n_embd : int = 768

class GPT(nn.Module):

    def __init__(self, config):
        super().__init__()

        # U-net design by @brendanh0gan
        self.num_encoder_layers = config.n_layer // 2 # Half of the layers for encoder
        self.num_decoder_layers = config.n_layer - self.num_encoder_layers # Remaining for decoder
        # Add learnable skip connection weights for decoder layers
        self.skip_weights = nn.Parameter(torch.ones(self.num_decoder_layers))

        self.transformer = nn.ModuleDict(dict(
            wte = nn.Embedding(config.vocab_size, config.n_embd),
            h = nn.ModuleList([Block(config) for _ in range(config.n_layer)]),
        ))
        self.lm_head = CastedLinear(config.n_embd, config.vocab_size, bias=False)
        self.lm_head.weight.data.zero_() # @Grad62304977

    def forward(self, idx, target, attn_blocksize):

        docs = (idx == 50256).cumsum(0)
        def document_causal_mask(b, h, q_idx, kv_idx):
          causal_mask = q_idx >= kv_idx
          document_mask = docs[q_idx] == docs[kv_idx]
          window_mask = q_idx - kv_idx < attn_blocksize
          return causal_mask & document_mask & window_mask

        S = len(idx)
        block_mask = create_block_mask(document_causal_mask, None, None, S, S, device="cuda", _compile=True)

        # forward the GPT model itself
        x = self.transformer.wte(idx[None]) # token embeddings of shape (b, t, n_embd)
        x = F.rms_norm(x, (x.size(-1),)) # @Grad62304977
        x0 = x
        v1 = None

        # Store outputs for U-Net skip connections
        skip_connections = []
        # Encoder pass - process only the first half of the blocks
        for i in range(self.num_encoder_layers):
            x, v1 = self.transformer.h[i](x, v1, x0, block_mask)
            skip_connections.append(x)
        # Decoder pass - process the remaining blocks with weighted skip connections
        for i in range(self.num_decoder_layers):
            x = x + self.skip_weights[i] * skip_connections.pop()
            x, v1 = self.transformer.h[self.num_encoder_layers + i](x, v1, x0, block_mask)

        x = F.rms_norm(x, (x.size(-1),))
        logits = self.lm_head(x)
        logits = 30 * torch.tanh(logits / 30) # @Grad62304977
        logits = logits.float()
        loss = F.cross_entropy(logits.view(-1, logits.size(-1)), target.view(-1))
        return loss

# -----------------------------------------------------------------------------
# Our own simple Distributed Data Loader

def _peek_data_shard(filename):
    # only reads the header, returns header data
    with open(filename, "rb") as f:
        # first read the header, which is 256 int32 integers (4 bytes each)
        header = np.frombuffer(f.read(256*4), dtype=np.int32)
    if header[0] != 20240520:
        print("ERROR: magic number mismatch in the data .bin file!")
        print("---> HINT: Are you passing in a correct file with --input_bin?")
        print("---> HINT: Dataset encoding changed recently, re-run data prepro or refer again to README")
        print("---> HINT: For example re-run: `python dev/data/tinyshakespeare.py`, then re-try")
        exit(1)
    assert header[1] == 1, "unsupported version"
    ntok = header[2] # number of tokens (claimed)
    return ntok # for now just return the number of tokens

def _load_data_shard(filename):
    with open(filename, "rb") as f:
        # first read the header, which is 256 int32 integers (4 bytes each)
        header = np.frombuffer(f.read(256*4), dtype=np.int32)
        assert header[0] == 20240520, "magic number mismatch in the data .bin file"
        assert header[1] == 1, "unsupported version"
        ntok = header[2] # number of tokens (claimed)
        # the rest of it are tokens, stored as uint16
        tokens = np.frombuffer(f.read(), dtype=np.uint16)
    assert len(tokens) == ntok, "number of tokens read does not match header?"
    return tokens

class DistributedDataLoader:
    def __init__(self, filename_pattern, B, T, process_rank, num_processes):
        self.process_rank = process_rank
        self.num_processes = num_processes
        self.B = B
        self.T = T

        # glob files that match the pattern
        self.files = sorted(glob.glob(filename_pattern))
        assert len(self.files) > 0, f"did not find any files that match the pattern {filename_pattern}"

        # load and validate all data shards, count number of tokens in total
        ntok_total = 0
        for fname in self.files:
            shard_ntok = _peek_data_shard(fname)
            assert shard_ntok >= num_processes * B * T + 1
            ntok_total += int(shard_ntok)
        self.ntok_total = ntok_total

        self.reset()

    def reset(self):
        self.current_shard = -1
        self.advance()

    def advance(self): # advance to next data shard
        self.current_shard = (self.current_shard + 1) % len(self.files)
        self.current_position = self.process_rank * self.B * self.T
        self.tokens = _load_data_shard(self.files[self.current_shard])

    def next_batch(self):
        batch_size = self.B * self.T * self.num_processes
        buf = self.tokens[self.current_position:self.current_position+self.B*self.T+1]
        buf = torch.tensor(buf.astype(np.int32), dtype=torch.long)
        x = buf[:-1] # inputs
        y = buf[1:] # targets
        # advance current position and load next shard if necessary
        self.current_position += batch_size
        if self.current_position + batch_size >= len(self.tokens):
            self.advance()
        return x.cuda(), y.cuda()

# -----------------------------------------------------------------------------
# int main

@dataclass
class Hyperparameters:
    # data hyperparams
    input_bin : str = 'data/fineweb10B/fineweb_train_*.bin' # input .bin to train on
    input_val_bin : str = 'data/fineweb10B/fineweb_val_*.bin' # input .bin to eval validation loss on
    # optimization hyperparams
    batch_size : int = 8 # batch size, in sequences, across all devices
    device_batch_size : int = 1 # batch size, in sequences, per device
    sequence_length : int = 64*1024 # sequence length, in tokens
    num_iterations : int = 1750 # number of iterations to run
    warmup_iters : int = 0
    cooldown_iters : int = 640 # number of iterations of linear warmup/cooldown for triangular or trapezoidal schedule
    weight_decay : float = 0
    # evaluation and logging hyperparams
    val_loss_every : int = 125 # every how many steps to evaluate val loss? 0 for only at the end
    val_tokens : int = 10485760 # how many tokens of validation data? it's important to keep this fixed for consistent comparisons
    save_every : int = 0 # every how many steps to save the checkpoint? 0 for only at the end
args = Hyperparameters()

# set up DDP (distributed data parallel). torchrun sets this env variable
assert torch.cuda.is_available()
dist.init_process_group(backend='nccl')
ddp_rank = int(os.environ['RANK'])
ddp_local_rank = int(os.environ['LOCAL_RANK'])
ddp_world_size = int(os.environ['WORLD_SIZE'])
device = f'cuda:{ddp_local_rank}'
torch.cuda.set_device(device)
print(f"using device: {device}")
master_process = (ddp_rank == 0) # this process will do logging, checkpointing etc.

# begin logging
logfile = None
if master_process:
    run_id = str(uuid.uuid4())
    logdir = 'logs/%s/' % run_id
    os.makedirs(logdir, exist_ok=True)
    logfile = 'logs/%s.txt' % run_id
    # create the log file
    with open(logfile, "w") as f:
        # begin the log by printing this file (the Python code)
        f.write('='*100 + '\n')
        f.write(code)
        f.write('='*100 + '\n')
def print0(s, logonly=False):
    if master_process:
        with open(logfile, "a") as f:
            if not logonly:
                print(s)
            f.write(s+'\n')
# log information about the hardware/software environment this is running on
# and print the full `nvidia-smi` to file
print0(f"Running pytorch {torch.version.__version__} compiled for CUDA {torch.version.cuda}\nnvidia-smi:")
import subprocess
result = subprocess.run(['nvidia-smi'], stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True)
print0(f'{result.stdout}', logonly=True)
print0('='*100, logonly=True)

# convenience variables
B, T = args.device_batch_size, args.sequence_length
# calculate the number of steps to take in the val loop.
assert args.val_tokens % (B * T * ddp_world_size) == 0
val_steps = args.val_tokens // (B * T * ddp_world_size)
# calculate the steps of gradient accumulation required to attain the desired global batch size.
assert args.batch_size % (B * ddp_world_size) == 0
train_accumulation_steps = args.batch_size // (B * ddp_world_size)

# load tokens
train_loader = DistributedDataLoader(args.input_bin, B, T, ddp_rank, ddp_world_size)
val_loader = DistributedDataLoader(args.input_val_bin, B, T, ddp_rank, ddp_world_size)
print0(f"Training DataLoader: total number of tokens: {train_loader.ntok_total} across {len(train_loader.files)} files")
print0(f"Validation DataLoader: total number of tokens: {val_loader.ntok_total} across {len(val_loader.files)} files")
print0('='*100, logonly=True)
x, y = train_loader.next_batch()

# there are only 50257 unique GPT-2 tokens; we extend to nearest multiple of 128 for efficiency. suggested to me by @Grad62304977.
# this originates from Karpathy's experiments.
num_vocab = 50304
model = GPT(GPTConfig(vocab_size=num_vocab, n_layer=12, n_head=6, n_embd=768))
model = model.cuda().bfloat16()
for m in model.modules():
    if isinstance(m, CastedLinear):
        m.float()
if hasattr(config, "coordinate_descent_tuning"):
    config.coordinate_descent_tuning = True # suggested by @Chillee
model = torch.compile(model)
# here we wrap model into DDP container
model = DDP(model, device_ids=[ddp_local_rank])
raw_model = model.module # always contains the "raw" unwrapped model

# CUDNN attention is ~4ms faster than Flash, but doesn't get selected by default in PyTorch 2.5.1
from torch.backends.cuda import enable_cudnn_sdp, enable_flash_sdp, enable_math_sdp, enable_mem_efficient_sdp
enable_cudnn_sdp(True)
enable_flash_sdp(False)
enable_mem_efficient_sdp(False)
enable_math_sdp(False)

# init the optimizer(s)
optimizer1 = torch.optim.Adam([raw_model.transformer.wte.weight], lr=0.6,   betas=(0.8, 0.95), fused=True)
optimizer2 = torch.optim.Adam([raw_model.lm_head.weight],         lr=0.008, betas=(0.8, 0.95), fused=True)
params = list(raw_model.transformer.h.parameters())
matrix_params = [p for p in params if p.ndim == 2]
scalar_params = [p for p in params if p.ndim < 2] + [raw_model.skip_weights]
optimizer3 = Muon(matrix_params, lr=0.05, momentum=0.95)
optimizer4 = torch.optim.Adam(scalar_params, lr=0.04, betas=(0.8, 0.95), fused=True) # note that this learning rate is neither sensitive nor tuned
optimizers = [optimizer1, optimizer2, optimizer3, optimizer4]
# learning rate decay scheduler (linear warmup and cooldown)
def get_lr(it):
    assert it <= args.num_iterations
    # 1) linear warmup for warmup_iters steps
    if it < args.warmup_iters:
        return (it+1) / args.warmup_iters
    # 2) constant lr for a while
    elif it < args.num_iterations - args.cooldown_iters:
        return 1.0
    # 3) linear cooldown
    else:
        decay_ratio = (args.num_iterations - it) / args.cooldown_iters
        return decay_ratio
schedulers = [torch.optim.lr_scheduler.LambdaLR(opt, get_lr) for opt in optimizers]

# Start training loop
training_time_ms = 0
# start the clock
torch.cuda.synchronize()
t0 = time.time()
# begin training
for step in range(args.num_iterations + 1):
    last_step = (step == args.num_iterations)
    # Set the attention blocksize for the current step, in chunks of 64
    attn_blocksize = torch.tensor(64*((step/args.num_iterations * (1792 - 64) + 64)//64), dtype=torch.int, device='cuda')
    # This effectively ignores timing first 10 steps, which are slower for weird reasons.
    # Alternately, and slightly more correctly in terms of benchmarking, we could do 10
    # steps with dummy data first, and then re-initialize the model and reset the loader.
    if step == 10:
        training_time_ms = 0
        t0 = time.time()
    timed_steps = float('nan') if step <= 11 else (step - 10) + 1 # <= 11 to avoid bug in val

    # once in a while evaluate the validation dataset
    if (last_step or (args.val_loss_every > 0 and step % args.val_loss_every == 0)):
        # stop the clock
        torch.cuda.synchronize()
        training_time_ms += 1000 * (time.time() - t0)
        # run validation batches
        model.eval()
        val_loader.reset()
        val_loss = 0.0
        for _ in range(val_steps):
            with torch.no_grad():
                x_val, y_val = val_loader.next_batch()
                val_loss += model(x_val, y_val, attn_blocksize=attn_blocksize)
        dist.all_reduce(val_loss, op=dist.ReduceOp.AVG)
        val_loss /= val_steps
        # log val loss to console and to logfile
        print0(f'step:{step}/{args.num_iterations} val_loss:{val_loss:.4f} train_time:{training_time_ms:.0f}ms step_avg:{training_time_ms/(timed_steps-1):.2f}ms')
        # start the clock again
        torch.cuda.synchronize()
        t0 = time.time()

    if master_process and (last_step or (args.save_every > 0 and step % args.save_every == 0)):
        # stop the clock
        torch.cuda.synchronize()
        training_time_ms += 1000 * (time.time() - t0)
        # save the state of the training process
        log = dict(step=step, code=code, model=raw_model.state_dict(), optimizers=[opt.state_dict() for opt in optimizers])
        torch.save(log, 'logs/%s/state_step%06d.pt' % (run_id, step))
        # start the clock again
        torch.cuda.synchronize()
        t0 = time.time()

    # bit confusing: we want to make sure to eval on 0th iteration
    # but also after the very last iteration. so we loop for step <= num_iterations
    # instead of just < num_iterations (one extra due to <=), only to do
    # the validation/sampling one last time, and then we break right here as we're done.
    if last_step:
        break

    # --------------- TRAINING SECTION BEGIN -----------------
    model.train()
    for i in range(1, train_accumulation_steps+1):
        # forward pass
        loss = model(x, y, attn_blocksize=attn_blocksize)
        train_loss = loss.detach()
        # advance the dataset for the next batch
        x, y = train_loader.next_batch()
        # backward pass
        if i < train_accumulation_steps:
            with model.no_sync(): # there's no need to sync gradients every accumulation step
                loss.backward()
        else:
            loss.backward() # just sync on the last step
    for p in model.parameters():
        p.grad /= train_accumulation_steps
    # momentum warmup for Muon
    frac = min(step/300, 1)
    optimizer3.param_groups[0]['momentum'] = (1 - frac) * 0.85 + frac * 0.95
    # step the optimizers and schedulers
    for opt, sched in zip(optimizers, schedulers):
        opt.step()
        sched.step()
    # null the gradients
    model.zero_grad(set_to_none=True)
    # --------------- TRAINING SECTION END -------------------
    # everything that follows now is just diagnostics, prints, logging, etc.

    #dist.all_reduce(train_loss, op=dist.ReduceOp.AVG) # all-reducing the training loss would be more correct in terms of logging, but slower
    approx_time = training_time_ms + 1000 * (time.time() - t0)
    print0(f"step:{step+1}/{args.num_iterations} train_loss:{train_loss.item():.4f} train_time:{approx_time:.0f}ms step_avg:{approx_time/timed_steps:.2f}ms")

if master_process:
    print(f"peak memory consumption: {torch.cuda.max_memory_allocated() // 1024 // 1024} MiB")

# -------------------------------------------------------------------------
# clean up nice
dist.destroy_process_group()
====================================================================================================
Running pytorch 2.6.0.dev20241124+cu124 compiled for CUDA 12.4
nvidia-smi:
Sun Nov 24 23:58:04 2024       
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 555.42.06              Driver Version: 555.42.06      CUDA Version: 12.5     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA H100 80GB HBM3          Off |   00000000:18:00.0 Off |                    0 |
| N/A   32C    P0             69W /  700W |       4MiB /  81559MiB |      0%      Default |
|                                         |                        |             Disabled |
+-----------------------------------------+------------------------+----------------------+
|   1  NVIDIA H100 80GB HBM3          Off |   00000000:2A:00.0 Off |                    0 |
| N/A   37C    P0            116W /  700W |      36MiB /  81559MiB |      0%      Default |
|                                         |                        |             Disabled |
+-----------------------------------------+------------------------+----------------------+
|   2  NVIDIA H100 80GB HBM3          Off |   00000000:3A:00.0 Off |                    0 |
| N/A   39C    P0            114W /  700W |     530MiB /  81559MiB |      0%      Default |
|                                         |                        |             Disabled |
+-----------------------------------------+------------------------+----------------------+
|   3  NVIDIA H100 80GB HBM3          Off |   00000000:5D:00.0 Off |                    0 |
| N/A   32C    P0            114W /  700W |     530MiB /  81559MiB |      0%      Default |
|                                         |                        |             Disabled |
+-----------------------------------------+------------------------+----------------------+
|   4  NVIDIA H100 80GB HBM3          Off |   00000000:84:00.0 Off |                    0 |
| N/A   32C    P0            112W /  700W |      36MiB /  81559MiB |      0%      Default |
|                                         |                        |             Disabled |
+-----------------------------------------+------------------------+----------------------+
|   5  NVIDIA H100 80GB HBM3          Off |   00000000:8B:00.0 Off |                    0 |
| N/A   38C    P0            117W /  700W |     530MiB /  81559MiB |      0%      Default |
|                                         |                        |             Disabled |
+-----------------------------------------+------------------------+----------------------+
|   6  NVIDIA H100 80GB HBM3          Off |   00000000:91:00.0 Off |                    0 |
| N/A   35C    P0            113W /  700W |     530MiB /  81559MiB |      0%      Default |
|                                         |                        |             Disabled |
+-----------------------------------------+------------------------+----------------------+
|   7  NVIDIA H100 80GB HBM3          Off |   00000000:E4:00.0 Off |                    0 |
| N/A   32C    P0            116W /  700W |     530MiB /  81559MiB |      0%      Default |
|                                         |                        |             Disabled |
+-----------------------------------------+------------------------+----------------------+
                                                                                         
+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI        PID   Type   Process name                              GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|    1   N/A  N/A     31614      C   /usr/bin/python3                                0MiB |
|    2   N/A  N/A     31615      C   /usr/bin/python3                                0MiB |
|    3   N/A  N/A     31616      C   /usr/bin/python3                                0MiB |
|    4   N/A  N/A     31617      C   /usr/bin/python3                                0MiB |
|    5   N/A  N/A     31618      C   /usr/bin/python3                                0MiB |
|    6   N/A  N/A     31619      C   /usr/bin/python3                                0MiB |
|    7   N/A  N/A     31620      C   /usr/bin/python3                                0MiB |
+-----------------------------------------------------------------------------------------+

====================================================================================================
Training DataLoader: total number of tokens: 1800000000 across 18 files
Validation DataLoader: total number of tokens: 100000000 across 1 files
====================================================================================================
step:0/1750 val_loss:10.8258 train_time:0ms step_avg:nanms
step:1/1750 train_loss:10.8258 train_time:24777ms step_avg:nanms
step:2/1750 train_loss:10.0696 train_time:24886ms step_avg:nanms
step:3/1750 train_loss:8.3467 train_time:25030ms step_avg:nanms
step:4/1750 train_loss:7.5907 train_time:25177ms step_avg:nanms
step:5/1750 train_loss:7.4986 train_time:25323ms step_avg:nanms
step:6/1750 train_loss:6.9888 train_time:25470ms step_avg:nanms
step:7/1750 train_loss:7.1867 train_time:25618ms step_avg:nanms
step:8/1750 train_loss:6.7462 train_time:25765ms step_avg:nanms
step:9/1750 train_loss:6.6415 train_time:25913ms step_avg:nanms
step:10/1750 train_loss:6.5114 train_time:26063ms step_avg:nanms
step:11/1750 train_loss:6.4809 train_time:110ms step_avg:nanms
step:12/1750 train_loss:6.3664 train_time:258ms step_avg:nanms
step:13/1750 train_loss:6.2703 train_time:404ms step_avg:134.66ms
step:14/1750 train_loss:6.1918 train_time:552ms step_avg:138.01ms
step:15/1750 train_loss:6.1907 train_time:699ms step_avg:139.89ms
step:16/1750 train_loss:6.1304 train_time:847ms step_avg:141.11ms
step:17/1750 train_loss:6.2163 train_time:994ms step_avg:141.96ms
step:18/1750 train_loss:6.0114 train_time:1143ms step_avg:142.85ms
step:19/1750 train_loss:6.0561 train_time:1289ms step_avg:143.18ms
step:20/1750 train_loss:5.7053 train_time:1438ms step_avg:143.76ms
step:21/1750 train_loss:6.0442 train_time:1584ms step_avg:144.00ms
step:22/1750 train_loss:6.2556 train_time:1732ms step_avg:144.29ms
step:23/1750 train_loss:5.9100 train_time:1879ms step_avg:144.55ms
step:24/1750 train_loss:6.0771 train_time:2026ms step_avg:144.71ms
step:25/1750 train_loss:5.7909 train_time:2173ms step_avg:144.89ms
step:26/1750 train_loss:5.7070 train_time:2321ms step_avg:145.09ms
step:27/1750 train_loss:5.9017 train_time:2468ms step_avg:145.17ms
step:28/1750 train_loss:5.5325 train_time:2617ms step_avg:145.36ms
step:29/1750 train_loss:5.7746 train_time:2765ms step_avg:145.51ms
step:30/1750 train_loss:5.5632 train_time:2911ms step_avg:145.53ms
step:31/1750 train_loss:5.5483 train_time:3060ms step_avg:145.69ms
step:32/1750 train_loss:5.3963 train_time:3205ms step_avg:145.70ms
step:33/1750 train_loss:5.7032 train_time:3353ms step_avg:145.79ms
step:34/1750 train_loss:5.6029 train_time:3500ms step_avg:145.84ms
step:35/1750 train_loss:5.7279 train_time:3647ms step_avg:145.87ms
step:36/1750 train_loss:5.6539 train_time:3794ms step_avg:145.92ms
step:37/1750 train_loss:5.5536 train_time:3942ms step_avg:145.99ms
step:38/1750 train_loss:5.4202 train_time:4088ms step_avg:146.01ms
step:39/1750 train_loss:5.4139 train_time:4237ms step_avg:146.11ms
step:40/1750 train_loss:5.3482 train_time:4384ms step_avg:146.15ms
step:41/1750 train_loss:5.3373 train_time:4532ms step_avg:146.19ms
step:42/1750 train_loss:5.2742 train_time:4679ms step_avg:146.23ms
step:43/1750 train_loss:5.3617 train_time:4826ms step_avg:146.26ms
step:44/1750 train_loss:5.3213 train_time:4975ms step_avg:146.33ms
step:45/1750 train_loss:5.4671 train_time:5122ms step_avg:146.36ms
step:46/1750 train_loss:5.2637 train_time:5269ms step_avg:146.36ms
step:47/1750 train_loss:5.1617 train_time:5417ms step_avg:146.41ms
step:48/1750 train_loss:5.2806 train_time:5564ms step_avg:146.42ms
step:49/1750 train_loss:5.2290 train_time:5711ms step_avg:146.42ms
step:50/1750 train_loss:5.3534 train_time:5859ms step_avg:146.48ms
step:51/1750 train_loss:5.2376 train_time:6006ms step_avg:146.48ms
step:52/1750 train_loss:5.1136 train_time:6154ms step_avg:146.52ms
step:53/1750 train_loss:5.2500 train_time:6301ms step_avg:146.52ms
step:54/1750 train_loss:5.1074 train_time:6447ms step_avg:146.52ms
step:55/1750 train_loss:5.4920 train_time:6594ms step_avg:146.54ms
step:56/1750 train_loss:5.1318 train_time:6742ms step_avg:146.57ms
step:57/1750 train_loss:4.9750 train_time:6889ms step_avg:146.57ms
step:58/1750 train_loss:5.0893 train_time:7037ms step_avg:146.61ms
step:59/1750 train_loss:5.0847 train_time:7185ms step_avg:146.63ms
step:60/1750 train_loss:5.2131 train_time:7332ms step_avg:146.65ms
step:61/1750 train_loss:4.9507 train_time:7480ms step_avg:146.66ms
step:62/1750 train_loss:5.0609 train_time:7627ms step_avg:146.67ms
step:63/1750 train_loss:5.0448 train_time:7775ms step_avg:146.71ms
step:64/1750 train_loss:4.9298 train_time:7924ms step_avg:146.73ms
step:65/1750 train_loss:4.8900 train_time:8070ms step_avg:146.73ms
step:66/1750 train_loss:5.0596 train_time:8219ms step_avg:146.77ms
step:67/1750 train_loss:4.9403 train_time:8367ms step_avg:146.79ms
step:68/1750 train_loss:5.2433 train_time:8514ms step_avg:146.80ms
step:69/1750 train_loss:4.8122 train_time:8662ms step_avg:146.81ms
step:70/1750 train_loss:4.9353 train_time:8808ms step_avg:146.81ms
step:71/1750 train_loss:5.0737 train_time:8957ms step_avg:146.84ms
step:72/1750 train_loss:4.9935 train_time:9104ms step_avg:146.83ms
step:73/1750 train_loss:4.8437 train_time:9251ms step_avg:146.84ms
step:74/1750 train_loss:5.0014 train_time:9398ms step_avg:146.85ms
step:75/1750 train_loss:4.9576 train_time:9546ms step_avg:146.86ms
step:76/1750 train_loss:4.8840 train_time:9693ms step_avg:146.87ms
step:77/1750 train_loss:5.0027 train_time:9841ms step_avg:146.88ms
step:78/1750 train_loss:5.1963 train_time:9988ms step_avg:146.88ms
step:79/1750 train_loss:4.8986 train_time:10136ms step_avg:146.90ms
step:80/1750 train_loss:4.9474 train_time:10283ms step_avg:146.90ms
step:81/1750 train_loss:4.7491 train_time:10429ms step_avg:146.89ms
step:82/1750 train_loss:4.8971 train_time:10577ms step_avg:146.90ms
step:83/1750 train_loss:4.8451 train_time:10724ms step_avg:146.91ms
step:84/1750 train_loss:4.8604 train_time:10871ms step_avg:146.91ms
step:85/1750 train_loss:4.7199 train_time:11020ms step_avg:146.93ms
step:86/1750 train_loss:4.9357 train_time:11167ms step_avg:146.93ms
step:87/1750 train_loss:4.8148 train_time:11315ms step_avg:146.95ms
step:88/1750 train_loss:4.8341 train_time:11462ms step_avg:146.95ms
step:89/1750 train_loss:4.8021 train_time:11608ms step_avg:146.94ms
step:90/1750 train_loss:4.7301 train_time:11757ms step_avg:146.96ms
step:91/1750 train_loss:4.7091 train_time:11903ms step_avg:146.96ms
step:92/1750 train_loss:4.8536 train_time:12051ms step_avg:146.96ms
step:93/1750 train_loss:4.6666 train_time:12199ms step_avg:146.97ms
step:94/1750 train_loss:4.6935 train_time:12346ms step_avg:146.98ms
step:95/1750 train_loss:4.7336 train_time:12494ms step_avg:146.98ms
step:96/1750 train_loss:4.6578 train_time:12642ms step_avg:147.00ms
step:97/1750 train_loss:4.6947 train_time:12789ms step_avg:147.00ms
step:98/1750 train_loss:4.6431 train_time:12937ms step_avg:147.01ms
step:99/1750 train_loss:4.7432 train_time:13084ms step_avg:147.01ms
step:100/1750 train_loss:4.7288 train_time:13231ms step_avg:147.01ms
step:101/1750 train_loss:4.5786 train_time:13379ms step_avg:147.03ms
step:102/1750 train_loss:4.7507 train_time:13527ms step_avg:147.03ms
step:103/1750 train_loss:4.6192 train_time:13674ms step_avg:147.03ms
step:104/1750 train_loss:4.5831 train_time:13821ms step_avg:147.04ms
step:105/1750 train_loss:4.6001 train_time:13968ms step_avg:147.03ms
step:106/1750 train_loss:4.6512 train_time:14117ms step_avg:147.05ms
step:107/1750 train_loss:4.5581 train_time:14264ms step_avg:147.05ms
step:108/1750 train_loss:4.3875 train_time:14410ms step_avg:147.04ms
step:109/1750 train_loss:4.5229 train_time:14559ms step_avg:147.06ms
step:110/1750 train_loss:4.5067 train_time:14705ms step_avg:147.05ms
step:111/1750 train_loss:4.4567 train_time:14853ms step_avg:147.06ms
step:112/1750 train_loss:4.6268 train_time:15000ms step_avg:147.05ms
step:113/1750 train_loss:4.5246 train_time:15147ms step_avg:147.06ms
step:114/1750 train_loss:4.3935 train_time:15294ms step_avg:147.06ms
step:115/1750 train_loss:4.5388 train_time:15442ms step_avg:147.07ms
step:116/1750 train_loss:4.5041 train_time:15589ms step_avg:147.06ms
step:117/1750 train_loss:4.4267 train_time:15737ms step_avg:147.08ms
step:118/1750 train_loss:4.6428 train_time:15884ms step_avg:147.08ms
step:119/1750 train_loss:4.5073 train_time:16031ms step_avg:147.07ms
step:120/1750 train_loss:4.4162 train_time:16179ms step_avg:147.08ms
step:121/1750 train_loss:4.3540 train_time:16326ms step_avg:147.08ms
step:122/1750 train_loss:4.4850 train_time:16472ms step_avg:147.07ms
step:123/1750 train_loss:4.3359 train_time:16620ms step_avg:147.08ms
step:124/1750 train_loss:4.6332 train_time:16767ms step_avg:147.08ms
step:125/1750 train_loss:4.5138 train_time:16914ms step_avg:147.08ms
step:125/1750 val_loss:4.4507 train_time:16953ms step_avg:147.41ms
step:126/1750 train_loss:4.4644 train_time:17065ms step_avg:147.11ms
step:127/1750 train_loss:4.4881 train_time:17212ms step_avg:147.11ms
step:128/1750 train_loss:4.4251 train_time:17361ms step_avg:147.13ms
step:129/1750 train_loss:4.7386 train_time:17508ms step_avg:147.13ms
step:130/1750 train_loss:4.4233 train_time:17655ms step_avg:147.12ms
step:131/1750 train_loss:4.4400 train_time:17805ms step_avg:147.15ms
step:132/1750 train_loss:4.3813 train_time:17956ms step_avg:147.18ms
step:133/1750 train_loss:4.4800 train_time:18106ms step_avg:147.21ms
step:134/1750 train_loss:4.3044 train_time:18257ms step_avg:147.23ms
step:135/1750 train_loss:4.4815 train_time:18407ms step_avg:147.26ms
step:136/1750 train_loss:4.2433 train_time:18557ms step_avg:147.28ms
step:137/1750 train_loss:4.4004 train_time:18708ms step_avg:147.31ms
step:138/1750 train_loss:4.3156 train_time:18858ms step_avg:147.33ms
step:139/1750 train_loss:4.3997 train_time:19008ms step_avg:147.35ms
step:140/1750 train_loss:4.4837 train_time:19159ms step_avg:147.37ms
step:141/1750 train_loss:4.3306 train_time:19309ms step_avg:147.40ms
step:142/1750 train_loss:4.3378 train_time:19460ms step_avg:147.43ms
step:143/1750 train_loss:4.2807 train_time:19611ms step_avg:147.45ms
step:144/1750 train_loss:4.3778 train_time:19762ms step_avg:147.48ms
step:145/1750 train_loss:4.3263 train_time:19913ms step_avg:147.50ms
step:146/1750 train_loss:4.1943 train_time:20065ms step_avg:147.53ms
step:147/1750 train_loss:4.3476 train_time:20215ms step_avg:147.55ms
step:148/1750 train_loss:4.3867 train_time:20366ms step_avg:147.58ms
step:149/1750 train_loss:4.3244 train_time:20516ms step_avg:147.60ms
step:150/1750 train_loss:4.4586 train_time:20667ms step_avg:147.62ms
step:151/1750 train_loss:4.2908 train_time:20816ms step_avg:147.63ms
step:152/1750 train_loss:4.2967 train_time:20968ms step_avg:147.66ms
step:153/1750 train_loss:4.3887 train_time:21118ms step_avg:147.68ms
step:154/1750 train_loss:4.3787 train_time:21269ms step_avg:147.70ms
step:155/1750 train_loss:4.3055 train_time:21418ms step_avg:147.71ms
step:156/1750 train_loss:4.3808 train_time:21569ms step_avg:147.74ms
step:157/1750 train_loss:4.4349 train_time:21719ms step_avg:147.75ms
step:158/1750 train_loss:4.2651 train_time:21870ms step_avg:147.77ms
step:159/1750 train_loss:4.3347 train_time:22019ms step_avg:147.78ms
step:160/1750 train_loss:4.1479 train_time:22170ms step_avg:147.80ms
step:161/1750 train_loss:4.3648 train_time:22319ms step_avg:147.81ms
step:162/1750 train_loss:4.3835 train_time:22470ms step_avg:147.83ms
step:163/1750 train_loss:4.3663 train_time:22620ms step_avg:147.84ms
step:164/1750 train_loss:4.2088 train_time:22771ms step_avg:147.87ms
step:165/1750 train_loss:4.3004 train_time:22922ms step_avg:147.88ms
step:166/1750 train_loss:4.3666 train_time:23073ms step_avg:147.90ms
step:167/1750 train_loss:4.2215 train_time:23224ms step_avg:147.92ms
step:168/1750 train_loss:4.3022 train_time:23374ms step_avg:147.94ms
step:169/1750 train_loss:4.1673 train_time:23525ms step_avg:147.96ms
step:170/1750 train_loss:4.0411 train_time:23676ms step_avg:147.97ms
step:171/1750 train_loss:4.2174 train_time:23826ms step_avg:147.99ms
step:172/1750 train_loss:4.2258 train_time:23976ms step_avg:148.00ms
step:173/1750 train_loss:4.2819 train_time:24126ms step_avg:148.02ms
step:174/1750 train_loss:4.4530 train_time:24276ms step_avg:148.03ms
step:175/1750 train_loss:4.2824 train_time:24426ms step_avg:148.04ms
step:176/1750 train_loss:4.1088 train_time:24577ms step_avg:148.05ms
step:177/1750 train_loss:4.0896 train_time:24727ms step_avg:148.06ms
step:178/1750 train_loss:4.2171 train_time:24877ms step_avg:148.08ms
step:179/1750 train_loss:4.1554 train_time:25028ms step_avg:148.09ms
step:180/1750 train_loss:4.1382 train_time:25178ms step_avg:148.10ms
step:181/1750 train_loss:4.3258 train_time:25328ms step_avg:148.11ms
step:182/1750 train_loss:4.1919 train_time:25478ms step_avg:148.13ms
step:183/1750 train_loss:4.1468 train_time:25628ms step_avg:148.14ms
step:184/1750 train_loss:4.1486 train_time:25778ms step_avg:148.15ms
step:185/1750 train_loss:4.2305 train_time:25929ms step_avg:148.17ms
step:186/1750 train_loss:4.1965 train_time:26080ms step_avg:148.18ms
step:187/1750 train_loss:4.2586 train_time:26230ms step_avg:148.19ms
step:188/1750 train_loss:4.2004 train_time:26498ms step_avg:148.86ms
step:189/1750 train_loss:4.1594 train_time:26787ms step_avg:149.65ms
step:190/1750 train_loss:4.2467 train_time:26935ms step_avg:149.64ms
step:191/1750 train_loss:4.1130 train_time:27088ms step_avg:149.66ms
step:192/1750 train_loss:4.0537 train_time:27238ms step_avg:149.66ms
step:193/1750 train_loss:4.2861 train_time:27389ms step_avg:149.67ms
step:194/1750 train_loss:4.2126 train_time:27539ms step_avg:149.67ms
step:195/1750 train_loss:4.3910 train_time:27689ms step_avg:149.67ms
step:196/1750 train_loss:4.2034 train_time:27837ms step_avg:149.66ms
step:197/1750 train_loss:4.0668 train_time:27988ms step_avg:149.67ms
step:198/1750 train_loss:4.1939 train_time:28138ms step_avg:149.67ms
step:199/1750 train_loss:4.0465 train_time:28287ms step_avg:149.67ms
step:200/1750 train_loss:4.1395 train_time:28436ms step_avg:149.66ms
step:201/1750 train_loss:4.0131 train_time:28587ms step_avg:149.67ms
step:202/1750 train_loss:4.2632 train_time:28735ms step_avg:149.66ms
step:203/1750 train_loss:4.0881 train_time:28886ms step_avg:149.67ms
step:204/1750 train_loss:4.2088 train_time:29034ms step_avg:149.66ms
step:205/1750 train_loss:4.2534 train_time:29185ms step_avg:149.67ms
step:206/1750 train_loss:3.9534 train_time:29335ms step_avg:149.67ms
step:207/1750 train_loss:4.1011 train_time:29485ms step_avg:149.67ms
step:208/1750 train_loss:4.1121 train_time:29634ms step_avg:149.67ms
step:209/1750 train_loss:4.2551 train_time:29785ms step_avg:149.67ms
step:210/1750 train_loss:4.2014 train_time:29934ms step_avg:149.67ms
step:211/1750 train_loss:4.0724 train_time:30085ms step_avg:149.68ms
step:212/1750 train_loss:4.1425 train_time:30233ms step_avg:149.67ms
step:213/1750 train_loss:4.0658 train_time:30384ms step_avg:149.67ms
step:214/1750 train_loss:4.1253 train_time:30532ms step_avg:149.67ms
step:215/1750 train_loss:3.9665 train_time:30682ms step_avg:149.67ms
step:216/1750 train_loss:4.0293 train_time:30831ms step_avg:149.66ms
step:217/1750 train_loss:4.0257 train_time:30981ms step_avg:149.67ms
step:218/1750 train_loss:4.0992 train_time:31129ms step_avg:149.66ms
step:219/1750 train_loss:4.0863 train_time:31279ms step_avg:149.66ms
step:220/1750 train_loss:4.0971 train_time:31428ms step_avg:149.66ms
step:221/1750 train_loss:4.1065 train_time:31577ms step_avg:149.65ms
step:222/1750 train_loss:4.0103 train_time:31727ms step_avg:149.65ms
step:223/1750 train_loss:4.0136 train_time:31877ms step_avg:149.66ms
step:224/1750 train_loss:4.3169 train_time:32026ms step_avg:149.66ms
step:225/1750 train_loss:3.9171 train_time:32175ms step_avg:149.65ms
step:226/1750 train_loss:4.0016 train_time:32325ms step_avg:149.65ms
step:227/1750 train_loss:4.0015 train_time:32474ms step_avg:149.65ms
step:228/1750 train_loss:4.1646 train_time:32624ms step_avg:149.65ms
step:229/1750 train_loss:3.9555 train_time:32773ms step_avg:149.65ms
step:230/1750 train_loss:4.0622 train_time:32923ms step_avg:149.65ms
step:231/1750 train_loss:3.9216 train_time:33072ms step_avg:149.65ms
step:232/1750 train_loss:3.9859 train_time:33222ms step_avg:149.65ms
step:233/1750 train_loss:4.1094 train_time:33372ms step_avg:149.65ms
step:234/1750 train_loss:4.0518 train_time:33521ms step_avg:149.65ms
step:235/1750 train_loss:3.9276 train_time:33670ms step_avg:149.65ms
step:236/1750 train_loss:4.1127 train_time:33820ms step_avg:149.65ms
step:237/1750 train_loss:4.1071 train_time:33969ms step_avg:149.65ms
step:238/1750 train_loss:3.9597 train_time:34118ms step_avg:149.64ms
step:239/1750 train_loss:4.1019 train_time:34268ms step_avg:149.64ms
step:240/1750 train_loss:4.1376 train_time:34416ms step_avg:149.63ms
step:241/1750 train_loss:3.9991 train_time:34567ms step_avg:149.64ms
step:242/1750 train_loss:4.1676 train_time:34716ms step_avg:149.64ms
step:243/1750 train_loss:4.0457 train_time:34867ms step_avg:149.64ms
step:244/1750 train_loss:4.0946 train_time:35016ms step_avg:149.64ms
step:245/1750 train_loss:4.1717 train_time:35166ms step_avg:149.64ms
step:246/1750 train_loss:4.0879 train_time:35314ms step_avg:149.64ms
step:247/1750 train_loss:4.0292 train_time:35465ms step_avg:149.64ms
step:248/1750 train_loss:4.1398 train_time:35614ms step_avg:149.64ms
step:249/1750 train_loss:3.9439 train_time:35765ms step_avg:149.65ms
step:250/1750 train_loss:3.9950 train_time:35914ms step_avg:149.64ms
step:250/1750 val_loss:4.0355 train_time:35954ms step_avg:149.81ms
step:251/1750 train_loss:4.1055 train_time:36068ms step_avg:149.66ms
step:252/1750 train_loss:4.1859 train_time:36219ms step_avg:149.66ms
step:253/1750 train_loss:3.9581 train_time:36370ms step_avg:149.67ms
step:254/1750 train_loss:3.8976 train_time:36519ms step_avg:149.67ms
step:255/1750 train_loss:4.0989 train_time:36670ms step_avg:149.67ms
step:256/1750 train_loss:4.0101 train_time:36819ms step_avg:149.67ms
step:257/1750 train_loss:4.0108 train_time:36968ms step_avg:149.67ms
step:258/1750 train_loss:4.0120 train_time:37118ms step_avg:149.67ms
step:259/1750 train_loss:4.0626 train_time:37268ms step_avg:149.67ms
step:260/1750 train_loss:4.0938 train_time:37416ms step_avg:149.66ms
step:261/1750 train_loss:4.0408 train_time:37569ms step_avg:149.68ms
step:262/1750 train_loss:4.0166 train_time:37722ms step_avg:149.69ms
step:263/1750 train_loss:3.9120 train_time:37875ms step_avg:149.70ms
step:264/1750 train_loss:4.0106 train_time:38028ms step_avg:149.72ms
step:265/1750 train_loss:3.8899 train_time:38182ms step_avg:149.73ms
step:266/1750 train_loss:3.9409 train_time:38333ms step_avg:149.74ms
step:267/1750 train_loss:3.9416 train_time:38488ms step_avg:149.76ms
step:268/1750 train_loss:3.9738 train_time:38640ms step_avg:149.77ms
step:269/1750 train_loss:3.8844 train_time:38793ms step_avg:149.78ms
step:270/1750 train_loss:4.1297 train_time:38946ms step_avg:149.79ms
step:271/1750 train_loss:3.9857 train_time:39098ms step_avg:149.80ms
step:272/1750 train_loss:3.9345 train_time:39250ms step_avg:149.81ms
step:273/1750 train_loss:3.9625 train_time:39403ms step_avg:149.82ms
step:274/1750 train_loss:4.0496 train_time:39555ms step_avg:149.83ms
step:275/1750 train_loss:4.0761 train_time:39709ms step_avg:149.84ms
step:276/1750 train_loss:4.2472 train_time:39862ms step_avg:149.86ms
step:277/1750 train_loss:4.0559 train_time:40014ms step_avg:149.86ms
step:278/1750 train_loss:4.1059 train_time:40167ms step_avg:149.88ms
step:279/1750 train_loss:4.0092 train_time:40318ms step_avg:149.88ms
step:280/1750 train_loss:4.1939 train_time:40472ms step_avg:149.90ms
step:281/1750 train_loss:3.9941 train_time:40625ms step_avg:149.91ms
step:282/1750 train_loss:3.9595 train_time:40778ms step_avg:149.92ms
step:283/1750 train_loss:3.9266 train_time:40929ms step_avg:149.92ms
step:284/1750 train_loss:4.0678 train_time:41083ms step_avg:149.94ms
step:285/1750 train_loss:4.0828 train_time:41237ms step_avg:149.95ms
step:286/1750 train_loss:4.1145 train_time:41390ms step_avg:149.97ms
step:287/1750 train_loss:3.9309 train_time:41544ms step_avg:149.98ms
step:288/1750 train_loss:4.0378 train_time:41696ms step_avg:149.98ms
step:289/1750 train_loss:3.8987 train_time:41849ms step_avg:149.99ms
step:290/1750 train_loss:3.8786 train_time:42002ms step_avg:150.01ms
step:291/1750 train_loss:3.9370 train_time:42155ms step_avg:150.02ms
step:292/1750 train_loss:3.8850 train_time:42308ms step_avg:150.03ms
step:293/1750 train_loss:3.9300 train_time:42460ms step_avg:150.03ms
step:294/1750 train_loss:3.9585 train_time:42612ms step_avg:150.04ms
step:295/1750 train_loss:3.8593 train_time:42765ms step_avg:150.05ms
step:296/1750 train_loss:3.8881 train_time:42918ms step_avg:150.06ms
step:297/1750 train_loss:3.8986 train_time:43070ms step_avg:150.07ms
step:298/1750 train_loss:3.9986 train_time:43223ms step_avg:150.08ms
step:299/1750 train_loss:3.8432 train_time:43375ms step_avg:150.09ms
step:300/1750 train_loss:3.9925 train_time:43529ms step_avg:150.10ms
step:301/1750 train_loss:4.0030 train_time:43681ms step_avg:150.11ms
step:302/1750 train_loss:3.9604 train_time:43834ms step_avg:150.12ms
step:303/1750 train_loss:4.0005 train_time:43988ms step_avg:150.13ms
step:304/1750 train_loss:3.9916 train_time:44139ms step_avg:150.13ms
step:305/1750 train_loss:4.4823 train_time:44292ms step_avg:150.14ms
step:306/1750 train_loss:3.9576 train_time:44444ms step_avg:150.15ms
step:307/1750 train_loss:3.8585 train_time:44597ms step_avg:150.16ms
step:308/1750 train_loss:4.0127 train_time:44749ms step_avg:150.17ms
step:309/1750 train_loss:3.8867 train_time:44903ms step_avg:150.18ms
step:310/1750 train_loss:4.1041 train_time:45054ms step_avg:150.18ms
step:311/1750 train_loss:3.9494 train_time:45208ms step_avg:150.19ms
step:312/1750 train_loss:3.8851 train_time:45361ms step_avg:150.20ms
step:313/1750 train_loss:3.9726 train_time:45514ms step_avg:150.21ms
step:314/1750 train_loss:4.0888 train_time:45667ms step_avg:150.22ms
step:315/1750 train_loss:3.9700 train_time:45819ms step_avg:150.22ms
step:316/1750 train_loss:3.8169 train_time:45971ms step_avg:150.23ms
step:317/1750 train_loss:3.8987 train_time:46124ms step_avg:150.24ms
step:318/1750 train_loss:3.9488 train_time:46276ms step_avg:150.25ms
step:319/1750 train_loss:3.9085 train_time:46430ms step_avg:150.26ms
step:320/1750 train_loss:4.0333 train_time:46585ms step_avg:150.27ms
step:321/1750 train_loss:3.9807 train_time:46737ms step_avg:150.28ms
step:322/1750 train_loss:3.9558 train_time:46890ms step_avg:150.29ms
step:323/1750 train_loss:4.0369 train_time:47042ms step_avg:150.29ms
step:324/1750 train_loss:3.9728 train_time:47195ms step_avg:150.30ms
step:325/1750 train_loss:4.0418 train_time:47347ms step_avg:150.31ms
step:326/1750 train_loss:3.9135 train_time:47499ms step_avg:150.31ms
step:327/1750 train_loss:4.4162 train_time:47650ms step_avg:150.32ms
step:328/1750 train_loss:4.0950 train_time:47803ms step_avg:150.32ms
step:329/1750 train_loss:3.8207 train_time:47955ms step_avg:150.33ms
step:330/1750 train_loss:3.7715 train_time:48108ms step_avg:150.34ms
step:331/1750 train_loss:4.0001 train_time:48260ms step_avg:150.34ms
step:332/1750 train_loss:3.9230 train_time:48412ms step_avg:150.35ms
step:333/1750 train_loss:3.9008 train_time:48565ms step_avg:150.35ms
step:334/1750 train_loss:3.8609 train_time:48714ms step_avg:150.35ms
step:335/1750 train_loss:4.0277 train_time:48867ms step_avg:150.36ms
step:336/1750 train_loss:3.9841 train_time:49018ms step_avg:150.36ms
step:337/1750 train_loss:4.4431 train_time:49171ms step_avg:150.37ms
step:338/1750 train_loss:3.9586 train_time:49323ms step_avg:150.37ms
step:339/1750 train_loss:3.8836 train_time:49474ms step_avg:150.38ms
step:340/1750 train_loss:3.9543 train_time:49628ms step_avg:150.39ms
step:341/1750 train_loss:3.8759 train_time:49780ms step_avg:150.39ms
step:342/1750 train_loss:3.8353 train_time:49931ms step_avg:150.39ms
step:343/1750 train_loss:3.8698 train_time:50084ms step_avg:150.40ms
step:344/1750 train_loss:4.0117 train_time:50235ms step_avg:150.41ms
step:345/1750 train_loss:3.8353 train_time:50389ms step_avg:150.42ms
step:346/1750 train_loss:3.7886 train_time:50540ms step_avg:150.42ms
step:347/1750 train_loss:3.8239 train_time:50692ms step_avg:150.42ms
step:348/1750 train_loss:3.8781 train_time:50844ms step_avg:150.43ms
step:349/1750 train_loss:3.8504 train_time:50996ms step_avg:150.43ms
step:350/1750 train_loss:3.5906 train_time:51147ms step_avg:150.43ms
step:351/1750 train_loss:3.8537 train_time:51300ms step_avg:150.44ms
step:352/1750 train_loss:4.2139 train_time:51450ms step_avg:150.44ms
step:353/1750 train_loss:3.6782 train_time:51603ms step_avg:150.45ms
step:354/1750 train_loss:3.9453 train_time:51754ms step_avg:150.45ms
step:355/1750 train_loss:3.8095 train_time:51907ms step_avg:150.46ms
step:356/1750 train_loss:3.9071 train_time:52059ms step_avg:150.46ms
step:357/1750 train_loss:3.7887 train_time:52211ms step_avg:150.46ms
step:358/1750 train_loss:3.8878 train_time:52363ms step_avg:150.47ms
step:359/1750 train_loss:3.8005 train_time:52514ms step_avg:150.47ms
step:360/1750 train_loss:3.4489 train_time:52667ms step_avg:150.48ms
step:361/1750 train_loss:4.0485 train_time:52818ms step_avg:150.48ms
step:362/1750 train_loss:3.9511 train_time:52970ms step_avg:150.48ms
step:363/1750 train_loss:3.8699 train_time:53120ms step_avg:150.48ms
step:364/1750 train_loss:3.7611 train_time:53273ms step_avg:150.49ms
step:365/1750 train_loss:3.9412 train_time:53425ms step_avg:150.49ms
step:366/1750 train_loss:3.8867 train_time:53576ms step_avg:150.50ms
step:367/1750 train_loss:3.8772 train_time:53729ms step_avg:150.50ms
step:368/1750 train_loss:3.8662 train_time:53880ms step_avg:150.50ms
step:369/1750 train_loss:3.7594 train_time:54031ms step_avg:150.51ms
step:370/1750 train_loss:3.9123 train_time:54184ms step_avg:150.51ms
step:371/1750 train_loss:3.7604 train_time:54337ms step_avg:150.52ms
step:372/1750 train_loss:3.7145 train_time:54489ms step_avg:150.52ms
step:373/1750 train_loss:3.9365 train_time:54641ms step_avg:150.53ms
step:374/1750 train_loss:3.8541 train_time:54792ms step_avg:150.53ms
step:375/1750 train_loss:3.8226 train_time:54944ms step_avg:150.53ms
step:375/1750 val_loss:3.8486 train_time:54984ms step_avg:150.64ms
step:376/1750 train_loss:3.8913 train_time:55098ms step_avg:150.54ms
step:377/1750 train_loss:3.8162 train_time:55366ms step_avg:150.86ms
step:378/1750 train_loss:3.8797 train_time:55525ms step_avg:150.88ms
step:379/1750 train_loss:3.8952 train_time:55817ms step_avg:151.27ms
step:380/1750 train_loss:3.9755 train_time:55968ms step_avg:151.27ms
step:381/1750 train_loss:3.8685 train_time:56119ms step_avg:151.26ms
step:382/1750 train_loss:3.8355 train_time:56271ms step_avg:151.27ms
step:383/1750 train_loss:3.8199 train_time:56423ms step_avg:151.27ms
step:384/1750 train_loss:3.8924 train_time:56575ms step_avg:151.27ms
step:385/1750 train_loss:3.8118 train_time:56727ms step_avg:151.27ms
step:386/1750 train_loss:3.9192 train_time:56878ms step_avg:151.27ms
step:387/1750 train_loss:4.0907 train_time:57030ms step_avg:151.27ms
step:388/1750 train_loss:3.8186 train_time:57182ms step_avg:151.28ms
step:389/1750 train_loss:3.8193 train_time:57334ms step_avg:151.28ms
step:390/1750 train_loss:3.9184 train_time:57488ms step_avg:151.28ms
step:391/1750 train_loss:3.8329 train_time:57642ms step_avg:151.29ms
step:392/1750 train_loss:3.9426 train_time:57796ms step_avg:151.30ms
step:393/1750 train_loss:3.7825 train_time:57950ms step_avg:151.31ms
step:394/1750 train_loss:3.9099 train_time:58105ms step_avg:151.31ms
step:395/1750 train_loss:3.6567 train_time:58258ms step_avg:151.32ms
step:396/1750 train_loss:3.8548 train_time:58414ms step_avg:151.33ms
step:397/1750 train_loss:3.8871 train_time:58569ms step_avg:151.34ms
step:398/1750 train_loss:3.9015 train_time:58723ms step_avg:151.35ms
step:399/1750 train_loss:3.7902 train_time:58876ms step_avg:151.35ms
step:400/1750 train_loss:3.8407 train_time:59033ms step_avg:151.37ms
step:401/1750 train_loss:3.9293 train_time:59187ms step_avg:151.37ms
step:402/1750 train_loss:3.8603 train_time:59341ms step_avg:151.38ms
step:403/1750 train_loss:3.9766 train_time:59496ms step_avg:151.39ms
step:404/1750 train_loss:3.7019 train_time:59650ms step_avg:151.40ms
step:405/1750 train_loss:3.8036 train_time:59807ms step_avg:151.41ms
step:406/1750 train_loss:4.1142 train_time:59960ms step_avg:151.41ms
step:407/1750 train_loss:3.8023 train_time:60115ms step_avg:151.42ms
step:408/1750 train_loss:3.8428 train_time:60268ms step_avg:151.43ms
step:409/1750 train_loss:3.8805 train_time:60422ms step_avg:151.43ms
step:410/1750 train_loss:3.7815 train_time:60576ms step_avg:151.44ms
step:411/1750 train_loss:3.7846 train_time:60731ms step_avg:151.45ms
step:412/1750 train_loss:4.2108 train_time:60885ms step_avg:151.45ms
step:413/1750 train_loss:3.6314 train_time:61038ms step_avg:151.46ms
step:414/1750 train_loss:4.0363 train_time:61193ms step_avg:151.47ms
step:415/1750 train_loss:3.7788 train_time:61348ms step_avg:151.48ms
step:416/1750 train_loss:3.7870 train_time:61502ms step_avg:151.48ms
step:417/1750 train_loss:3.9755 train_time:61656ms step_avg:151.49ms
step:418/1750 train_loss:3.7070 train_time:61812ms step_avg:151.50ms
step:419/1750 train_loss:3.8260 train_time:61966ms step_avg:151.51ms
step:420/1750 train_loss:3.7287 train_time:62120ms step_avg:151.51ms
step:421/1750 train_loss:3.6723 train_time:62274ms step_avg:151.52ms
step:422/1750 train_loss:3.8063 train_time:62430ms step_avg:151.53ms
step:423/1750 train_loss:3.8906 train_time:62585ms step_avg:151.54ms
step:424/1750 train_loss:3.6412 train_time:62739ms step_avg:151.54ms
step:425/1750 train_loss:3.8231 train_time:62894ms step_avg:151.55ms
step:426/1750 train_loss:3.6763 train_time:63047ms step_avg:151.56ms
step:427/1750 train_loss:3.9122 train_time:63202ms step_avg:151.56ms
step:428/1750 train_loss:3.8358 train_time:63355ms step_avg:151.57ms
step:429/1750 train_loss:3.7793 train_time:63511ms step_avg:151.58ms
step:430/1750 train_loss:3.7384 train_time:63666ms step_avg:151.59ms
step:431/1750 train_loss:3.6392 train_time:63820ms step_avg:151.59ms
step:432/1750 train_loss:3.7891 train_time:63976ms step_avg:151.60ms
step:433/1750 train_loss:3.8407 train_time:64130ms step_avg:151.61ms
step:434/1750 train_loss:3.7939 train_time:64285ms step_avg:151.62ms
step:435/1750 train_loss:3.8345 train_time:64438ms step_avg:151.62ms
step:436/1750 train_loss:3.8574 train_time:64593ms step_avg:151.63ms
step:437/1750 train_loss:3.7301 train_time:64747ms step_avg:151.63ms
step:438/1750 train_loss:3.7284 train_time:64901ms step_avg:151.64ms
step:439/1750 train_loss:3.7287 train_time:65056ms step_avg:151.64ms
step:440/1750 train_loss:3.9135 train_time:65211ms step_avg:151.65ms
step:441/1750 train_loss:3.7876 train_time:65365ms step_avg:151.66ms
step:442/1750 train_loss:3.7634 train_time:65519ms step_avg:151.66ms
step:443/1750 train_loss:3.6439 train_time:65673ms step_avg:151.67ms
step:444/1750 train_loss:3.9457 train_time:65827ms step_avg:151.68ms
step:445/1750 train_loss:3.8663 train_time:65982ms step_avg:151.68ms
step:446/1750 train_loss:3.8545 train_time:66137ms step_avg:151.69ms
step:447/1750 train_loss:3.7778 train_time:66293ms step_avg:151.70ms
step:448/1750 train_loss:3.8731 train_time:66447ms step_avg:151.71ms
step:449/1750 train_loss:3.7080 train_time:66603ms step_avg:151.72ms
step:450/1750 train_loss:3.7561 train_time:66756ms step_avg:151.72ms
step:451/1750 train_loss:3.6026 train_time:66913ms step_avg:151.73ms
step:452/1750 train_loss:3.7382 train_time:67066ms step_avg:151.73ms
step:453/1750 train_loss:3.7001 train_time:67219ms step_avg:151.74ms
step:454/1750 train_loss:3.6587 train_time:67374ms step_avg:151.74ms
step:455/1750 train_loss:3.8633 train_time:67529ms step_avg:151.75ms
step:456/1750 train_loss:3.7446 train_time:67683ms step_avg:151.76ms
step:457/1750 train_loss:3.8078 train_time:67837ms step_avg:151.76ms
step:458/1750 train_loss:3.8542 train_time:67992ms step_avg:151.77ms
step:459/1750 train_loss:3.6534 train_time:68146ms step_avg:151.77ms
step:460/1750 train_loss:3.8169 train_time:68298ms step_avg:151.77ms
step:461/1750 train_loss:3.7154 train_time:68453ms step_avg:151.78ms
step:462/1750 train_loss:3.7619 train_time:68607ms step_avg:151.79ms
step:463/1750 train_loss:3.7974 train_time:68761ms step_avg:151.79ms
step:464/1750 train_loss:3.7351 train_time:68915ms step_avg:151.79ms
step:465/1750 train_loss:3.7407 train_time:69069ms step_avg:151.80ms
step:466/1750 train_loss:3.8210 train_time:69222ms step_avg:151.80ms
step:467/1750 train_loss:3.8442 train_time:69376ms step_avg:151.81ms
step:468/1750 train_loss:3.8177 train_time:69530ms step_avg:151.81ms
step:469/1750 train_loss:3.7065 train_time:69684ms step_avg:151.82ms
step:470/1750 train_loss:3.7904 train_time:69838ms step_avg:151.82ms
step:471/1750 train_loss:3.8402 train_time:69992ms step_avg:151.83ms
step:472/1750 train_loss:3.8023 train_time:70147ms step_avg:151.83ms
step:473/1750 train_loss:3.7368 train_time:70299ms step_avg:151.83ms
step:474/1750 train_loss:3.6066 train_time:70453ms step_avg:151.84ms
step:475/1750 train_loss:4.0365 train_time:70607ms step_avg:151.84ms
step:476/1750 train_loss:3.7847 train_time:70760ms step_avg:151.84ms
step:477/1750 train_loss:3.6023 train_time:70914ms step_avg:151.85ms
step:478/1750 train_loss:3.8478 train_time:71067ms step_avg:151.85ms
step:479/1750 train_loss:3.7965 train_time:71220ms step_avg:151.85ms
step:480/1750 train_loss:3.9408 train_time:71375ms step_avg:151.86ms
step:481/1750 train_loss:3.7533 train_time:71530ms step_avg:151.87ms
step:482/1750 train_loss:3.5525 train_time:71684ms step_avg:151.87ms
step:483/1750 train_loss:3.8330 train_time:71836ms step_avg:151.87ms
step:484/1750 train_loss:3.6857 train_time:71992ms step_avg:151.88ms
step:485/1750 train_loss:3.6884 train_time:72146ms step_avg:151.89ms
step:486/1750 train_loss:3.6028 train_time:72300ms step_avg:151.89ms
step:487/1750 train_loss:3.6978 train_time:72453ms step_avg:151.89ms
step:488/1750 train_loss:3.9012 train_time:72608ms step_avg:151.90ms
step:489/1750 train_loss:3.7308 train_time:72760ms step_avg:151.90ms
step:490/1750 train_loss:3.6133 train_time:72914ms step_avg:151.90ms
step:491/1750 train_loss:3.6402 train_time:73067ms step_avg:151.91ms
step:492/1750 train_loss:3.7564 train_time:73220ms step_avg:151.91ms
step:493/1750 train_loss:3.6022 train_time:73375ms step_avg:151.91ms
step:494/1750 train_loss:3.7258 train_time:73529ms step_avg:151.92ms
step:495/1750 train_loss:3.6794 train_time:73684ms step_avg:151.93ms
step:496/1750 train_loss:3.5461 train_time:73837ms step_avg:151.93ms
step:497/1750 train_loss:3.7559 train_time:73992ms step_avg:151.93ms
step:498/1750 train_loss:3.8132 train_time:74145ms step_avg:151.94ms
step:499/1750 train_loss:3.8436 train_time:74299ms step_avg:151.94ms
step:500/1750 train_loss:3.7537 train_time:74452ms step_avg:151.94ms
step:500/1750 val_loss:3.7296 train_time:74491ms step_avg:152.02ms
step:501/1750 train_loss:3.8309 train_time:74606ms step_avg:151.95ms
step:502/1750 train_loss:3.7686 train_time:74763ms step_avg:151.96ms
step:503/1750 train_loss:3.7964 train_time:74919ms step_avg:151.96ms
step:504/1750 train_loss:3.7516 train_time:75071ms step_avg:151.97ms
step:505/1750 train_loss:3.8285 train_time:75225ms step_avg:151.97ms
step:506/1750 train_loss:3.6724 train_time:75381ms step_avg:151.98ms
step:507/1750 train_loss:3.7879 train_time:75533ms step_avg:151.98ms
step:508/1750 train_loss:3.8496 train_time:75687ms step_avg:151.98ms
step:509/1750 train_loss:3.7961 train_time:75840ms step_avg:151.98ms
step:510/1750 train_loss:3.6045 train_time:75995ms step_avg:151.99ms
step:511/1750 train_loss:3.7989 train_time:76147ms step_avg:151.99ms
step:512/1750 train_loss:3.7533 train_time:76302ms step_avg:152.00ms
step:513/1750 train_loss:3.6911 train_time:76456ms step_avg:152.00ms
step:514/1750 train_loss:3.8385 train_time:76611ms step_avg:152.01ms
step:515/1750 train_loss:3.7597 train_time:76765ms step_avg:152.01ms
step:516/1750 train_loss:4.0961 train_time:76920ms step_avg:152.02ms
step:517/1750 train_loss:3.7090 train_time:77074ms step_avg:152.02ms
step:518/1750 train_loss:3.8025 train_time:77227ms step_avg:152.02ms
step:519/1750 train_loss:3.6847 train_time:77382ms step_avg:152.03ms
step:520/1750 train_loss:3.7115 train_time:77538ms step_avg:152.04ms
step:521/1750 train_loss:3.6836 train_time:77694ms step_avg:152.04ms
step:522/1750 train_loss:3.6692 train_time:77850ms step_avg:152.05ms
step:523/1750 train_loss:4.3149 train_time:78007ms step_avg:152.06ms
step:524/1750 train_loss:3.7617 train_time:78162ms step_avg:152.07ms
step:525/1750 train_loss:3.7026 train_time:78318ms step_avg:152.07ms
step:526/1750 train_loss:3.7230 train_time:78476ms step_avg:152.09ms
step:527/1750 train_loss:3.6805 train_time:78632ms step_avg:152.09ms
step:528/1750 train_loss:3.6466 train_time:78787ms step_avg:152.10ms
step:529/1750 train_loss:3.8722 train_time:78942ms step_avg:152.10ms
step:530/1750 train_loss:3.6667 train_time:79100ms step_avg:152.12ms
step:531/1750 train_loss:3.9400 train_time:79258ms step_avg:152.13ms
step:532/1750 train_loss:3.7569 train_time:79413ms step_avg:152.13ms
step:533/1750 train_loss:3.6776 train_time:79569ms step_avg:152.14ms
step:534/1750 train_loss:3.6910 train_time:79724ms step_avg:152.14ms
step:535/1750 train_loss:3.6270 train_time:79882ms step_avg:152.16ms
step:536/1750 train_loss:3.7738 train_time:80039ms step_avg:152.16ms
step:537/1750 train_loss:3.7497 train_time:80195ms step_avg:152.17ms
step:538/1750 train_loss:3.6469 train_time:80352ms step_avg:152.18ms
step:539/1750 train_loss:4.1372 train_time:80510ms step_avg:152.19ms
step:540/1750 train_loss:3.6942 train_time:80665ms step_avg:152.20ms
step:541/1750 train_loss:3.8095 train_time:80820ms step_avg:152.20ms
step:542/1750 train_loss:3.6165 train_time:80976ms step_avg:152.21ms
step:543/1750 train_loss:3.6081 train_time:81132ms step_avg:152.22ms
step:544/1750 train_loss:3.6687 train_time:81289ms step_avg:152.23ms
step:545/1750 train_loss:3.6076 train_time:81444ms step_avg:152.23ms
step:546/1750 train_loss:3.6525 train_time:81602ms step_avg:152.24ms
step:547/1750 train_loss:3.6712 train_time:81759ms step_avg:152.25ms
step:548/1750 train_loss:3.6364 train_time:81917ms step_avg:152.26ms
step:549/1750 train_loss:3.7477 train_time:82073ms step_avg:152.27ms
step:550/1750 train_loss:3.6379 train_time:82231ms step_avg:152.28ms
step:551/1750 train_loss:3.6593 train_time:82386ms step_avg:152.29ms
step:552/1750 train_loss:3.9619 train_time:82543ms step_avg:152.29ms
step:553/1750 train_loss:3.7820 train_time:82700ms step_avg:152.30ms
step:554/1750 train_loss:3.7381 train_time:82855ms step_avg:152.31ms
step:555/1750 train_loss:3.6526 train_time:83011ms step_avg:152.31ms
step:556/1750 train_loss:3.7212 train_time:83167ms step_avg:152.32ms
step:557/1750 train_loss:3.3420 train_time:83324ms step_avg:152.33ms
step:558/1750 train_loss:3.6319 train_time:83481ms step_avg:152.34ms
step:559/1750 train_loss:3.6749 train_time:83637ms step_avg:152.34ms
step:560/1750 train_loss:3.7212 train_time:83795ms step_avg:152.35ms
step:561/1750 train_loss:3.6338 train_time:83949ms step_avg:152.36ms
step:562/1750 train_loss:3.5794 train_time:84107ms step_avg:152.37ms
step:563/1750 train_loss:3.7879 train_time:84263ms step_avg:152.37ms
step:564/1750 train_loss:3.5975 train_time:84419ms step_avg:152.38ms
step:565/1750 train_loss:3.7086 train_time:84575ms step_avg:152.39ms
step:566/1750 train_loss:3.6426 train_time:84855ms step_avg:152.62ms
step:567/1750 train_loss:3.6192 train_time:85019ms step_avg:152.64ms
step:568/1750 train_loss:3.7098 train_time:85173ms step_avg:152.64ms
step:569/1750 train_loss:3.6721 train_time:85480ms step_avg:152.92ms
step:570/1750 train_loss:3.7147 train_time:85637ms step_avg:152.92ms
step:571/1750 train_loss:3.7823 train_time:85795ms step_avg:152.93ms
step:572/1750 train_loss:3.7487 train_time:85950ms step_avg:152.94ms
step:573/1750 train_loss:3.7565 train_time:86107ms step_avg:152.94ms
step:574/1750 train_loss:3.7974 train_time:86265ms step_avg:152.95ms
step:575/1750 train_loss:3.7523 train_time:86421ms step_avg:152.96ms
step:576/1750 train_loss:3.7806 train_time:86577ms step_avg:152.96ms
step:577/1750 train_loss:3.6906 train_time:86733ms step_avg:152.97ms
step:578/1750 train_loss:3.7012 train_time:86890ms step_avg:152.97ms
step:579/1750 train_loss:3.6937 train_time:87045ms step_avg:152.98ms
step:580/1750 train_loss:3.6251 train_time:87203ms step_avg:152.99ms
step:581/1750 train_loss:3.6567 train_time:87360ms step_avg:152.99ms
step:582/1750 train_loss:3.8727 train_time:87515ms step_avg:153.00ms
step:583/1750 train_loss:3.6504 train_time:87672ms step_avg:153.00ms
step:584/1750 train_loss:3.6159 train_time:87829ms step_avg:153.01ms
step:585/1750 train_loss:3.8110 train_time:87984ms step_avg:153.02ms
step:586/1750 train_loss:3.5372 train_time:88139ms step_avg:153.02ms
step:587/1750 train_loss:3.6901 train_time:88294ms step_avg:153.02ms
step:588/1750 train_loss:3.6775 train_time:88448ms step_avg:153.02ms
step:589/1750 train_loss:4.0185 train_time:88606ms step_avg:153.03ms
step:590/1750 train_loss:3.8041 train_time:88763ms step_avg:153.04ms
step:591/1750 train_loss:3.5317 train_time:88917ms step_avg:153.04ms
step:592/1750 train_loss:3.5558 train_time:89073ms step_avg:153.05ms
step:593/1750 train_loss:3.5217 train_time:89228ms step_avg:153.05ms
step:594/1750 train_loss:3.5771 train_time:89384ms step_avg:153.05ms
step:595/1750 train_loss:3.9537 train_time:89539ms step_avg:153.06ms
step:596/1750 train_loss:3.6698 train_time:89695ms step_avg:153.06ms
step:597/1750 train_loss:3.6099 train_time:89849ms step_avg:153.07ms
step:598/1750 train_loss:3.6850 train_time:90006ms step_avg:153.07ms
step:599/1750 train_loss:3.4953 train_time:90163ms step_avg:153.08ms
step:600/1750 train_loss:3.6234 train_time:90318ms step_avg:153.08ms
step:601/1750 train_loss:3.6726 train_time:90474ms step_avg:153.09ms
step:602/1750 train_loss:3.6843 train_time:90629ms step_avg:153.09ms
step:603/1750 train_loss:3.8052 train_time:90786ms step_avg:153.10ms
step:604/1750 train_loss:3.6317 train_time:90941ms step_avg:153.10ms
step:605/1750 train_loss:3.6360 train_time:91098ms step_avg:153.11ms
step:606/1750 train_loss:3.5978 train_time:91254ms step_avg:153.11ms
step:607/1750 train_loss:3.8590 train_time:91410ms step_avg:153.12ms
step:608/1750 train_loss:3.6602 train_time:91566ms step_avg:153.12ms
step:609/1750 train_loss:3.6322 train_time:91721ms step_avg:153.12ms
step:610/1750 train_loss:3.7274 train_time:91878ms step_avg:153.13ms
step:611/1750 train_loss:3.6225 train_time:92035ms step_avg:153.14ms
step:612/1750 train_loss:3.5948 train_time:92192ms step_avg:153.14ms
step:613/1750 train_loss:3.7845 train_time:92347ms step_avg:153.15ms
step:614/1750 train_loss:3.7260 train_time:92503ms step_avg:153.15ms
step:615/1750 train_loss:3.7206 train_time:92657ms step_avg:153.15ms
step:616/1750 train_loss:3.6569 train_time:92812ms step_avg:153.15ms
step:617/1750 train_loss:3.5798 train_time:92969ms step_avg:153.16ms
step:618/1750 train_loss:3.7096 train_time:93125ms step_avg:153.17ms
step:619/1750 train_loss:3.5809 train_time:93282ms step_avg:153.17ms
step:620/1750 train_loss:3.6089 train_time:93437ms step_avg:153.18ms
step:621/1750 train_loss:3.9493 train_time:93593ms step_avg:153.18ms
step:622/1750 train_loss:3.5883 train_time:93748ms step_avg:153.18ms
step:623/1750 train_loss:3.6215 train_time:93904ms step_avg:153.19ms
step:624/1750 train_loss:3.7155 train_time:94060ms step_avg:153.19ms
step:625/1750 train_loss:3.7296 train_time:94216ms step_avg:153.20ms
step:625/1750 val_loss:3.6469 train_time:94257ms step_avg:153.26ms
step:626/1750 train_loss:3.7654 train_time:94373ms step_avg:153.20ms
step:627/1750 train_loss:3.7407 train_time:94529ms step_avg:153.21ms
step:628/1750 train_loss:3.7881 train_time:94684ms step_avg:153.21ms
step:629/1750 train_loss:3.6172 train_time:94840ms step_avg:153.22ms
step:630/1750 train_loss:3.7482 train_time:94994ms step_avg:153.22ms
step:631/1750 train_loss:3.7696 train_time:95149ms step_avg:153.22ms
step:632/1750 train_loss:3.6762 train_time:95305ms step_avg:153.22ms
step:633/1750 train_loss:3.6298 train_time:95462ms step_avg:153.23ms
step:634/1750 train_loss:3.7221 train_time:95618ms step_avg:153.23ms
step:635/1750 train_loss:3.9760 train_time:95772ms step_avg:153.24ms
step:636/1750 train_loss:3.5741 train_time:95928ms step_avg:153.24ms
step:637/1750 train_loss:3.3733 train_time:96084ms step_avg:153.24ms
step:638/1750 train_loss:3.6136 train_time:96239ms step_avg:153.25ms
step:639/1750 train_loss:3.6535 train_time:96393ms step_avg:153.25ms
step:640/1750 train_loss:3.5928 train_time:96549ms step_avg:153.25ms
step:641/1750 train_loss:3.6076 train_time:96704ms step_avg:153.26ms
step:642/1750 train_loss:3.6510 train_time:96859ms step_avg:153.26ms
step:643/1750 train_loss:3.6302 train_time:97016ms step_avg:153.26ms
step:644/1750 train_loss:3.5883 train_time:97171ms step_avg:153.27ms
step:645/1750 train_loss:3.8026 train_time:97327ms step_avg:153.27ms
step:646/1750 train_loss:3.7012 train_time:97484ms step_avg:153.28ms
step:647/1750 train_loss:3.6852 train_time:97638ms step_avg:153.28ms
step:648/1750 train_loss:3.7368 train_time:97793ms step_avg:153.28ms
step:649/1750 train_loss:3.7835 train_time:97949ms step_avg:153.28ms
step:650/1750 train_loss:3.6483 train_time:98107ms step_avg:153.29ms
step:651/1750 train_loss:3.7940 train_time:98267ms step_avg:153.30ms
step:652/1750 train_loss:3.6091 train_time:98424ms step_avg:153.31ms
step:653/1750 train_loss:3.6838 train_time:98582ms step_avg:153.32ms
step:654/1750 train_loss:3.4531 train_time:98740ms step_avg:153.32ms
step:655/1750 train_loss:3.5995 train_time:98897ms step_avg:153.33ms
step:656/1750 train_loss:3.6043 train_time:99053ms step_avg:153.33ms
step:657/1750 train_loss:3.5250 train_time:99212ms step_avg:153.34ms
step:658/1750 train_loss:3.7122 train_time:99371ms step_avg:153.35ms
step:659/1750 train_loss:3.6130 train_time:99528ms step_avg:153.36ms
step:660/1750 train_loss:3.7051 train_time:99688ms step_avg:153.37ms
step:661/1750 train_loss:3.7876 train_time:99847ms step_avg:153.37ms
step:662/1750 train_loss:3.6921 train_time:100003ms step_avg:153.38ms
step:663/1750 train_loss:3.5723 train_time:100159ms step_avg:153.38ms
step:664/1750 train_loss:3.6329 train_time:100317ms step_avg:153.39ms
step:665/1750 train_loss:3.5146 train_time:100476ms step_avg:153.40ms
step:666/1750 train_loss:3.8112 train_time:100632ms step_avg:153.40ms
step:667/1750 train_loss:3.6386 train_time:100791ms step_avg:153.41ms
step:668/1750 train_loss:3.6661 train_time:100949ms step_avg:153.42ms
step:669/1750 train_loss:3.5063 train_time:101107ms step_avg:153.43ms
step:670/1750 train_loss:3.6210 train_time:101265ms step_avg:153.43ms
step:671/1750 train_loss:3.5826 train_time:101423ms step_avg:153.44ms
step:672/1750 train_loss:3.5956 train_time:101582ms step_avg:153.45ms
step:673/1750 train_loss:3.8698 train_time:101741ms step_avg:153.46ms
step:674/1750 train_loss:3.6520 train_time:101899ms step_avg:153.46ms
step:675/1750 train_loss:3.7382 train_time:102058ms step_avg:153.47ms
step:676/1750 train_loss:3.5081 train_time:102218ms step_avg:153.48ms
step:677/1750 train_loss:3.6214 train_time:102375ms step_avg:153.49ms
step:678/1750 train_loss:3.5751 train_time:102532ms step_avg:153.49ms
step:679/1750 train_loss:3.6985 train_time:102692ms step_avg:153.50ms
step:680/1750 train_loss:3.6129 train_time:102850ms step_avg:153.51ms
step:681/1750 train_loss:3.6417 train_time:103007ms step_avg:153.51ms
step:682/1750 train_loss:3.6935 train_time:103166ms step_avg:153.52ms
step:683/1750 train_loss:3.7550 train_time:103324ms step_avg:153.53ms
step:684/1750 train_loss:3.6704 train_time:103484ms step_avg:153.54ms
step:685/1750 train_loss:3.7177 train_time:103643ms step_avg:153.54ms
step:686/1750 train_loss:3.6573 train_time:103802ms step_avg:153.55ms
step:687/1750 train_loss:3.6950 train_time:103959ms step_avg:153.56ms
step:688/1750 train_loss:3.2441 train_time:104120ms step_avg:153.57ms
step:689/1750 train_loss:3.4295 train_time:104278ms step_avg:153.58ms
step:690/1750 train_loss:3.5723 train_time:104438ms step_avg:153.59ms
step:691/1750 train_loss:3.4395 train_time:104595ms step_avg:153.59ms
step:692/1750 train_loss:3.6512 train_time:104752ms step_avg:153.60ms
step:693/1750 train_loss:3.6713 train_time:104910ms step_avg:153.60ms
step:694/1750 train_loss:3.5789 train_time:105068ms step_avg:153.61ms
step:695/1750 train_loss:3.5581 train_time:105225ms step_avg:153.61ms
step:696/1750 train_loss:3.8785 train_time:105384ms step_avg:153.62ms
step:697/1750 train_loss:3.6134 train_time:105543ms step_avg:153.63ms
step:698/1750 train_loss:3.6709 train_time:105700ms step_avg:153.63ms
step:699/1750 train_loss:3.7916 train_time:105860ms step_avg:153.64ms
step:700/1750 train_loss:3.5911 train_time:106017ms step_avg:153.65ms
step:701/1750 train_loss:3.5695 train_time:106174ms step_avg:153.65ms
step:702/1750 train_loss:3.5419 train_time:106333ms step_avg:153.66ms
step:703/1750 train_loss:3.5181 train_time:106492ms step_avg:153.67ms
step:704/1750 train_loss:3.6000 train_time:106650ms step_avg:153.67ms
step:705/1750 train_loss:3.5888 train_time:106810ms step_avg:153.68ms
step:706/1750 train_loss:3.6079 train_time:106972ms step_avg:153.69ms
step:707/1750 train_loss:3.6722 train_time:107130ms step_avg:153.70ms
step:708/1750 train_loss:3.6354 train_time:107289ms step_avg:153.71ms
step:709/1750 train_loss:3.6059 train_time:107448ms step_avg:153.72ms
step:710/1750 train_loss:3.5696 train_time:107606ms step_avg:153.72ms
step:711/1750 train_loss:3.6200 train_time:107766ms step_avg:153.73ms
step:712/1750 train_loss:3.6767 train_time:107926ms step_avg:153.74ms
step:713/1750 train_loss:3.6801 train_time:108087ms step_avg:153.75ms
step:714/1750 train_loss:3.5798 train_time:108243ms step_avg:153.75ms
step:715/1750 train_loss:3.5936 train_time:108400ms step_avg:153.76ms
step:716/1750 train_loss:3.6079 train_time:108558ms step_avg:153.76ms
step:717/1750 train_loss:3.7392 train_time:108716ms step_avg:153.77ms
step:718/1750 train_loss:3.6196 train_time:108872ms step_avg:153.77ms
step:719/1750 train_loss:3.6999 train_time:109029ms step_avg:153.78ms
step:720/1750 train_loss:3.8742 train_time:109187ms step_avg:153.79ms
step:721/1750 train_loss:3.4894 train_time:109344ms step_avg:153.79ms
step:722/1750 train_loss:3.7578 train_time:109502ms step_avg:153.79ms
step:723/1750 train_loss:3.7955 train_time:109659ms step_avg:153.80ms
step:724/1750 train_loss:3.5901 train_time:109817ms step_avg:153.81ms
step:725/1750 train_loss:3.6728 train_time:109976ms step_avg:153.81ms
step:726/1750 train_loss:3.5604 train_time:110135ms step_avg:153.82ms
step:727/1750 train_loss:3.5981 train_time:110294ms step_avg:153.83ms
step:728/1750 train_loss:3.7619 train_time:110451ms step_avg:153.83ms
step:729/1750 train_loss:3.6934 train_time:110609ms step_avg:153.84ms
step:730/1750 train_loss:3.6999 train_time:110767ms step_avg:153.84ms
step:731/1750 train_loss:3.5842 train_time:110923ms step_avg:153.85ms
step:732/1750 train_loss:3.6199 train_time:111080ms step_avg:153.85ms
step:733/1750 train_loss:3.8554 train_time:111236ms step_avg:153.85ms
step:734/1750 train_loss:3.5845 train_time:111394ms step_avg:153.86ms
step:735/1750 train_loss:3.6351 train_time:111551ms step_avg:153.86ms
step:736/1750 train_loss:3.7611 train_time:111709ms step_avg:153.87ms
step:737/1750 train_loss:3.6985 train_time:111867ms step_avg:153.87ms
step:738/1750 train_loss:3.6199 train_time:112023ms step_avg:153.88ms
step:739/1750 train_loss:3.5344 train_time:112180ms step_avg:153.88ms
step:740/1750 train_loss:4.1345 train_time:112340ms step_avg:153.89ms
step:741/1750 train_loss:3.5186 train_time:112496ms step_avg:153.89ms
step:742/1750 train_loss:3.5877 train_time:112654ms step_avg:153.90ms
step:743/1750 train_loss:3.6071 train_time:112811ms step_avg:153.90ms
step:744/1750 train_loss:3.6714 train_time:112969ms step_avg:153.91ms
step:745/1750 train_loss:3.6121 train_time:113127ms step_avg:153.91ms
step:746/1750 train_loss:3.6172 train_time:113284ms step_avg:153.92ms
step:747/1750 train_loss:3.6744 train_time:113440ms step_avg:153.92ms
step:748/1750 train_loss:3.5999 train_time:113601ms step_avg:153.93ms
step:749/1750 train_loss:3.5857 train_time:113762ms step_avg:153.94ms
step:750/1750 train_loss:3.6257 train_time:113918ms step_avg:153.94ms
step:750/1750 val_loss:3.5925 train_time:113961ms step_avg:154.00ms
step:751/1750 train_loss:3.5960 train_time:114079ms step_avg:153.95ms
step:752/1750 train_loss:3.6368 train_time:114238ms step_avg:153.96ms
step:753/1750 train_loss:3.6376 train_time:114395ms step_avg:153.96ms
step:754/1750 train_loss:3.6160 train_time:114551ms step_avg:153.97ms
step:755/1750 train_loss:3.7023 train_time:114828ms step_avg:154.13ms
step:756/1750 train_loss:3.4863 train_time:114993ms step_avg:154.15ms
step:757/1750 train_loss:3.7466 train_time:115155ms step_avg:154.16ms
step:758/1750 train_loss:3.6745 train_time:115310ms step_avg:154.16ms
step:759/1750 train_loss:3.6193 train_time:115604ms step_avg:154.34ms
step:760/1750 train_loss:3.7275 train_time:115759ms step_avg:154.35ms
step:761/1750 train_loss:3.4211 train_time:115917ms step_avg:154.35ms
step:762/1750 train_loss:3.5731 train_time:116074ms step_avg:154.35ms
step:763/1750 train_loss:3.6923 train_time:116229ms step_avg:154.35ms
step:764/1750 train_loss:3.3448 train_time:116388ms step_avg:154.36ms
step:765/1750 train_loss:3.7586 train_time:116545ms step_avg:154.36ms
step:766/1750 train_loss:3.5976 train_time:116704ms step_avg:154.37ms
step:767/1750 train_loss:3.5857 train_time:116862ms step_avg:154.38ms
step:768/1750 train_loss:3.5956 train_time:117021ms step_avg:154.38ms
step:769/1750 train_loss:3.6094 train_time:117179ms step_avg:154.39ms
step:770/1750 train_loss:3.6675 train_time:117336ms step_avg:154.39ms
step:771/1750 train_loss:3.9095 train_time:117494ms step_avg:154.39ms
step:772/1750 train_loss:3.4705 train_time:117652ms step_avg:154.40ms
step:773/1750 train_loss:3.6598 train_time:117808ms step_avg:154.40ms
step:774/1750 train_loss:3.6706 train_time:117965ms step_avg:154.40ms
step:775/1750 train_loss:3.6298 train_time:118121ms step_avg:154.41ms
step:776/1750 train_loss:3.4236 train_time:118278ms step_avg:154.41ms
step:777/1750 train_loss:3.4062 train_time:118437ms step_avg:154.42ms
step:778/1750 train_loss:3.5103 train_time:118594ms step_avg:154.42ms
step:779/1750 train_loss:3.6009 train_time:118752ms step_avg:154.42ms
step:780/1750 train_loss:3.6183 train_time:118911ms step_avg:154.43ms
step:781/1750 train_loss:3.6918 train_time:119069ms step_avg:154.43ms
step:782/1750 train_loss:3.6156 train_time:119230ms step_avg:154.44ms
step:783/1750 train_loss:3.5918 train_time:119387ms step_avg:154.45ms
step:784/1750 train_loss:3.6256 train_time:119547ms step_avg:154.45ms
step:785/1750 train_loss:3.5863 train_time:119706ms step_avg:154.46ms
step:786/1750 train_loss:3.4643 train_time:119866ms step_avg:154.47ms
step:787/1750 train_loss:3.7563 train_time:120026ms step_avg:154.47ms
step:788/1750 train_loss:3.5218 train_time:120185ms step_avg:154.48ms
step:789/1750 train_loss:3.5742 train_time:120344ms step_avg:154.49ms
step:790/1750 train_loss:3.6485 train_time:120503ms step_avg:154.49ms
step:791/1750 train_loss:3.7954 train_time:120664ms step_avg:154.50ms
step:792/1750 train_loss:3.7907 train_time:120825ms step_avg:154.51ms
step:793/1750 train_loss:3.4922 train_time:120983ms step_avg:154.51ms
step:794/1750 train_loss:3.6233 train_time:121144ms step_avg:154.52ms
step:795/1750 train_loss:3.6975 train_time:121304ms step_avg:154.53ms
step:796/1750 train_loss:3.7605 train_time:121465ms step_avg:154.54ms
step:797/1750 train_loss:3.5463 train_time:121624ms step_avg:154.54ms
step:798/1750 train_loss:3.6685 train_time:121785ms step_avg:154.55ms
step:799/1750 train_loss:3.5687 train_time:121946ms step_avg:154.56ms
step:800/1750 train_loss:3.5572 train_time:122104ms step_avg:154.56ms
step:801/1750 train_loss:3.6573 train_time:122264ms step_avg:154.57ms
step:802/1750 train_loss:3.5198 train_time:122427ms step_avg:154.58ms
step:803/1750 train_loss:3.5300 train_time:122585ms step_avg:154.58ms
step:804/1750 train_loss:3.6609 train_time:122744ms step_avg:154.59ms
step:805/1750 train_loss:3.5457 train_time:122906ms step_avg:154.60ms
step:806/1750 train_loss:3.5826 train_time:123064ms step_avg:154.60ms
step:807/1750 train_loss:3.6725 train_time:123224ms step_avg:154.61ms
step:808/1750 train_loss:3.5742 train_time:123384ms step_avg:154.62ms
step:809/1750 train_loss:3.5215 train_time:123542ms step_avg:154.62ms
step:810/1750 train_loss:3.5871 train_time:123703ms step_avg:154.63ms
step:811/1750 train_loss:3.6114 train_time:123863ms step_avg:154.64ms
step:812/1750 train_loss:3.6212 train_time:124022ms step_avg:154.64ms
step:813/1750 train_loss:3.6445 train_time:124181ms step_avg:154.65ms
step:814/1750 train_loss:3.5910 train_time:124341ms step_avg:154.65ms
step:815/1750 train_loss:3.5912 train_time:124500ms step_avg:154.66ms
step:816/1750 train_loss:3.7108 train_time:124662ms step_avg:154.67ms
step:817/1750 train_loss:3.7942 train_time:124822ms step_avg:154.67ms
step:818/1750 train_loss:3.5492 train_time:124979ms step_avg:154.68ms
step:819/1750 train_loss:3.7476 train_time:125139ms step_avg:154.68ms
step:820/1750 train_loss:3.5303 train_time:125298ms step_avg:154.69ms
step:821/1750 train_loss:3.5819 train_time:125458ms step_avg:154.69ms
step:822/1750 train_loss:3.7196 train_time:125618ms step_avg:154.70ms
step:823/1750 train_loss:3.6055 train_time:125776ms step_avg:154.71ms
step:824/1750 train_loss:3.5353 train_time:125935ms step_avg:154.71ms
step:825/1750 train_loss:3.6466 train_time:126097ms step_avg:154.72ms
step:826/1750 train_loss:3.5060 train_time:126260ms step_avg:154.73ms
step:827/1750 train_loss:3.7578 train_time:126419ms step_avg:154.74ms
step:828/1750 train_loss:3.6438 train_time:126579ms step_avg:154.74ms
step:829/1750 train_loss:3.6606 train_time:126738ms step_avg:154.75ms
step:830/1750 train_loss:3.5519 train_time:126898ms step_avg:154.75ms
step:831/1750 train_loss:3.6272 train_time:127058ms step_avg:154.76ms
step:832/1750 train_loss:3.5396 train_time:127218ms step_avg:154.77ms
step:833/1750 train_loss:3.6754 train_time:127379ms step_avg:154.77ms
step:834/1750 train_loss:3.5007 train_time:127539ms step_avg:154.78ms
step:835/1750 train_loss:3.4861 train_time:127698ms step_avg:154.79ms
step:836/1750 train_loss:3.7504 train_time:127860ms step_avg:154.79ms
step:837/1750 train_loss:3.4264 train_time:128018ms step_avg:154.80ms
step:838/1750 train_loss:3.6118 train_time:128178ms step_avg:154.80ms
step:839/1750 train_loss:3.4419 train_time:128338ms step_avg:154.81ms
step:840/1750 train_loss:3.4953 train_time:128495ms step_avg:154.81ms
step:841/1750 train_loss:3.5915 train_time:128653ms step_avg:154.82ms
step:842/1750 train_loss:3.6035 train_time:128815ms step_avg:154.83ms
step:843/1750 train_loss:3.5885 train_time:128974ms step_avg:154.83ms
step:844/1750 train_loss:3.4480 train_time:129133ms step_avg:154.84ms
step:845/1750 train_loss:3.6837 train_time:129294ms step_avg:154.84ms
step:846/1750 train_loss:3.5475 train_time:129455ms step_avg:154.85ms
step:847/1750 train_loss:3.5208 train_time:129614ms step_avg:154.86ms
step:848/1750 train_loss:3.6628 train_time:129772ms step_avg:154.86ms
step:849/1750 train_loss:3.5172 train_time:129930ms step_avg:154.86ms
step:850/1750 train_loss:3.4663 train_time:130089ms step_avg:154.87ms
step:851/1750 train_loss:3.7617 train_time:130253ms step_avg:154.88ms
step:852/1750 train_loss:3.4746 train_time:130411ms step_avg:154.88ms
step:853/1750 train_loss:3.5907 train_time:130570ms step_avg:154.89ms
step:854/1750 train_loss:3.6774 train_time:130730ms step_avg:154.89ms
step:855/1750 train_loss:3.5460 train_time:130891ms step_avg:154.90ms
step:856/1750 train_loss:3.5658 train_time:131050ms step_avg:154.91ms
step:857/1750 train_loss:3.6292 train_time:131209ms step_avg:154.91ms
step:858/1750 train_loss:3.5011 train_time:131371ms step_avg:154.92ms
step:859/1750 train_loss:3.5913 train_time:131529ms step_avg:154.92ms
step:860/1750 train_loss:3.6215 train_time:131686ms step_avg:154.92ms
step:861/1750 train_loss:3.6648 train_time:131849ms step_avg:154.93ms
step:862/1750 train_loss:3.6208 train_time:132010ms step_avg:154.94ms
step:863/1750 train_loss:3.5979 train_time:132171ms step_avg:154.95ms
step:864/1750 train_loss:3.4083 train_time:132329ms step_avg:154.95ms
step:865/1750 train_loss:3.6210 train_time:132487ms step_avg:154.96ms
step:866/1750 train_loss:3.9109 train_time:132648ms step_avg:154.96ms
step:867/1750 train_loss:3.4811 train_time:132805ms step_avg:154.97ms
step:868/1750 train_loss:3.6672 train_time:132963ms step_avg:154.97ms
step:869/1750 train_loss:3.6451 train_time:133122ms step_avg:154.97ms
step:870/1750 train_loss:3.4771 train_time:133284ms step_avg:154.98ms
step:871/1750 train_loss:3.4431 train_time:133444ms step_avg:154.99ms
step:872/1750 train_loss:3.6796 train_time:133602ms step_avg:154.99ms
step:873/1750 train_loss:3.4801 train_time:133762ms step_avg:155.00ms
step:874/1750 train_loss:3.2400 train_time:133922ms step_avg:155.00ms
step:875/1750 train_loss:3.6624 train_time:134080ms step_avg:155.01ms
step:875/1750 val_loss:3.5468 train_time:134120ms step_avg:155.05ms
step:876/1750 train_loss:3.4662 train_time:134240ms step_avg:155.01ms
step:877/1750 train_loss:3.6462 train_time:134402ms step_avg:155.02ms
step:878/1750 train_loss:3.4938 train_time:134561ms step_avg:155.02ms
step:879/1750 train_loss:3.6728 train_time:134722ms step_avg:155.03ms
step:880/1750 train_loss:3.3325 train_time:134880ms step_avg:155.03ms
step:881/1750 train_loss:3.5129 train_time:135038ms step_avg:155.04ms
step:882/1750 train_loss:3.7138 train_time:135195ms step_avg:155.04ms
step:883/1750 train_loss:3.8660 train_time:135356ms step_avg:155.05ms
step:884/1750 train_loss:3.5917 train_time:135516ms step_avg:155.05ms
step:885/1750 train_loss:3.5156 train_time:135674ms step_avg:155.06ms
step:886/1750 train_loss:3.5972 train_time:135833ms step_avg:155.06ms
step:887/1750 train_loss:4.1078 train_time:135995ms step_avg:155.07ms
step:888/1750 train_loss:3.8656 train_time:136159ms step_avg:155.08ms
step:889/1750 train_loss:3.5532 train_time:136317ms step_avg:155.08ms
step:890/1750 train_loss:3.5579 train_time:136476ms step_avg:155.09ms
step:891/1750 train_loss:3.3883 train_time:136636ms step_avg:155.09ms
step:892/1750 train_loss:3.7500 train_time:136795ms step_avg:155.10ms
step:893/1750 train_loss:3.4484 train_time:136951ms step_avg:155.10ms
step:894/1750 train_loss:3.6555 train_time:137112ms step_avg:155.10ms
step:895/1750 train_loss:3.7112 train_time:137271ms step_avg:155.11ms
step:896/1750 train_loss:3.5288 train_time:137431ms step_avg:155.11ms
step:897/1750 train_loss:3.5694 train_time:137591ms step_avg:155.12ms
step:898/1750 train_loss:3.6206 train_time:137750ms step_avg:155.12ms
step:899/1750 train_loss:3.5038 train_time:137909ms step_avg:155.13ms
step:900/1750 train_loss:3.4443 train_time:138067ms step_avg:155.13ms
step:901/1750 train_loss:3.6413 train_time:138225ms step_avg:155.14ms
step:902/1750 train_loss:3.6653 train_time:138382ms step_avg:155.14ms
step:903/1750 train_loss:3.5649 train_time:138544ms step_avg:155.14ms
step:904/1750 train_loss:3.5202 train_time:138705ms step_avg:155.15ms
step:905/1750 train_loss:3.5289 train_time:138863ms step_avg:155.15ms
step:906/1750 train_loss:3.7410 train_time:139024ms step_avg:155.16ms
step:907/1750 train_loss:3.5422 train_time:139184ms step_avg:155.17ms
step:908/1750 train_loss:3.5990 train_time:139341ms step_avg:155.17ms
step:909/1750 train_loss:3.4794 train_time:139503ms step_avg:155.18ms
step:910/1750 train_loss:3.5532 train_time:139667ms step_avg:155.19ms
step:911/1750 train_loss:3.6702 train_time:139827ms step_avg:155.19ms
step:912/1750 train_loss:3.6215 train_time:139990ms step_avg:155.20ms
step:913/1750 train_loss:3.4820 train_time:140152ms step_avg:155.21ms
step:914/1750 train_loss:3.7773 train_time:140314ms step_avg:155.22ms
step:915/1750 train_loss:3.5599 train_time:140477ms step_avg:155.22ms
step:916/1750 train_loss:3.6465 train_time:140639ms step_avg:155.23ms
step:917/1750 train_loss:3.6260 train_time:140799ms step_avg:155.24ms
step:918/1750 train_loss:4.8565 train_time:140962ms step_avg:155.24ms
step:919/1750 train_loss:3.5185 train_time:141124ms step_avg:155.25ms
step:920/1750 train_loss:3.6142 train_time:141283ms step_avg:155.26ms
step:921/1750 train_loss:3.5663 train_time:141446ms step_avg:155.26ms
step:922/1750 train_loss:3.6115 train_time:141610ms step_avg:155.27ms
step:923/1750 train_loss:3.6360 train_time:141768ms step_avg:155.28ms
step:924/1750 train_loss:3.7064 train_time:141929ms step_avg:155.28ms
step:925/1750 train_loss:3.6695 train_time:142090ms step_avg:155.29ms
step:926/1750 train_loss:3.5799 train_time:142251ms step_avg:155.30ms
step:927/1750 train_loss:3.5769 train_time:142413ms step_avg:155.30ms
step:928/1750 train_loss:3.8058 train_time:142575ms step_avg:155.31ms
step:929/1750 train_loss:3.6378 train_time:142736ms step_avg:155.32ms
step:930/1750 train_loss:3.4306 train_time:142898ms step_avg:155.32ms
step:931/1750 train_loss:3.5188 train_time:143057ms step_avg:155.33ms
step:932/1750 train_loss:3.6788 train_time:143220ms step_avg:155.34ms
step:933/1750 train_loss:3.4047 train_time:143381ms step_avg:155.34ms
step:934/1750 train_loss:3.6105 train_time:143542ms step_avg:155.35ms
step:935/1750 train_loss:3.4671 train_time:143706ms step_avg:155.36ms
step:936/1750 train_loss:3.5467 train_time:143868ms step_avg:155.37ms
step:937/1750 train_loss:3.6478 train_time:144033ms step_avg:155.38ms
step:938/1750 train_loss:3.5662 train_time:144192ms step_avg:155.38ms
step:939/1750 train_loss:3.6996 train_time:144358ms step_avg:155.39ms
step:940/1750 train_loss:3.5078 train_time:144518ms step_avg:155.40ms
step:941/1750 train_loss:3.5800 train_time:144678ms step_avg:155.40ms
step:942/1750 train_loss:3.3867 train_time:144840ms step_avg:155.41ms
step:943/1750 train_loss:3.7430 train_time:145005ms step_avg:155.42ms
step:944/1750 train_loss:3.4347 train_time:145292ms step_avg:155.56ms
step:945/1750 train_loss:3.4517 train_time:145460ms step_avg:155.57ms
step:946/1750 train_loss:5.0868 train_time:145626ms step_avg:155.58ms
step:947/1750 train_loss:3.6326 train_time:145786ms step_avg:155.59ms
step:948/1750 train_loss:3.5093 train_time:145946ms step_avg:155.59ms
step:949/1750 train_loss:3.4133 train_time:146251ms step_avg:155.75ms
step:950/1750 train_loss:3.4672 train_time:146410ms step_avg:155.76ms
step:951/1750 train_loss:3.4323 train_time:146572ms step_avg:155.76ms
step:952/1750 train_loss:3.5019 train_time:146733ms step_avg:155.77ms
step:953/1750 train_loss:3.5953 train_time:146897ms step_avg:155.78ms
step:954/1750 train_loss:3.4703 train_time:147061ms step_avg:155.79ms
step:955/1750 train_loss:3.5031 train_time:147222ms step_avg:155.79ms
step:956/1750 train_loss:3.4713 train_time:147384ms step_avg:155.80ms
step:957/1750 train_loss:3.5290 train_time:147546ms step_avg:155.80ms
step:958/1750 train_loss:3.5319 train_time:147709ms step_avg:155.81ms
step:959/1750 train_loss:3.5444 train_time:147869ms step_avg:155.82ms
step:960/1750 train_loss:3.4322 train_time:148031ms step_avg:155.82ms
step:961/1750 train_loss:3.6792 train_time:148190ms step_avg:155.83ms
step:962/1750 train_loss:3.6311 train_time:148351ms step_avg:155.83ms
step:963/1750 train_loss:3.5439 train_time:148515ms step_avg:155.84ms
step:964/1750 train_loss:3.4534 train_time:148678ms step_avg:155.85ms
step:965/1750 train_loss:3.5057 train_time:148837ms step_avg:155.85ms
step:966/1750 train_loss:3.7415 train_time:148999ms step_avg:155.86ms
step:967/1750 train_loss:3.5537 train_time:149158ms step_avg:155.86ms
step:968/1750 train_loss:3.5531 train_time:149320ms step_avg:155.87ms
step:969/1750 train_loss:3.6130 train_time:149482ms step_avg:155.87ms
step:970/1750 train_loss:3.4052 train_time:149641ms step_avg:155.88ms
step:971/1750 train_loss:3.5662 train_time:149800ms step_avg:155.88ms
step:972/1750 train_loss:3.5130 train_time:149959ms step_avg:155.88ms
step:973/1750 train_loss:3.5749 train_time:150120ms step_avg:155.89ms
step:974/1750 train_loss:3.6247 train_time:150282ms step_avg:155.89ms
step:975/1750 train_loss:3.5049 train_time:150443ms step_avg:155.90ms
step:976/1750 train_loss:3.7015 train_time:150601ms step_avg:155.90ms
step:977/1750 train_loss:3.6076 train_time:150760ms step_avg:155.90ms
step:978/1750 train_loss:3.3928 train_time:150921ms step_avg:155.91ms
step:979/1750 train_loss:3.6574 train_time:151080ms step_avg:155.91ms
step:980/1750 train_loss:3.4464 train_time:151241ms step_avg:155.92ms
step:981/1750 train_loss:3.6053 train_time:151403ms step_avg:155.92ms
step:982/1750 train_loss:3.5853 train_time:151562ms step_avg:155.93ms
step:983/1750 train_loss:3.5509 train_time:151723ms step_avg:155.93ms
step:984/1750 train_loss:3.5274 train_time:151881ms step_avg:155.94ms
step:985/1750 train_loss:3.6202 train_time:152042ms step_avg:155.94ms
step:986/1750 train_loss:3.4528 train_time:152204ms step_avg:155.95ms
step:987/1750 train_loss:3.5208 train_time:152362ms step_avg:155.95ms
step:988/1750 train_loss:3.5400 train_time:152523ms step_avg:155.95ms
step:989/1750 train_loss:3.4498 train_time:152683ms step_avg:155.96ms
step:990/1750 train_loss:3.6832 train_time:152844ms step_avg:155.96ms
step:991/1750 train_loss:3.5048 train_time:153003ms step_avg:155.97ms
step:992/1750 train_loss:3.4800 train_time:153169ms step_avg:155.98ms
step:993/1750 train_loss:3.5371 train_time:153333ms step_avg:155.98ms
step:994/1750 train_loss:3.6300 train_time:153492ms step_avg:155.99ms
step:995/1750 train_loss:3.5722 train_time:153651ms step_avg:155.99ms
step:996/1750 train_loss:3.4926 train_time:153809ms step_avg:155.99ms
step:997/1750 train_loss:3.8057 train_time:153968ms step_avg:156.00ms
step:998/1750 train_loss:3.4843 train_time:154128ms step_avg:156.00ms
step:999/1750 train_loss:3.6307 train_time:154287ms step_avg:156.00ms
step:1000/1750 train_loss:3.4865 train_time:154448ms step_avg:156.01ms
step:1000/1750 val_loss:3.5074 train_time:154491ms step_avg:156.05ms
step:1001/1750 train_loss:3.5414 train_time:154609ms step_avg:156.01ms
step:1002/1750 train_loss:3.4158 train_time:154771ms step_avg:156.02ms
step:1003/1750 train_loss:3.5963 train_time:154932ms step_avg:156.02ms
step:1004/1750 train_loss:3.6421 train_time:155094ms step_avg:156.03ms
step:1005/1750 train_loss:3.4249 train_time:155253ms step_avg:156.03ms
step:1006/1750 train_loss:3.5013 train_time:155412ms step_avg:156.04ms
step:1007/1750 train_loss:3.4800 train_time:155573ms step_avg:156.04ms
step:1008/1750 train_loss:3.6022 train_time:155734ms step_avg:156.05ms
step:1009/1750 train_loss:3.6999 train_time:155897ms step_avg:156.05ms
step:1010/1750 train_loss:3.5962 train_time:156056ms step_avg:156.06ms
step:1011/1750 train_loss:3.5765 train_time:156216ms step_avg:156.06ms
step:1012/1750 train_loss:3.4299 train_time:156376ms step_avg:156.06ms
step:1013/1750 train_loss:3.5797 train_time:156536ms step_avg:156.07ms
step:1014/1750 train_loss:3.6653 train_time:156697ms step_avg:156.07ms
step:1015/1750 train_loss:3.3748 train_time:156860ms step_avg:156.08ms
step:1016/1750 train_loss:3.4545 train_time:157022ms step_avg:156.09ms
step:1017/1750 train_loss:3.4475 train_time:157186ms step_avg:156.09ms
step:1018/1750 train_loss:3.4342 train_time:157345ms step_avg:156.10ms
step:1019/1750 train_loss:3.5621 train_time:157507ms step_avg:156.10ms
step:1020/1750 train_loss:3.4276 train_time:157670ms step_avg:156.11ms
step:1021/1750 train_loss:3.3945 train_time:157827ms step_avg:156.11ms
step:1022/1750 train_loss:3.5260 train_time:157988ms step_avg:156.11ms
step:1023/1750 train_loss:3.5551 train_time:158147ms step_avg:156.12ms
step:1024/1750 train_loss:3.5181 train_time:158307ms step_avg:156.12ms
step:1025/1750 train_loss:3.5255 train_time:158468ms step_avg:156.13ms
step:1026/1750 train_loss:3.6694 train_time:158626ms step_avg:156.13ms
step:1027/1750 train_loss:3.3641 train_time:158786ms step_avg:156.13ms
step:1028/1750 train_loss:3.4355 train_time:158949ms step_avg:156.14ms
step:1029/1750 train_loss:3.3666 train_time:159113ms step_avg:156.15ms
step:1030/1750 train_loss:3.5760 train_time:159273ms step_avg:156.15ms
step:1031/1750 train_loss:3.5594 train_time:159432ms step_avg:156.15ms
step:1032/1750 train_loss:3.7411 train_time:159594ms step_avg:156.16ms
step:1033/1750 train_loss:3.5419 train_time:159752ms step_avg:156.16ms
step:1034/1750 train_loss:3.4524 train_time:159914ms step_avg:156.17ms
step:1035/1750 train_loss:3.4906 train_time:160075ms step_avg:156.17ms
step:1036/1750 train_loss:3.5356 train_time:160234ms step_avg:156.17ms
step:1037/1750 train_loss:3.8422 train_time:160395ms step_avg:156.18ms
step:1038/1750 train_loss:3.6718 train_time:160556ms step_avg:156.18ms
step:1039/1750 train_loss:3.5584 train_time:160719ms step_avg:156.19ms
step:1040/1750 train_loss:3.4581 train_time:160882ms step_avg:156.20ms
step:1041/1750 train_loss:3.5304 train_time:161045ms step_avg:156.20ms
step:1042/1750 train_loss:3.5697 train_time:161204ms step_avg:156.21ms
step:1043/1750 train_loss:3.4905 train_time:161365ms step_avg:156.21ms
step:1044/1750 train_loss:3.5050 train_time:161526ms step_avg:156.21ms
step:1045/1750 train_loss:3.5672 train_time:161690ms step_avg:156.22ms
step:1046/1750 train_loss:3.4730 train_time:161850ms step_avg:156.23ms
step:1047/1750 train_loss:3.6885 train_time:162011ms step_avg:156.23ms
step:1048/1750 train_loss:3.5510 train_time:162175ms step_avg:156.24ms
step:1049/1750 train_loss:3.4515 train_time:162337ms step_avg:156.24ms
step:1050/1750 train_loss:3.4355 train_time:162502ms step_avg:156.25ms
step:1051/1750 train_loss:3.5469 train_time:162666ms step_avg:156.26ms
step:1052/1750 train_loss:3.4105 train_time:162828ms step_avg:156.26ms
step:1053/1750 train_loss:3.7459 train_time:162989ms step_avg:156.27ms
step:1054/1750 train_loss:3.5918 train_time:163152ms step_avg:156.28ms
step:1055/1750 train_loss:3.4267 train_time:163311ms step_avg:156.28ms
step:1056/1750 train_loss:3.5449 train_time:163472ms step_avg:156.28ms
step:1057/1750 train_loss:3.6274 train_time:163634ms step_avg:156.29ms
step:1058/1750 train_loss:3.3548 train_time:163798ms step_avg:156.30ms
step:1059/1750 train_loss:3.4267 train_time:163966ms step_avg:156.31ms
step:1060/1750 train_loss:3.4965 train_time:164126ms step_avg:156.31ms
step:1061/1750 train_loss:3.4678 train_time:164286ms step_avg:156.31ms
step:1062/1750 train_loss:3.4341 train_time:164448ms step_avg:156.32ms
step:1063/1750 train_loss:3.5164 train_time:164608ms step_avg:156.32ms
step:1064/1750 train_loss:3.4352 train_time:164769ms step_avg:156.33ms
step:1065/1750 train_loss:3.4105 train_time:164931ms step_avg:156.33ms
step:1066/1750 train_loss:3.4627 train_time:165094ms step_avg:156.34ms
step:1067/1750 train_loss:3.3338 train_time:165259ms step_avg:156.35ms
step:1068/1750 train_loss:3.4887 train_time:165420ms step_avg:156.35ms
step:1069/1750 train_loss:3.3533 train_time:165586ms step_avg:156.36ms
step:1070/1750 train_loss:3.6221 train_time:165747ms step_avg:156.36ms
step:1071/1750 train_loss:3.5633 train_time:165911ms step_avg:156.37ms
step:1072/1750 train_loss:3.4936 train_time:166072ms step_avg:156.38ms
step:1073/1750 train_loss:3.5771 train_time:166232ms step_avg:156.38ms
step:1074/1750 train_loss:3.4997 train_time:166395ms step_avg:156.39ms
step:1075/1750 train_loss:3.4500 train_time:166559ms step_avg:156.39ms
step:1076/1750 train_loss:3.8484 train_time:166722ms step_avg:156.40ms
step:1077/1750 train_loss:3.4911 train_time:166884ms step_avg:156.40ms
step:1078/1750 train_loss:3.1483 train_time:167053ms step_avg:156.42ms
step:1079/1750 train_loss:3.5923 train_time:167215ms step_avg:156.42ms
step:1080/1750 train_loss:3.4865 train_time:167379ms step_avg:156.43ms
step:1081/1750 train_loss:3.5679 train_time:167540ms step_avg:156.43ms
step:1082/1750 train_loss:3.6567 train_time:167702ms step_avg:156.44ms
step:1083/1750 train_loss:3.5614 train_time:167864ms step_avg:156.44ms
step:1084/1750 train_loss:3.5352 train_time:168025ms step_avg:156.45ms
step:1085/1750 train_loss:3.4909 train_time:168188ms step_avg:156.45ms
step:1086/1750 train_loss:3.6891 train_time:168350ms step_avg:156.46ms
step:1087/1750 train_loss:3.5670 train_time:168510ms step_avg:156.46ms
step:1088/1750 train_loss:3.4303 train_time:168675ms step_avg:156.47ms
step:1089/1750 train_loss:3.4437 train_time:168842ms step_avg:156.48ms
step:1090/1750 train_loss:3.5461 train_time:169006ms step_avg:156.49ms
step:1091/1750 train_loss:3.3444 train_time:169168ms step_avg:156.49ms
step:1092/1750 train_loss:3.5511 train_time:169329ms step_avg:156.50ms
step:1093/1750 train_loss:3.6707 train_time:169492ms step_avg:156.50ms
step:1094/1750 train_loss:3.5103 train_time:169652ms step_avg:156.51ms
step:1095/1750 train_loss:3.4813 train_time:169813ms step_avg:156.51ms
step:1096/1750 train_loss:3.4901 train_time:169979ms step_avg:156.52ms
step:1097/1750 train_loss:3.5505 train_time:170143ms step_avg:156.53ms
step:1098/1750 train_loss:3.6202 train_time:170307ms step_avg:156.53ms
step:1099/1750 train_loss:3.5876 train_time:170470ms step_avg:156.54ms
step:1100/1750 train_loss:3.4995 train_time:170632ms step_avg:156.54ms
step:1101/1750 train_loss:3.3538 train_time:170794ms step_avg:156.55ms
step:1102/1750 train_loss:3.3717 train_time:170960ms step_avg:156.56ms
step:1103/1750 train_loss:3.5105 train_time:171125ms step_avg:156.56ms
step:1104/1750 train_loss:3.3799 train_time:171286ms step_avg:156.57ms
step:1105/1750 train_loss:4.1256 train_time:171448ms step_avg:156.57ms
step:1106/1750 train_loss:3.2934 train_time:171608ms step_avg:156.58ms
step:1107/1750 train_loss:3.6304 train_time:171770ms step_avg:156.58ms
step:1108/1750 train_loss:3.4063 train_time:171928ms step_avg:156.58ms
step:1109/1750 train_loss:3.5663 train_time:172090ms step_avg:156.59ms
step:1110/1750 train_loss:3.4910 train_time:172250ms step_avg:156.59ms
step:1111/1750 train_loss:3.5479 train_time:172412ms step_avg:156.60ms
step:1112/1750 train_loss:3.6284 train_time:172574ms step_avg:156.60ms
step:1113/1750 train_loss:3.4994 train_time:172740ms step_avg:156.61ms
step:1114/1750 train_loss:3.4300 train_time:172904ms step_avg:156.62ms
step:1115/1750 train_loss:3.3145 train_time:173068ms step_avg:156.62ms
step:1116/1750 train_loss:3.4926 train_time:173227ms step_avg:156.62ms
step:1117/1750 train_loss:3.6610 train_time:173391ms step_avg:156.63ms
step:1118/1750 train_loss:3.6842 train_time:173554ms step_avg:156.64ms
step:1119/1750 train_loss:3.5422 train_time:173715ms step_avg:156.64ms
step:1120/1750 train_loss:3.5597 train_time:173879ms step_avg:156.65ms
step:1121/1750 train_loss:3.4537 train_time:174043ms step_avg:156.65ms
step:1122/1750 train_loss:3.5239 train_time:174203ms step_avg:156.66ms
step:1123/1750 train_loss:3.6569 train_time:174364ms step_avg:156.66ms
step:1124/1750 train_loss:3.4117 train_time:174524ms step_avg:156.66ms
step:1125/1750 train_loss:3.2901 train_time:174686ms step_avg:156.67ms
step:1125/1750 val_loss:3.4777 train_time:174727ms step_avg:156.71ms
step:1126/1750 train_loss:3.5425 train_time:174846ms step_avg:156.67ms
step:1127/1750 train_loss:3.7493 train_time:175010ms step_avg:156.68ms
step:1128/1750 train_loss:3.3004 train_time:175173ms step_avg:156.68ms
step:1129/1750 train_loss:3.6290 train_time:175338ms step_avg:156.69ms
step:1130/1750 train_loss:3.4490 train_time:175500ms step_avg:156.70ms
step:1131/1750 train_loss:3.4688 train_time:175665ms step_avg:156.70ms
step:1132/1750 train_loss:3.4337 train_time:175824ms step_avg:156.71ms
step:1133/1750 train_loss:3.5688 train_time:176102ms step_avg:156.81ms
step:1134/1750 train_loss:3.5162 train_time:176271ms step_avg:156.82ms
step:1135/1750 train_loss:3.5921 train_time:176431ms step_avg:156.83ms
step:1136/1750 train_loss:3.6186 train_time:176593ms step_avg:156.83ms
step:1137/1750 train_loss:3.5283 train_time:176756ms step_avg:156.84ms
step:1138/1750 train_loss:3.4185 train_time:176919ms step_avg:156.84ms
step:1139/1750 train_loss:3.7251 train_time:177214ms step_avg:156.97ms
step:1140/1750 train_loss:3.5302 train_time:177373ms step_avg:156.97ms
step:1141/1750 train_loss:3.6651 train_time:177537ms step_avg:156.97ms
step:1142/1750 train_loss:3.5231 train_time:177696ms step_avg:156.98ms
step:1143/1750 train_loss:3.4326 train_time:177857ms step_avg:156.98ms
step:1144/1750 train_loss:3.5111 train_time:178018ms step_avg:156.98ms
step:1145/1750 train_loss:3.6639 train_time:178177ms step_avg:156.98ms
step:1146/1750 train_loss:3.6203 train_time:178341ms step_avg:156.99ms
step:1147/1750 train_loss:3.5647 train_time:178502ms step_avg:156.99ms
step:1148/1750 train_loss:3.5685 train_time:178663ms step_avg:157.00ms
step:1149/1750 train_loss:3.4009 train_time:178826ms step_avg:157.00ms
step:1150/1750 train_loss:3.4388 train_time:178986ms step_avg:157.01ms
step:1151/1750 train_loss:3.3879 train_time:179150ms step_avg:157.01ms
step:1152/1750 train_loss:3.4785 train_time:179316ms step_avg:157.02ms
step:1153/1750 train_loss:3.4921 train_time:179478ms step_avg:157.02ms
step:1154/1750 train_loss:3.5913 train_time:179638ms step_avg:157.03ms
step:1155/1750 train_loss:3.3966 train_time:179801ms step_avg:157.03ms
step:1156/1750 train_loss:3.6023 train_time:179965ms step_avg:157.04ms
step:1157/1750 train_loss:3.5656 train_time:180128ms step_avg:157.04ms
step:1158/1750 train_loss:3.3296 train_time:180288ms step_avg:157.04ms
step:1159/1750 train_loss:3.4110 train_time:180450ms step_avg:157.05ms
step:1160/1750 train_loss:3.4002 train_time:180611ms step_avg:157.05ms
step:1161/1750 train_loss:3.1647 train_time:180774ms step_avg:157.06ms
step:1162/1750 train_loss:3.4917 train_time:180936ms step_avg:157.06ms
step:1163/1750 train_loss:3.4546 train_time:181097ms step_avg:157.07ms
step:1164/1750 train_loss:3.3543 train_time:181258ms step_avg:157.07ms
step:1165/1750 train_loss:3.3201 train_time:181417ms step_avg:157.07ms
step:1166/1750 train_loss:3.4530 train_time:181579ms step_avg:157.08ms
step:1167/1750 train_loss:3.4657 train_time:181740ms step_avg:157.08ms
step:1168/1750 train_loss:3.7944 train_time:181901ms step_avg:157.08ms
step:1169/1750 train_loss:3.4461 train_time:182065ms step_avg:157.09ms
step:1170/1750 train_loss:3.4552 train_time:182229ms step_avg:157.09ms
step:1171/1750 train_loss:3.3839 train_time:182392ms step_avg:157.10ms
step:1172/1750 train_loss:3.4927 train_time:182554ms step_avg:157.10ms
step:1173/1750 train_loss:3.6110 train_time:182722ms step_avg:157.11ms
step:1174/1750 train_loss:3.4508 train_time:182892ms step_avg:157.12ms
step:1175/1750 train_loss:3.4455 train_time:183057ms step_avg:157.13ms
step:1176/1750 train_loss:3.4890 train_time:183222ms step_avg:157.14ms
step:1177/1750 train_loss:3.5170 train_time:183390ms step_avg:157.15ms
step:1178/1750 train_loss:3.5690 train_time:183552ms step_avg:157.15ms
step:1179/1750 train_loss:3.4776 train_time:183713ms step_avg:157.15ms
step:1180/1750 train_loss:3.4268 train_time:183883ms step_avg:157.16ms
step:1181/1750 train_loss:3.4100 train_time:184044ms step_avg:157.17ms
step:1182/1750 train_loss:3.4613 train_time:184207ms step_avg:157.17ms
step:1183/1750 train_loss:3.4014 train_time:184372ms step_avg:157.18ms
step:1184/1750 train_loss:3.5787 train_time:184537ms step_avg:157.19ms
step:1185/1750 train_loss:3.6157 train_time:184702ms step_avg:157.19ms
step:1186/1750 train_loss:3.4306 train_time:184867ms step_avg:157.20ms
step:1187/1750 train_loss:3.4787 train_time:185039ms step_avg:157.21ms
step:1188/1750 train_loss:3.5150 train_time:185201ms step_avg:157.22ms
step:1189/1750 train_loss:3.3391 train_time:185366ms step_avg:157.22ms
step:1190/1750 train_loss:3.5173 train_time:185529ms step_avg:157.23ms
step:1191/1750 train_loss:3.6506 train_time:185694ms step_avg:157.23ms
step:1192/1750 train_loss:3.4630 train_time:185855ms step_avg:157.24ms
step:1193/1750 train_loss:3.3416 train_time:186017ms step_avg:157.24ms
step:1194/1750 train_loss:3.6264 train_time:186181ms step_avg:157.25ms
step:1195/1750 train_loss:3.4415 train_time:186347ms step_avg:157.25ms
step:1196/1750 train_loss:3.4568 train_time:186519ms step_avg:157.27ms
step:1197/1750 train_loss:3.3590 train_time:186684ms step_avg:157.27ms
step:1198/1750 train_loss:3.3717 train_time:186857ms step_avg:157.29ms
step:1199/1750 train_loss:3.4122 train_time:187021ms step_avg:157.29ms
step:1200/1750 train_loss:3.5159 train_time:187183ms step_avg:157.30ms
step:1201/1750 train_loss:3.5531 train_time:187347ms step_avg:157.30ms
step:1202/1750 train_loss:3.7302 train_time:187521ms step_avg:157.32ms
step:1203/1750 train_loss:3.4771 train_time:187686ms step_avg:157.32ms
step:1204/1750 train_loss:3.3791 train_time:187852ms step_avg:157.33ms
step:1205/1750 train_loss:3.4991 train_time:188014ms step_avg:157.33ms
step:1206/1750 train_loss:3.5378 train_time:188177ms step_avg:157.34ms
step:1207/1750 train_loss:3.5914 train_time:188341ms step_avg:157.34ms
step:1208/1750 train_loss:3.4697 train_time:188501ms step_avg:157.35ms
step:1209/1750 train_loss:3.3058 train_time:188666ms step_avg:157.35ms
step:1210/1750 train_loss:3.3735 train_time:188832ms step_avg:157.36ms
step:1211/1750 train_loss:3.4718 train_time:188997ms step_avg:157.37ms
step:1212/1750 train_loss:3.4614 train_time:189160ms step_avg:157.37ms
step:1213/1750 train_loss:3.4853 train_time:189324ms step_avg:157.38ms
step:1214/1750 train_loss:3.3326 train_time:189489ms step_avg:157.38ms
step:1215/1750 train_loss:3.4599 train_time:189653ms step_avg:157.39ms
step:1216/1750 train_loss:3.3984 train_time:189817ms step_avg:157.39ms
step:1217/1750 train_loss:3.3922 train_time:189981ms step_avg:157.40ms
step:1218/1750 train_loss:3.4849 train_time:190144ms step_avg:157.40ms
step:1219/1750 train_loss:3.3319 train_time:190313ms step_avg:157.41ms
step:1220/1750 train_loss:3.5432 train_time:190476ms step_avg:157.42ms
step:1221/1750 train_loss:3.5773 train_time:190639ms step_avg:157.42ms
step:1222/1750 train_loss:3.5101 train_time:190800ms step_avg:157.43ms
step:1223/1750 train_loss:3.3662 train_time:190964ms step_avg:157.43ms
step:1224/1750 train_loss:3.3314 train_time:191131ms step_avg:157.44ms
step:1225/1750 train_loss:3.4401 train_time:191293ms step_avg:157.44ms
step:1226/1750 train_loss:3.3938 train_time:191458ms step_avg:157.45ms
step:1227/1750 train_loss:3.3395 train_time:191622ms step_avg:157.45ms
step:1228/1750 train_loss:3.5166 train_time:191782ms step_avg:157.46ms
step:1229/1750 train_loss:3.4409 train_time:191948ms step_avg:157.46ms
step:1230/1750 train_loss:3.4620 train_time:192120ms step_avg:157.48ms
step:1231/1750 train_loss:3.6460 train_time:192285ms step_avg:157.48ms
step:1232/1750 train_loss:3.5627 train_time:192451ms step_avg:157.49ms
step:1233/1750 train_loss:3.4928 train_time:192615ms step_avg:157.49ms
step:1234/1750 train_loss:3.6547 train_time:192778ms step_avg:157.50ms
step:1235/1750 train_loss:3.3986 train_time:192942ms step_avg:157.50ms
step:1236/1750 train_loss:3.3601 train_time:193104ms step_avg:157.51ms
step:1237/1750 train_loss:3.3424 train_time:193267ms step_avg:157.51ms
step:1238/1750 train_loss:3.3578 train_time:193439ms step_avg:157.52ms
step:1239/1750 train_loss:3.3946 train_time:193601ms step_avg:157.53ms
step:1240/1750 train_loss:3.4518 train_time:193762ms step_avg:157.53ms
step:1241/1750 train_loss:3.4955 train_time:193926ms step_avg:157.54ms
step:1242/1750 train_loss:3.3649 train_time:194088ms step_avg:157.54ms
step:1243/1750 train_loss:3.4795 train_time:194254ms step_avg:157.55ms
step:1244/1750 train_loss:3.4789 train_time:194415ms step_avg:157.55ms
step:1245/1750 train_loss:3.4880 train_time:194577ms step_avg:157.55ms
step:1246/1750 train_loss:3.3081 train_time:194741ms step_avg:157.56ms
step:1247/1750 train_loss:3.4455 train_time:194902ms step_avg:157.56ms
step:1248/1750 train_loss:3.5089 train_time:195064ms step_avg:157.56ms
step:1249/1750 train_loss:3.4920 train_time:195225ms step_avg:157.57ms
step:1250/1750 train_loss:3.3700 train_time:195387ms step_avg:157.57ms
step:1250/1750 val_loss:3.4245 train_time:195432ms step_avg:157.61ms
step:1251/1750 train_loss:3.5666 train_time:195556ms step_avg:157.58ms
step:1252/1750 train_loss:3.4377 train_time:195716ms step_avg:157.58ms
step:1253/1750 train_loss:3.3770 train_time:195877ms step_avg:157.58ms
step:1254/1750 train_loss:3.4812 train_time:196041ms step_avg:157.59ms
step:1255/1750 train_loss:3.5898 train_time:196210ms step_avg:157.60ms
step:1256/1750 train_loss:3.3768 train_time:196375ms step_avg:157.60ms
step:1257/1750 train_loss:3.4363 train_time:196536ms step_avg:157.61ms
step:1258/1750 train_loss:3.4224 train_time:196702ms step_avg:157.61ms
step:1259/1750 train_loss:3.4005 train_time:196863ms step_avg:157.62ms
step:1260/1750 train_loss:3.2766 train_time:197025ms step_avg:157.62ms
step:1261/1750 train_loss:3.3696 train_time:197191ms step_avg:157.63ms
step:1262/1750 train_loss:3.3914 train_time:197356ms step_avg:157.63ms
step:1263/1750 train_loss:3.2965 train_time:197519ms step_avg:157.64ms
step:1264/1750 train_loss:3.5121 train_time:197681ms step_avg:157.64ms
step:1265/1750 train_loss:3.4992 train_time:197842ms step_avg:157.64ms
step:1266/1750 train_loss:3.5129 train_time:198007ms step_avg:157.65ms
step:1267/1750 train_loss:3.4367 train_time:198170ms step_avg:157.65ms
step:1268/1750 train_loss:3.4746 train_time:198333ms step_avg:157.66ms
step:1269/1750 train_loss:3.3227 train_time:198498ms step_avg:157.66ms
step:1270/1750 train_loss:3.1622 train_time:198660ms step_avg:157.67ms
step:1271/1750 train_loss:3.4635 train_time:198823ms step_avg:157.67ms
step:1272/1750 train_loss:3.4248 train_time:198981ms step_avg:157.67ms
step:1273/1750 train_loss:3.4578 train_time:199143ms step_avg:157.67ms
step:1274/1750 train_loss:3.4201 train_time:199306ms step_avg:157.68ms
step:1275/1750 train_loss:3.5012 train_time:199468ms step_avg:157.68ms
step:1276/1750 train_loss:3.5412 train_time:199629ms step_avg:157.68ms
step:1277/1750 train_loss:3.4782 train_time:199793ms step_avg:157.69ms
step:1278/1750 train_loss:3.4708 train_time:199953ms step_avg:157.69ms
step:1279/1750 train_loss:3.3264 train_time:200117ms step_avg:157.70ms
step:1280/1750 train_loss:3.4404 train_time:200284ms step_avg:157.70ms
step:1281/1750 train_loss:3.4936 train_time:200448ms step_avg:157.71ms
step:1282/1750 train_loss:3.5391 train_time:200609ms step_avg:157.71ms
step:1283/1750 train_loss:3.3971 train_time:200772ms step_avg:157.72ms
step:1284/1750 train_loss:3.4409 train_time:200933ms step_avg:157.72ms
step:1285/1750 train_loss:3.4261 train_time:201094ms step_avg:157.72ms
step:1286/1750 train_loss:3.4030 train_time:201257ms step_avg:157.72ms
step:1287/1750 train_loss:3.5565 train_time:201420ms step_avg:157.73ms
step:1288/1750 train_loss:3.3711 train_time:201585ms step_avg:157.73ms
step:1289/1750 train_loss:3.4522 train_time:201756ms step_avg:157.75ms
step:1290/1750 train_loss:3.5296 train_time:201924ms step_avg:157.75ms
step:1291/1750 train_loss:3.4527 train_time:202089ms step_avg:157.76ms
step:1292/1750 train_loss:3.5425 train_time:202254ms step_avg:157.76ms
step:1293/1750 train_loss:3.5828 train_time:202418ms step_avg:157.77ms
step:1294/1750 train_loss:3.5357 train_time:202582ms step_avg:157.77ms
step:1295/1750 train_loss:3.3564 train_time:202744ms step_avg:157.78ms
step:1296/1750 train_loss:3.4412 train_time:202908ms step_avg:157.78ms
step:1297/1750 train_loss:3.3470 train_time:203072ms step_avg:157.79ms
step:1298/1750 train_loss:3.3473 train_time:203236ms step_avg:157.79ms
step:1299/1750 train_loss:3.4616 train_time:203399ms step_avg:157.80ms
step:1300/1750 train_loss:3.4777 train_time:203560ms step_avg:157.80ms
step:1301/1750 train_loss:3.4710 train_time:203723ms step_avg:157.80ms
step:1302/1750 train_loss:3.6416 train_time:203891ms step_avg:157.81ms
step:1303/1750 train_loss:3.3702 train_time:204060ms step_avg:157.82ms
step:1304/1750 train_loss:3.5746 train_time:204225ms step_avg:157.82ms
step:1305/1750 train_loss:3.3371 train_time:204387ms step_avg:157.83ms
step:1306/1750 train_loss:3.5161 train_time:204555ms step_avg:157.84ms
step:1307/1750 train_loss:3.5292 train_time:204716ms step_avg:157.84ms
step:1308/1750 train_loss:3.3643 train_time:204881ms step_avg:157.84ms
step:1309/1750 train_loss:3.3714 train_time:205047ms step_avg:157.85ms
step:1310/1750 train_loss:3.3710 train_time:205211ms step_avg:157.85ms
step:1311/1750 train_loss:3.3610 train_time:205373ms step_avg:157.86ms
step:1312/1750 train_loss:3.4528 train_time:205539ms step_avg:157.86ms
step:1313/1750 train_loss:3.4156 train_time:205702ms step_avg:157.87ms
step:1314/1750 train_loss:3.1057 train_time:205869ms step_avg:157.87ms
step:1315/1750 train_loss:3.3500 train_time:206031ms step_avg:157.88ms
step:1316/1750 train_loss:3.4609 train_time:206192ms step_avg:157.88ms
step:1317/1750 train_loss:3.4868 train_time:206356ms step_avg:157.89ms
step:1318/1750 train_loss:3.3620 train_time:206526ms step_avg:157.89ms
step:1319/1750 train_loss:3.4932 train_time:206690ms step_avg:157.90ms
step:1320/1750 train_loss:3.5239 train_time:206857ms step_avg:157.91ms
step:1321/1750 train_loss:3.4284 train_time:207022ms step_avg:157.91ms
step:1322/1750 train_loss:3.3849 train_time:207308ms step_avg:158.01ms
step:1323/1750 train_loss:3.3942 train_time:207481ms step_avg:158.02ms
step:1324/1750 train_loss:3.4974 train_time:207646ms step_avg:158.03ms
step:1325/1750 train_loss:3.5609 train_time:207817ms step_avg:158.04ms
step:1326/1750 train_loss:3.2886 train_time:207985ms step_avg:158.04ms
step:1327/1750 train_loss:3.2308 train_time:208147ms step_avg:158.05ms
step:1328/1750 train_loss:3.5561 train_time:208311ms step_avg:158.05ms
step:1329/1750 train_loss:3.3653 train_time:208620ms step_avg:158.17ms
step:1330/1750 train_loss:3.4959 train_time:208784ms step_avg:158.17ms
step:1331/1750 train_loss:3.4021 train_time:208945ms step_avg:158.17ms
step:1332/1750 train_loss:3.8079 train_time:209111ms step_avg:158.18ms
step:1333/1750 train_loss:3.5399 train_time:209277ms step_avg:158.18ms
step:1334/1750 train_loss:3.4345 train_time:209442ms step_avg:158.19ms
step:1335/1750 train_loss:3.3627 train_time:209606ms step_avg:158.19ms
step:1336/1750 train_loss:3.3582 train_time:209777ms step_avg:158.20ms
step:1337/1750 train_loss:3.6193 train_time:209943ms step_avg:158.21ms
step:1338/1750 train_loss:3.5874 train_time:210108ms step_avg:158.21ms
step:1339/1750 train_loss:3.4094 train_time:210275ms step_avg:158.22ms
step:1340/1750 train_loss:3.3501 train_time:210437ms step_avg:158.22ms
step:1341/1750 train_loss:3.6607 train_time:210600ms step_avg:158.23ms
step:1342/1750 train_loss:3.4269 train_time:210767ms step_avg:158.23ms
step:1343/1750 train_loss:3.4290 train_time:210929ms step_avg:158.24ms
step:1344/1750 train_loss:3.4836 train_time:211094ms step_avg:158.24ms
step:1345/1750 train_loss:3.4533 train_time:211261ms step_avg:158.25ms
step:1346/1750 train_loss:3.3658 train_time:211425ms step_avg:158.25ms
step:1347/1750 train_loss:3.3349 train_time:211589ms step_avg:158.26ms
step:1348/1750 train_loss:3.4097 train_time:211753ms step_avg:158.26ms
step:1349/1750 train_loss:3.3414 train_time:211913ms step_avg:158.26ms
step:1350/1750 train_loss:3.4558 train_time:212079ms step_avg:158.27ms
step:1351/1750 train_loss:3.3130 train_time:212240ms step_avg:158.27ms
step:1352/1750 train_loss:3.3677 train_time:212404ms step_avg:158.27ms
step:1353/1750 train_loss:3.4743 train_time:212572ms step_avg:158.28ms
step:1354/1750 train_loss:3.3269 train_time:212735ms step_avg:158.28ms
step:1355/1750 train_loss:3.2547 train_time:212895ms step_avg:158.29ms
step:1356/1750 train_loss:3.5776 train_time:213061ms step_avg:158.29ms
step:1357/1750 train_loss:3.4884 train_time:213226ms step_avg:158.30ms
step:1358/1750 train_loss:3.2424 train_time:213391ms step_avg:158.30ms
step:1359/1750 train_loss:3.5100 train_time:213555ms step_avg:158.31ms
step:1360/1750 train_loss:3.4227 train_time:213720ms step_avg:158.31ms
step:1361/1750 train_loss:3.2051 train_time:213889ms step_avg:158.32ms
step:1362/1750 train_loss:3.4570 train_time:214054ms step_avg:158.32ms
step:1363/1750 train_loss:3.3438 train_time:214223ms step_avg:158.33ms
step:1364/1750 train_loss:3.3726 train_time:214383ms step_avg:158.33ms
step:1365/1750 train_loss:3.3804 train_time:214543ms step_avg:158.33ms
step:1366/1750 train_loss:3.4909 train_time:214707ms step_avg:158.34ms
step:1367/1750 train_loss:3.4604 train_time:214871ms step_avg:158.34ms
step:1368/1750 train_loss:3.4197 train_time:215036ms step_avg:158.35ms
step:1369/1750 train_loss:3.3297 train_time:215207ms step_avg:158.36ms
step:1370/1750 train_loss:3.6657 train_time:215373ms step_avg:158.36ms
step:1371/1750 train_loss:3.3833 train_time:215536ms step_avg:158.37ms
step:1372/1750 train_loss:3.4312 train_time:215703ms step_avg:158.37ms
step:1373/1750 train_loss:3.4343 train_time:215866ms step_avg:158.38ms
step:1374/1750 train_loss:3.2226 train_time:216032ms step_avg:158.38ms
step:1375/1750 train_loss:3.6132 train_time:216195ms step_avg:158.38ms
step:1375/1750 val_loss:3.3780 train_time:216236ms step_avg:158.41ms
step:1376/1750 train_loss:3.4044 train_time:216358ms step_avg:158.39ms
step:1377/1750 train_loss:3.5486 train_time:216522ms step_avg:158.39ms
step:1378/1750 train_loss:3.5426 train_time:216684ms step_avg:158.39ms
step:1379/1750 train_loss:3.1909 train_time:216851ms step_avg:158.40ms
step:1380/1750 train_loss:3.3739 train_time:217014ms step_avg:158.40ms
step:1381/1750 train_loss:3.7762 train_time:217182ms step_avg:158.41ms
step:1382/1750 train_loss:3.2815 train_time:217344ms step_avg:158.41ms
step:1383/1750 train_loss:3.4610 train_time:217509ms step_avg:158.42ms
step:1384/1750 train_loss:3.5426 train_time:217674ms step_avg:158.42ms
step:1385/1750 train_loss:3.4703 train_time:217834ms step_avg:158.42ms
step:1386/1750 train_loss:3.4070 train_time:217997ms step_avg:158.43ms
step:1387/1750 train_loss:3.2622 train_time:218159ms step_avg:158.43ms
step:1388/1750 train_loss:3.4113 train_time:218321ms step_avg:158.43ms
step:1389/1750 train_loss:3.3828 train_time:218486ms step_avg:158.44ms
step:1390/1750 train_loss:3.6387 train_time:218647ms step_avg:158.44ms
step:1391/1750 train_loss:3.3588 train_time:218812ms step_avg:158.44ms
step:1392/1750 train_loss:3.3571 train_time:218975ms step_avg:158.45ms
step:1393/1750 train_loss:3.3139 train_time:219138ms step_avg:158.45ms
step:1394/1750 train_loss:3.5718 train_time:219300ms step_avg:158.45ms
step:1395/1750 train_loss:3.4668 train_time:219461ms step_avg:158.46ms
step:1396/1750 train_loss:3.4730 train_time:219622ms step_avg:158.46ms
step:1397/1750 train_loss:3.3689 train_time:219784ms step_avg:158.46ms
step:1398/1750 train_loss:3.3184 train_time:219947ms step_avg:158.46ms
step:1399/1750 train_loss:3.3907 train_time:220109ms step_avg:158.47ms
step:1400/1750 train_loss:3.3831 train_time:220276ms step_avg:158.47ms
step:1401/1750 train_loss:3.4080 train_time:220436ms step_avg:158.47ms
step:1402/1750 train_loss:3.3613 train_time:220600ms step_avg:158.48ms
step:1403/1750 train_loss:3.5603 train_time:220767ms step_avg:158.48ms
step:1404/1750 train_loss:3.3427 train_time:220930ms step_avg:158.49ms
step:1405/1750 train_loss:3.3772 train_time:221096ms step_avg:158.49ms
step:1406/1750 train_loss:3.3785 train_time:221261ms step_avg:158.50ms
step:1407/1750 train_loss:3.2390 train_time:221425ms step_avg:158.50ms
step:1408/1750 train_loss:3.3730 train_time:221588ms step_avg:158.50ms
step:1409/1750 train_loss:3.3595 train_time:221757ms step_avg:158.51ms
step:1410/1750 train_loss:3.3483 train_time:221918ms step_avg:158.51ms
step:1411/1750 train_loss:3.4311 train_time:222079ms step_avg:158.51ms
step:1412/1750 train_loss:3.3909 train_time:222243ms step_avg:158.52ms
step:1413/1750 train_loss:3.4221 train_time:222407ms step_avg:158.52ms
step:1414/1750 train_loss:3.3978 train_time:222572ms step_avg:158.53ms
step:1415/1750 train_loss:3.4811 train_time:222738ms step_avg:158.53ms
step:1416/1750 train_loss:3.2913 train_time:222906ms step_avg:158.54ms
step:1417/1750 train_loss:3.3489 train_time:223071ms step_avg:158.54ms
step:1418/1750 train_loss:3.4568 train_time:223234ms step_avg:158.55ms
step:1419/1750 train_loss:3.4078 train_time:223399ms step_avg:158.55ms
step:1420/1750 train_loss:3.4221 train_time:223565ms step_avg:158.56ms
step:1421/1750 train_loss:3.4423 train_time:223731ms step_avg:158.56ms
step:1422/1750 train_loss:3.4069 train_time:223895ms step_avg:158.57ms
step:1423/1750 train_loss:3.3824 train_time:224057ms step_avg:158.57ms
step:1424/1750 train_loss:3.3899 train_time:224223ms step_avg:158.57ms
step:1425/1750 train_loss:3.2449 train_time:224395ms step_avg:158.58ms
step:1426/1750 train_loss:3.3913 train_time:224556ms step_avg:158.58ms
step:1427/1750 train_loss:3.3405 train_time:224723ms step_avg:158.59ms
step:1428/1750 train_loss:3.4390 train_time:224888ms step_avg:158.60ms
step:1429/1750 train_loss:3.4185 train_time:225048ms step_avg:158.60ms
step:1430/1750 train_loss:3.3241 train_time:225215ms step_avg:158.60ms
step:1431/1750 train_loss:3.3816 train_time:225380ms step_avg:158.61ms
step:1432/1750 train_loss:3.4027 train_time:225546ms step_avg:158.61ms
step:1433/1750 train_loss:3.2097 train_time:225716ms step_avg:158.62ms
step:1434/1750 train_loss:3.3506 train_time:225884ms step_avg:158.63ms
step:1435/1750 train_loss:3.1773 train_time:226048ms step_avg:158.63ms
step:1436/1750 train_loss:3.2860 train_time:226214ms step_avg:158.63ms
step:1437/1750 train_loss:3.4759 train_time:226376ms step_avg:158.64ms
step:1438/1750 train_loss:3.4457 train_time:226537ms step_avg:158.64ms
step:1439/1750 train_loss:3.3771 train_time:226702ms step_avg:158.64ms
step:1440/1750 train_loss:3.2453 train_time:226866ms step_avg:158.65ms
step:1441/1750 train_loss:3.4029 train_time:227031ms step_avg:158.65ms
step:1442/1750 train_loss:3.4507 train_time:227200ms step_avg:158.66ms
step:1443/1750 train_loss:3.5409 train_time:227374ms step_avg:158.67ms
step:1444/1750 train_loss:3.5088 train_time:227537ms step_avg:158.67ms
step:1445/1750 train_loss:3.3965 train_time:227699ms step_avg:158.68ms
step:1446/1750 train_loss:3.2666 train_time:227867ms step_avg:158.68ms
step:1447/1750 train_loss:3.3589 train_time:228034ms step_avg:158.69ms
step:1448/1750 train_loss:3.3590 train_time:228199ms step_avg:158.69ms
step:1449/1750 train_loss:3.4574 train_time:228363ms step_avg:158.70ms
step:1450/1750 train_loss:3.4489 train_time:228528ms step_avg:158.70ms
step:1451/1750 train_loss:3.2706 train_time:228693ms step_avg:158.70ms
step:1452/1750 train_loss:3.3898 train_time:228858ms step_avg:158.71ms
step:1453/1750 train_loss:3.3226 train_time:229019ms step_avg:158.71ms
step:1454/1750 train_loss:3.3451 train_time:229184ms step_avg:158.71ms
step:1455/1750 train_loss:3.3874 train_time:229353ms step_avg:158.72ms
step:1456/1750 train_loss:3.3361 train_time:229518ms step_avg:158.73ms
step:1457/1750 train_loss:3.2165 train_time:229682ms step_avg:158.73ms
step:1458/1750 train_loss:3.4815 train_time:229847ms step_avg:158.73ms
step:1459/1750 train_loss:3.3351 train_time:230015ms step_avg:158.74ms
step:1460/1750 train_loss:3.3796 train_time:230179ms step_avg:158.74ms
step:1461/1750 train_loss:3.4956 train_time:230346ms step_avg:158.75ms
step:1462/1750 train_loss:3.3180 train_time:230512ms step_avg:158.75ms
step:1463/1750 train_loss:3.5246 train_time:230680ms step_avg:158.76ms
step:1464/1750 train_loss:3.4198 train_time:230845ms step_avg:158.77ms
step:1465/1750 train_loss:3.4152 train_time:231009ms step_avg:158.77ms
step:1466/1750 train_loss:3.3423 train_time:231173ms step_avg:158.77ms
step:1467/1750 train_loss:3.4568 train_time:231338ms step_avg:158.78ms
step:1468/1750 train_loss:3.3439 train_time:231501ms step_avg:158.78ms
step:1469/1750 train_loss:3.3241 train_time:231666ms step_avg:158.78ms
step:1470/1750 train_loss:3.3898 train_time:231836ms step_avg:158.79ms
step:1471/1750 train_loss:3.3081 train_time:232008ms step_avg:158.80ms
step:1472/1750 train_loss:3.2972 train_time:232177ms step_avg:158.81ms
step:1473/1750 train_loss:3.4930 train_time:232339ms step_avg:158.81ms
step:1474/1750 train_loss:3.3704 train_time:232509ms step_avg:158.82ms
step:1475/1750 train_loss:3.2023 train_time:232678ms step_avg:158.82ms
step:1476/1750 train_loss:3.3224 train_time:232841ms step_avg:158.83ms
step:1477/1750 train_loss:3.3006 train_time:233014ms step_avg:158.84ms
step:1478/1750 train_loss:3.3662 train_time:233182ms step_avg:158.84ms
step:1479/1750 train_loss:3.4563 train_time:233349ms step_avg:158.85ms
step:1480/1750 train_loss:3.3354 train_time:233514ms step_avg:158.85ms
step:1481/1750 train_loss:3.5105 train_time:233681ms step_avg:158.86ms
step:1482/1750 train_loss:3.4281 train_time:233855ms step_avg:158.87ms
step:1483/1750 train_loss:3.3347 train_time:234028ms step_avg:158.88ms
step:1484/1750 train_loss:3.3177 train_time:234198ms step_avg:158.89ms
step:1485/1750 train_loss:3.3369 train_time:234363ms step_avg:158.89ms
step:1486/1750 train_loss:3.2824 train_time:234533ms step_avg:158.90ms
step:1487/1750 train_loss:3.3945 train_time:234699ms step_avg:158.90ms
step:1488/1750 train_loss:3.2885 train_time:234866ms step_avg:158.91ms
step:1489/1750 train_loss:3.3773 train_time:235030ms step_avg:158.91ms
step:1490/1750 train_loss:3.3020 train_time:235194ms step_avg:158.92ms
step:1491/1750 train_loss:3.2164 train_time:235359ms step_avg:158.92ms
step:1492/1750 train_loss:3.3207 train_time:235523ms step_avg:158.92ms
step:1493/1750 train_loss:3.4902 train_time:235686ms step_avg:158.92ms
step:1494/1750 train_loss:3.3527 train_time:235849ms step_avg:158.93ms
step:1495/1750 train_loss:3.0866 train_time:236018ms step_avg:158.93ms
step:1496/1750 train_loss:3.4084 train_time:236183ms step_avg:158.94ms
step:1497/1750 train_loss:3.3678 train_time:236349ms step_avg:158.94ms
step:1498/1750 train_loss:3.4010 train_time:236517ms step_avg:158.95ms
step:1499/1750 train_loss:3.3686 train_time:236686ms step_avg:158.96ms
step:1500/1750 train_loss:3.3505 train_time:236860ms step_avg:158.97ms
step:1500/1750 val_loss:3.3406 train_time:236904ms step_avg:159.00ms
step:1501/1750 train_loss:3.1455 train_time:237029ms step_avg:158.97ms
step:1502/1750 train_loss:3.4144 train_time:237204ms step_avg:158.98ms
step:1503/1750 train_loss:3.2982 train_time:237367ms step_avg:158.99ms
step:1504/1750 train_loss:3.3027 train_time:237533ms step_avg:158.99ms
step:1505/1750 train_loss:3.2624 train_time:237699ms step_avg:159.00ms
step:1506/1750 train_loss:3.3332 train_time:237865ms step_avg:159.00ms
step:1507/1750 train_loss:3.2265 train_time:238040ms step_avg:159.01ms
step:1508/1750 train_loss:3.5330 train_time:238205ms step_avg:159.02ms
step:1509/1750 train_loss:3.3320 train_time:238368ms step_avg:159.02ms
step:1510/1750 train_loss:3.3271 train_time:238534ms step_avg:159.02ms
step:1511/1750 train_loss:3.4663 train_time:238815ms step_avg:159.10ms
step:1512/1750 train_loss:3.4737 train_time:238984ms step_avg:159.11ms
step:1513/1750 train_loss:3.3212 train_time:239152ms step_avg:159.12ms
step:1514/1750 train_loss:3.1431 train_time:239318ms step_avg:159.12ms
step:1515/1750 train_loss:3.2893 train_time:239483ms step_avg:159.12ms
step:1516/1750 train_loss:3.3030 train_time:239652ms step_avg:159.13ms
step:1517/1750 train_loss:3.3557 train_time:239816ms step_avg:159.13ms
step:1518/1750 train_loss:3.2561 train_time:239982ms step_avg:159.14ms
step:1519/1750 train_loss:3.5519 train_time:240291ms step_avg:159.24ms
step:1520/1750 train_loss:3.1843 train_time:240458ms step_avg:159.24ms
step:1521/1750 train_loss:3.2633 train_time:240619ms step_avg:159.24ms
step:1522/1750 train_loss:3.4034 train_time:240785ms step_avg:159.25ms
step:1523/1750 train_loss:3.2776 train_time:240947ms step_avg:159.25ms
step:1524/1750 train_loss:3.3949 train_time:241111ms step_avg:159.25ms
step:1525/1750 train_loss:3.3821 train_time:241280ms step_avg:159.26ms
step:1526/1750 train_loss:3.3271 train_time:241452ms step_avg:159.27ms
step:1527/1750 train_loss:3.3354 train_time:241617ms step_avg:159.27ms
step:1528/1750 train_loss:3.4564 train_time:241781ms step_avg:159.28ms
step:1529/1750 train_loss:3.4563 train_time:241942ms step_avg:159.28ms
step:1530/1750 train_loss:3.2863 train_time:242103ms step_avg:159.28ms
step:1531/1750 train_loss:3.2401 train_time:242269ms step_avg:159.28ms
step:1532/1750 train_loss:3.3974 train_time:242435ms step_avg:159.29ms
step:1533/1750 train_loss:3.3252 train_time:242604ms step_avg:159.29ms
step:1534/1750 train_loss:3.3258 train_time:242774ms step_avg:159.30ms
step:1535/1750 train_loss:3.3330 train_time:242941ms step_avg:159.31ms
step:1536/1750 train_loss:3.2764 train_time:243107ms step_avg:159.31ms
step:1537/1750 train_loss:3.3222 train_time:243271ms step_avg:159.31ms
step:1538/1750 train_loss:3.4772 train_time:243441ms step_avg:159.32ms
step:1539/1750 train_loss:3.4436 train_time:243611ms step_avg:159.33ms
step:1540/1750 train_loss:3.3238 train_time:243776ms step_avg:159.33ms
step:1541/1750 train_loss:3.2816 train_time:243939ms step_avg:159.33ms
step:1542/1750 train_loss:3.2980 train_time:244105ms step_avg:159.34ms
step:1543/1750 train_loss:3.1980 train_time:244272ms step_avg:159.34ms
step:1544/1750 train_loss:3.3415 train_time:244434ms step_avg:159.34ms
step:1545/1750 train_loss:3.3094 train_time:244600ms step_avg:159.35ms
step:1546/1750 train_loss:3.3065 train_time:244772ms step_avg:159.36ms
step:1547/1750 train_loss:3.2682 train_time:244939ms step_avg:159.36ms
step:1548/1750 train_loss:3.3094 train_time:245108ms step_avg:159.37ms
step:1549/1750 train_loss:3.3816 train_time:245272ms step_avg:159.37ms
step:1550/1750 train_loss:3.3392 train_time:245435ms step_avg:159.37ms
step:1551/1750 train_loss:3.2515 train_time:245602ms step_avg:159.38ms
step:1552/1750 train_loss:3.2688 train_time:245770ms step_avg:159.38ms
step:1553/1750 train_loss:3.2724 train_time:245933ms step_avg:159.39ms
step:1554/1750 train_loss:3.4059 train_time:246098ms step_avg:159.39ms
step:1555/1750 train_loss:3.3842 train_time:246263ms step_avg:159.39ms
step:1556/1750 train_loss:3.3260 train_time:246425ms step_avg:159.40ms
step:1557/1750 train_loss:3.3720 train_time:246588ms step_avg:159.40ms
step:1558/1750 train_loss:3.3065 train_time:246755ms step_avg:159.40ms
step:1559/1750 train_loss:3.1782 train_time:246928ms step_avg:159.41ms
step:1560/1750 train_loss:3.4812 train_time:247091ms step_avg:159.41ms
step:1561/1750 train_loss:3.2738 train_time:247257ms step_avg:159.42ms
step:1562/1750 train_loss:3.2614 train_time:247422ms step_avg:159.42ms
step:1563/1750 train_loss:3.3714 train_time:247588ms step_avg:159.43ms
step:1564/1750 train_loss:3.2000 train_time:247759ms step_avg:159.43ms
step:1565/1750 train_loss:3.2119 train_time:247926ms step_avg:159.44ms
step:1566/1750 train_loss:3.4102 train_time:248092ms step_avg:159.44ms
step:1567/1750 train_loss:3.2822 train_time:248256ms step_avg:159.45ms
step:1568/1750 train_loss:3.2875 train_time:248426ms step_avg:159.45ms
step:1569/1750 train_loss:3.3706 train_time:248601ms step_avg:159.46ms
step:1570/1750 train_loss:3.3326 train_time:248769ms step_avg:159.47ms
step:1571/1750 train_loss:3.2069 train_time:248938ms step_avg:159.47ms
step:1572/1750 train_loss:3.2436 train_time:249103ms step_avg:159.48ms
step:1573/1750 train_loss:3.3607 train_time:249271ms step_avg:159.48ms
step:1574/1750 train_loss:3.2146 train_time:249434ms step_avg:159.48ms
step:1575/1750 train_loss:3.3737 train_time:249599ms step_avg:159.49ms
step:1576/1750 train_loss:3.2755 train_time:249763ms step_avg:159.49ms
step:1577/1750 train_loss:3.3331 train_time:249932ms step_avg:159.50ms
step:1578/1750 train_loss:3.3130 train_time:250098ms step_avg:159.50ms
step:1579/1750 train_loss:3.2855 train_time:250266ms step_avg:159.51ms
step:1580/1750 train_loss:3.2485 train_time:250434ms step_avg:159.51ms
step:1581/1750 train_loss:3.4446 train_time:250605ms step_avg:159.52ms
step:1582/1750 train_loss:3.2700 train_time:250781ms step_avg:159.53ms
step:1583/1750 train_loss:3.4246 train_time:250952ms step_avg:159.54ms
step:1584/1750 train_loss:3.2414 train_time:251116ms step_avg:159.54ms
step:1585/1750 train_loss:3.4173 train_time:251286ms step_avg:159.55ms
step:1586/1750 train_loss:3.1956 train_time:251453ms step_avg:159.55ms
step:1587/1750 train_loss:3.3997 train_time:251618ms step_avg:159.55ms
step:1588/1750 train_loss:3.2693 train_time:251784ms step_avg:159.56ms
step:1589/1750 train_loss:3.4407 train_time:251949ms step_avg:159.56ms
step:1590/1750 train_loss:3.2836 train_time:252118ms step_avg:159.57ms
step:1591/1750 train_loss:3.2921 train_time:252282ms step_avg:159.57ms
step:1592/1750 train_loss:3.3661 train_time:252448ms step_avg:159.57ms
step:1593/1750 train_loss:3.3359 train_time:252620ms step_avg:159.58ms
step:1594/1750 train_loss:3.3074 train_time:252785ms step_avg:159.59ms
step:1595/1750 train_loss:3.4538 train_time:252954ms step_avg:159.59ms
step:1596/1750 train_loss:3.1588 train_time:253128ms step_avg:159.60ms
step:1597/1750 train_loss:3.3340 train_time:253299ms step_avg:159.61ms
step:1598/1750 train_loss:3.3822 train_time:253468ms step_avg:159.61ms
step:1599/1750 train_loss:3.4476 train_time:253641ms step_avg:159.62ms
step:1600/1750 train_loss:3.2767 train_time:253809ms step_avg:159.63ms
step:1601/1750 train_loss:3.5857 train_time:253975ms step_avg:159.63ms
step:1602/1750 train_loss:3.4518 train_time:254144ms step_avg:159.64ms
step:1603/1750 train_loss:3.2234 train_time:254318ms step_avg:159.65ms
step:1604/1750 train_loss:3.2726 train_time:254486ms step_avg:159.65ms
step:1605/1750 train_loss:3.1575 train_time:254659ms step_avg:159.66ms
step:1606/1750 train_loss:3.4703 train_time:254832ms step_avg:159.67ms
step:1607/1750 train_loss:3.3024 train_time:254998ms step_avg:159.67ms
step:1608/1750 train_loss:3.3028 train_time:255168ms step_avg:159.68ms
step:1609/1750 train_loss:3.2433 train_time:255339ms step_avg:159.69ms
step:1610/1750 train_loss:3.7505 train_time:255517ms step_avg:159.70ms
step:1611/1750 train_loss:3.5043 train_time:255687ms step_avg:159.70ms
step:1612/1750 train_loss:3.3974 train_time:255861ms step_avg:159.71ms
step:1613/1750 train_loss:3.2671 train_time:256037ms step_avg:159.72ms
step:1614/1750 train_loss:3.2926 train_time:256204ms step_avg:159.73ms
step:1615/1750 train_loss:3.3150 train_time:256374ms step_avg:159.73ms
step:1616/1750 train_loss:3.2848 train_time:256553ms step_avg:159.75ms
step:1617/1750 train_loss:3.3592 train_time:256726ms step_avg:159.76ms
step:1618/1750 train_loss:3.2879 train_time:256889ms step_avg:159.76ms
step:1619/1750 train_loss:3.1871 train_time:257056ms step_avg:159.76ms
step:1620/1750 train_loss:3.4576 train_time:257221ms step_avg:159.76ms
step:1621/1750 train_loss:3.3837 train_time:257391ms step_avg:159.77ms
step:1622/1750 train_loss:3.1578 train_time:257558ms step_avg:159.78ms
step:1623/1750 train_loss:3.2565 train_time:257727ms step_avg:159.78ms
step:1624/1750 train_loss:3.2168 train_time:257891ms step_avg:159.78ms
step:1625/1750 train_loss:3.3214 train_time:258057ms step_avg:159.79ms
step:1625/1750 val_loss:3.3023 train_time:258099ms step_avg:159.81ms
step:1626/1750 train_loss:3.2403 train_time:258221ms step_avg:159.79ms
step:1627/1750 train_loss:3.2409 train_time:258383ms step_avg:159.79ms
step:1628/1750 train_loss:3.3671 train_time:258548ms step_avg:159.79ms
step:1629/1750 train_loss:3.2505 train_time:258714ms step_avg:159.80ms
step:1630/1750 train_loss:3.3250 train_time:258883ms step_avg:159.80ms
step:1631/1750 train_loss:3.1776 train_time:259061ms step_avg:159.82ms
step:1632/1750 train_loss:3.1479 train_time:259227ms step_avg:159.82ms
step:1633/1750 train_loss:3.2984 train_time:259394ms step_avg:159.82ms
step:1634/1750 train_loss:3.3123 train_time:259559ms step_avg:159.83ms
step:1635/1750 train_loss:3.2506 train_time:259733ms step_avg:159.84ms
step:1636/1750 train_loss:3.3304 train_time:259899ms step_avg:159.84ms
step:1637/1750 train_loss:3.3801 train_time:260067ms step_avg:159.84ms
step:1638/1750 train_loss:3.4031 train_time:260236ms step_avg:159.85ms
step:1639/1750 train_loss:3.5734 train_time:260408ms step_avg:159.86ms
step:1640/1750 train_loss:3.3470 train_time:260578ms step_avg:159.86ms
step:1641/1750 train_loss:3.2978 train_time:260748ms step_avg:159.87ms
step:1642/1750 train_loss:3.4093 train_time:260915ms step_avg:159.87ms
step:1643/1750 train_loss:3.2777 train_time:261088ms step_avg:159.88ms
step:1644/1750 train_loss:3.3163 train_time:261253ms step_avg:159.89ms
step:1645/1750 train_loss:3.3195 train_time:261415ms step_avg:159.89ms
step:1646/1750 train_loss:3.0705 train_time:261582ms step_avg:159.89ms
step:1647/1750 train_loss:3.3252 train_time:261751ms step_avg:159.90ms
step:1648/1750 train_loss:3.2169 train_time:261916ms step_avg:159.90ms
step:1649/1750 train_loss:3.2889 train_time:262079ms step_avg:159.90ms
step:1650/1750 train_loss:3.2682 train_time:262245ms step_avg:159.91ms
step:1651/1750 train_loss:3.3416 train_time:262412ms step_avg:159.91ms
step:1652/1750 train_loss:3.2601 train_time:262579ms step_avg:159.91ms
step:1653/1750 train_loss:3.3949 train_time:262750ms step_avg:159.92ms
step:1654/1750 train_loss:3.3848 train_time:262913ms step_avg:159.92ms
step:1655/1750 train_loss:3.1781 train_time:263085ms step_avg:159.93ms
step:1656/1750 train_loss:3.3374 train_time:263259ms step_avg:159.94ms
step:1657/1750 train_loss:3.2510 train_time:263427ms step_avg:159.94ms
step:1658/1750 train_loss:3.2251 train_time:263590ms step_avg:159.95ms
step:1659/1750 train_loss:3.3060 train_time:263755ms step_avg:159.95ms
step:1660/1750 train_loss:3.3487 train_time:263921ms step_avg:159.95ms
step:1661/1750 train_loss:3.2507 train_time:264088ms step_avg:159.96ms
step:1662/1750 train_loss:3.3605 train_time:264254ms step_avg:159.96ms
step:1663/1750 train_loss:3.3512 train_time:264423ms step_avg:159.97ms
step:1664/1750 train_loss:3.4085 train_time:264601ms step_avg:159.98ms
step:1665/1750 train_loss:3.3289 train_time:264771ms step_avg:159.98ms
step:1666/1750 train_loss:3.5021 train_time:264933ms step_avg:159.98ms
step:1667/1750 train_loss:3.2077 train_time:265100ms step_avg:159.99ms
step:1668/1750 train_loss:3.2916 train_time:265270ms step_avg:159.99ms
step:1669/1750 train_loss:3.2087 train_time:265436ms step_avg:160.00ms
step:1670/1750 train_loss:3.2208 train_time:265603ms step_avg:160.00ms
step:1671/1750 train_loss:3.3718 train_time:265770ms step_avg:160.01ms
step:1672/1750 train_loss:3.5776 train_time:265936ms step_avg:160.01ms
step:1673/1750 train_loss:3.2772 train_time:266105ms step_avg:160.02ms
step:1674/1750 train_loss:3.2583 train_time:266272ms step_avg:160.02ms
step:1675/1750 train_loss:3.1320 train_time:266442ms step_avg:160.02ms
step:1676/1750 train_loss:3.3518 train_time:266612ms step_avg:160.03ms
step:1677/1750 train_loss:3.2785 train_time:266780ms step_avg:160.04ms
step:1678/1750 train_loss:3.2943 train_time:266950ms step_avg:160.04ms
step:1679/1750 train_loss:3.2999 train_time:267115ms step_avg:160.04ms
step:1680/1750 train_loss:3.0918 train_time:267289ms step_avg:160.05ms
step:1681/1750 train_loss:3.3018 train_time:267457ms step_avg:160.06ms
step:1682/1750 train_loss:3.2938 train_time:267625ms step_avg:160.06ms
step:1683/1750 train_loss:3.3061 train_time:267792ms step_avg:160.07ms
step:1684/1750 train_loss:3.3411 train_time:267955ms step_avg:160.07ms
step:1685/1750 train_loss:3.2445 train_time:268120ms step_avg:160.07ms
step:1686/1750 train_loss:3.3670 train_time:268289ms step_avg:160.08ms
step:1687/1750 train_loss:3.2458 train_time:268455ms step_avg:160.08ms
step:1688/1750 train_loss:3.3120 train_time:268629ms step_avg:160.09ms
step:1689/1750 train_loss:3.2238 train_time:268798ms step_avg:160.09ms
step:1690/1750 train_loss:3.0740 train_time:268972ms step_avg:160.10ms
step:1691/1750 train_loss:3.3078 train_time:269137ms step_avg:160.11ms
step:1692/1750 train_loss:3.2947 train_time:269302ms step_avg:160.11ms
step:1693/1750 train_loss:3.2155 train_time:269467ms step_avg:160.11ms
step:1694/1750 train_loss:3.6113 train_time:269641ms step_avg:160.12ms
step:1695/1750 train_loss:3.3335 train_time:269813ms step_avg:160.13ms
step:1696/1750 train_loss:3.3380 train_time:269979ms step_avg:160.13ms
step:1697/1750 train_loss:3.2560 train_time:270145ms step_avg:160.13ms
step:1698/1750 train_loss:3.1232 train_time:270313ms step_avg:160.14ms
step:1699/1750 train_loss:3.2299 train_time:270480ms step_avg:160.14ms
step:1700/1750 train_loss:3.2488 train_time:270770ms step_avg:160.22ms
step:1701/1750 train_loss:3.3240 train_time:270941ms step_avg:160.23ms
step:1702/1750 train_loss:3.2371 train_time:271106ms step_avg:160.23ms
step:1703/1750 train_loss:3.4167 train_time:271270ms step_avg:160.23ms
step:1704/1750 train_loss:3.2125 train_time:271436ms step_avg:160.23ms
step:1705/1750 train_loss:3.4359 train_time:271601ms step_avg:160.24ms
step:1706/1750 train_loss:3.2525 train_time:271764ms step_avg:160.24ms
step:1707/1750 train_loss:3.0552 train_time:271934ms step_avg:160.24ms
step:1708/1750 train_loss:3.3968 train_time:272099ms step_avg:160.25ms
step:1709/1750 train_loss:3.2974 train_time:272403ms step_avg:160.33ms
step:1710/1750 train_loss:3.2777 train_time:272577ms step_avg:160.34ms
step:1711/1750 train_loss:3.2898 train_time:272743ms step_avg:160.34ms
step:1712/1750 train_loss:3.3211 train_time:272912ms step_avg:160.35ms
step:1713/1750 train_loss:3.3387 train_time:273080ms step_avg:160.35ms
step:1714/1750 train_loss:3.2268 train_time:273251ms step_avg:160.36ms
step:1715/1750 train_loss:3.2876 train_time:273427ms step_avg:160.37ms
step:1716/1750 train_loss:3.1004 train_time:273592ms step_avg:160.37ms
step:1717/1750 train_loss:3.2507 train_time:273755ms step_avg:160.37ms
step:1718/1750 train_loss:3.2620 train_time:273923ms step_avg:160.38ms
step:1719/1750 train_loss:3.2207 train_time:274092ms step_avg:160.38ms
step:1720/1750 train_loss:3.3779 train_time:274266ms step_avg:160.39ms
step:1721/1750 train_loss:3.1716 train_time:274448ms step_avg:160.40ms
step:1722/1750 train_loss:3.3156 train_time:274614ms step_avg:160.41ms
step:1723/1750 train_loss:3.4097 train_time:274790ms step_avg:160.41ms
step:1724/1750 train_loss:3.2613 train_time:274957ms step_avg:160.42ms
step:1725/1750 train_loss:3.4887 train_time:275132ms step_avg:160.43ms
step:1726/1750 train_loss:3.2587 train_time:275306ms step_avg:160.43ms
step:1727/1750 train_loss:3.3311 train_time:275471ms step_avg:160.44ms
step:1728/1750 train_loss:3.3012 train_time:275638ms step_avg:160.44ms
step:1729/1750 train_loss:3.2811 train_time:275811ms step_avg:160.45ms
step:1730/1750 train_loss:3.6543 train_time:275981ms step_avg:160.45ms
step:1731/1750 train_loss:3.2949 train_time:276147ms step_avg:160.46ms
step:1732/1750 train_loss:3.4332 train_time:276314ms step_avg:160.46ms
step:1733/1750 train_loss:3.2106 train_time:276476ms step_avg:160.46ms
step:1734/1750 train_loss:3.2463 train_time:276644ms step_avg:160.47ms
step:1735/1750 train_loss:3.2760 train_time:276813ms step_avg:160.47ms
step:1736/1750 train_loss:3.2569 train_time:276983ms step_avg:160.48ms
step:1737/1750 train_loss:3.3858 train_time:277155ms step_avg:160.48ms
step:1738/1750 train_loss:3.2191 train_time:277333ms step_avg:160.49ms
step:1739/1750 train_loss:3.2866 train_time:277506ms step_avg:160.50ms
step:1740/1750 train_loss:3.3725 train_time:277677ms step_avg:160.51ms
step:1741/1750 train_loss:3.1619 train_time:277842ms step_avg:160.51ms
step:1742/1750 train_loss:3.0627 train_time:278012ms step_avg:160.51ms
step:1743/1750 train_loss:2.9626 train_time:278187ms step_avg:160.52ms
step:1744/1750 train_loss:3.2944 train_time:278352ms step_avg:160.53ms
step:1745/1750 train_loss:3.3169 train_time:278515ms step_avg:160.53ms
step:1746/1750 train_loss:3.2721 train_time:278679ms step_avg:160.53ms
step:1747/1750 train_loss:3.2966 train_time:278852ms step_avg:160.54ms
step:1748/1750 train_loss:3.5020 train_time:279032ms step_avg:160.55ms
step:1749/1750 train_loss:3.2266 train_time:279199ms step_avg:160.55ms
step:1750/1750 train_loss:3.2819 train_time:279371ms step_avg:160.56ms
step:1750/1750 val_loss:3.2810 train_time:279420ms step_avg:160.59ms