-
Notifications
You must be signed in to change notification settings - Fork 36
/
wandb_client.py
78 lines (65 loc) · 2.15 KB
/
wandb_client.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
import netrc
from pathlib import Path
from typing import Any, Sequence
from contextlib import suppress
import wandb
from wandb.sdk.wandb_settings import Settings
def logout_wandb():
netrc_path = Path("/root/.netrc")
if not netrc_path.exists():
return
n = netrc.netrc(netrc_path)
if "api.wandb.ai" in n.hosts:
del n.hosts["api.wandb.ai"]
netrc_path.write_text(repr(n))
class WeightsAndBiasesClient:
def __init__(
self,
api_key: str,
project: str,
config: dict,
sample_prompts: list[str],
entity: str | None,
name: str | None,
):
self.api_key = api_key
self.sample_prompts = sample_prompts
wandb.login(key=self.api_key, verify=True)
try:
self.run = wandb.init(
project=project,
entity=entity,
name=name,
config=config,
save_code=False,
settings=Settings(_disable_machine_info=True),
)
except Exception as e:
raise ValueError(f"Failed to log in to Weights & Biases: {e}")
def log_loss(self, loss_dict: dict[str, Any], step: int | None):
try:
wandb.log(data=loss_dict, step=step)
except Exception as e:
print(f"Failed to log to Weights & Biases: {e}")
def log_samples(self, image_paths: Sequence[Path], step: int | None):
data = {
f"samples/{truncate(prompt)}": wandb.Image(str(path))
for prompt, path in zip(self.sample_prompts, image_paths)
}
try:
wandb.log(data=data, step=step)
except Exception as e:
print(f"Failed to log to Weights & Biases: {e}")
def save_weights(self, lora_path: Path):
try:
wandb.save(lora_path)
except Exception as e:
print(f"Failed to save to Weights & Biases: {e}")
def finish(self):
with suppress(Exception):
wandb.finish()
def truncate(text, max_chars=50):
if len(text) <= max_chars:
return text
half = (max_chars - 3) // 2
return f"{text[:half]}...{text[-half:]}"