fixed some style, removed broken links in example/huggingface/README.md

examples/huggingface/README.md

@@ -17,7 +17,7 @@ Use the included requirements.txt file to include all extra dependencies:
 - [`HFGlobalPowerLimitOptimizer`](https://ml.energy/zeus/reference/optimizer/power_limit/#zeus.optimizer.power_limit.HFGlobalPowerLimitOptimizer): Online-profiles each power limit with `ZeusMonitor` and finds the cost-optimal power limit. Calls GlobalPowerLimitOptimizer under the hood.
 
 ## Integration with HuggingFace 🤗 Trainer
-For easy use with [HuggingFace 🤗 Transformers](https://huggingface.co/docs/transformers/en/index), [`HFGlobalPowerLimitOptimizer`](zeus.optimizer.power_limit.HFGlobalPowerLimitOptimizer) is a drop-in compatible [HuggingFace 🤗 Trainer Callback](https://huggingface.co/docs/transformers/en/main_classes/callback). When initializing a [HuggingFace 🤗 Trainer](https://huggingface.co/docs/transformers/main_classes/trainer), initialize and pass in [`HFGlobalPowerLimitOptimizer`](zeus.optimizer.power_limit.HFGlobalPowerLimitOptimizer) as shown below:
+For easy use with [HuggingFace 🤗 Transformers](https://huggingface.co/docs/transformers/en/index), `HFGlobalPowerLimitOptimizer` is a drop-in compatible [HuggingFace 🤗 Trainer Callback](https://huggingface.co/docs/transformers/en/main_classes/callback). When initializing a [HuggingFace 🤗 Trainer](https://huggingface.co/docs/transformers/main_classes/trainer), initialize and pass in `HFGlobalPowerLimitOptimizer` as shown below:
 
 ```python
     monitor = ZeusMonitor()
@@ -44,3 +44,4 @@ python run_clm.py \
     --do_train \
     --do_eval \
     --output_dir /tmp/test-clm
+```

examples/huggingface/run_clm.py

@@ -34,7 +34,6 @@
 import evaluate
 import torch
 from datasets import load_dataset
-
 import transformers
 from transformers import (
     CONFIG_MAPPING,
@@ -290,20 +289,6 @@ def __post_init__(self):
                 ], "`validation_file` should be a csv, a json or a txt file."
 
 
-@dataclass
-class ZeusTrainingArguments:
-    """
-    Arguments pertaining to the zeus training.
-    """
-
-    gpu_indices: Optional[str] = field(
-        default=None,
-        metadata={
-            "help": 'The indices of the gpus to use for training (ex. "0,1,2"). If none, all gpus will be used.'
-        },
-    )
-
-
 def main():
     # See all possible arguments in src/transformers/training_args.py
     # or by passing the --help flag to this script.
@@ -314,21 +299,19 @@ def main():
             ModelArguments,
             DataTrainingArguments,
             TrainingArguments,
-            ZeusTrainingArguments,
         )
     )
     if len(sys.argv) == 2 and sys.argv[1].endswith(".json"):
         # If we pass only one argument to the script and it's the path to a json file,
         # let's parse it to get our arguments.
-        model_args, data_args, training_args, zeus_args = parser.parse_json_file(
+        model_args, data_args, training_args = parser.parse_json_file(
             json_file=os.path.abspath(sys.argv[1])
         )
     else:
         (
             model_args,
             data_args,
             training_args,
-            zeus_args,
         ) = parser.parse_args_into_dataclasses()
 
     if model_args.use_auth_token is not None:
@@ -689,13 +672,8 @@ def compute_metrics(eval_preds):
             preds = preds[:, :-1].reshape(-1)
             return metric.compute(predictions=preds, references=labels)
 
-    # See if gpu_indices is set, and pass it to the monitor
-    gpu_indices = None
-    if zeus_args.gpu_indices is not None:
-        gpu_indices = [int(i) for i in zeus_args.gpu_indices.split(",")]
-
     # Initialize our ZeusMonitor and HFGlobalPowerLimitOptimizer
-    monitor = ZeusMonitor(gpu_indices=gpu_indices)
+    monitor = ZeusMonitor()
     optimizer = HFGlobalPowerLimitOptimizer(monitor)
 
     # Initialize our Trainer

tests/optimizer/test_power_limit_optimizer.py

@@ -14,6 +14,7 @@
 
 from __future__ import annotations
 
+import inspect
 import os
 import typing
 from typing import Generator, Iterable
@@ -38,8 +39,6 @@
     from pathlib import Path
     from pytest_mock import MockerFixture
 
-import inspect
-
 
 PROFILE_DATA_DIR = "tests/profile_data/"
 
@@ -279,4 +278,4 @@ def test_HFGPLO_signature_equivalence() -> None:
     gplo_signature = inspect.signature(GlobalPowerLimitOptimizer.__init__)
     hfgplo_signature = inspect.signature(HFGlobalPowerLimitOptimizer.__init__)
 
-    assert gplo_signature == hfgplo_signature, "GlobalPowerLimitOptimizer and HFGlobalPowerLimitOptimizer signatures do not match."
+    assert gplo_signature == hfgplo_signature, "GlobalPowerLimitOptimizer and HFGlobalPowerLimitOptimizer signatures do not match."