Integration with Hugging Face , the demo should add report_to="none" to TrainingArguments

[dshwei]

🐛 Bug

  File "/demo/project/pre_train_ft_7b.py", line 200, in run_training
    trainer = Trainer(
  File "/demo/miniconda3/envs/sqlcode/lib/python3.9/site-packages/transformers/trainer.py", line 503, in __init__
    self.callback_handler = CallbackHandler(
  File "/demo/miniconda3/envs/sqlcode/lib/python3.9/site-packages/transformers/trainer_callback.py", line 313, in __init__
    self.add_callback(cb)
  File "/demo/miniconda3/envs/sqlcode/lib/python3.9/site-packages/transformers/trainer_callback.py", line 330, in add_callback
    cb = callback() if isinstance(callback, type) else callback
  File "/demo/miniconda3/envs/sqlcode/lib/python3.9/site-packages/transformers/integrations/integration_utils.py", line 597, in __init__
    from torch.utils.tensorboard import SummaryWriter  # noqa: F401
  File "/demo/miniconda3/envs/sqlcode/lib/python3.9/site-packages/torch/utils/tensorboard/__init__.py", line 12, in <module>
    from .writer import FileWriter, SummaryWriter  # noqa: F401
  File "/demo/miniconda3/envs/sqlcode/lib/python3.9/site-packages/torch/utils/tensorboard/writer.py", line 10, in <module>
    from tensorboard.compat.proto import event_pb2
  File "/demo/miniconda3/envs/sqlcode/lib/python3.9/site-packages/tensorboard/compat/proto/event_pb2.py", line 17, in <module>
    from tensorboard.compat.proto import summary_pb2 as tensorboard_dot_compat_dot_proto_dot_summary__pb2
  File "/demo/miniconda3/envs/sqlcode/lib/python3.9/site-packages/tensorboard/compat/proto/summary_pb2.py", line 17, in <module>
    from tensorboard.compat.proto import tensor_pb2 as tensorboard_dot_compat_dot_proto_dot_tensor__pb2
  File "/demo/miniconda3/envs/sqlcode/lib/python3.9/site-packages/tensorboard/compat/proto/tensor_pb2.py", line 16, in <module>
    from tensorboard.compat.proto import resource_handle_pb2 as tensorboard_dot_compat_dot_proto_dot_resource__handle__pb2
  File "/demo/miniconda3/envs/sqlcode/lib/python3.9/site-packages/tensorboard/compat/proto/resource_handle_pb2.py", line 16, in <module>
    from tensorboard.compat.proto import tensor_shape_pb2 as tensorboard_dot_compat_dot_proto_dot_tensor__shape__pb2
  File "/demo/miniconda3/envs/sqlcode/lib/python3.9/site-packages/tensorboard/compat/proto/tensor_shape_pb2.py", line 36, in <module>
    _descriptor.FieldDescriptor(
  File "/demo/miniconda3/envs/sqlcode/lib/python3.9/site-packages/google/protobuf/descriptor.py", line 553, in __new__
    _message.Message._CheckCalledFromGeneratedFile()
TypeError: Descriptors cannot be created directly.
If this call came from a _pb2.py file, your generated code is out of date and must be regenerated with protoc >= 3.19.0.
If you cannot immediately regenerate your protos, some other possible workarounds are:
 1. Downgrade the protobuf package to 3.20.x or lower.
 2. Set PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION=python (but this will use pure-Python parsing and will be much slower).

To reproduce

Expected behavior

training_args = TrainingArguments(output_dir="test_trainer") should be modified

training_args = TrainingArguments(output_dir="test_trainer",report_to="none")

Environment

• Aim Version (e.g., 3.0.1)
• Python version 3.9
• pip version 3
• OS (e.g., Linux) unbuntu
• Any other relevant information

Additional context

import evaluate
import numpy as np
from aim.hugging_face import AimCallback
from datasets import load_dataset
from transformers import (AutoModelForSequenceClassification, AutoTokenizer,
                          Trainer, TrainingArguments)


def tokenize_function(examples):
    return tokenizer(examples['text'], padding='max_length', truncation=True)


def compute_metrics(eval_pred):
    logits, labels = eval_pred
    predictions = np.argmax(logits, axis=-1)
    return metric.compute(predictions=predictions, references=labels)


dataset = load_dataset('yelp_review_full')


tokenizer = AutoTokenizer.from_pretrained('bert-base-cased')

tokenized_datasets = dataset.map(tokenize_function, batched=True)

small_train_dataset = tokenized_datasets['train'].shuffle(seed=42).select(range(1000))
small_eval_dataset = tokenized_datasets['test'].shuffle(seed=42).select(range(1000))

metric = evaluate.load('accuracy')

model = AutoModelForSequenceClassification.from_pretrained(
    "bert-base-cased", num_labels=5
)

training_args = TrainingArguments(output_dir="test_trainer")

aim_callback = AimCallback(experiment="example_experiment")


trainer = Trainer(
    model=model,
    args=training_args,
    train_dataset=small_train_dataset,
    eval_dataset=small_eval_dataset,
    compute_metrics=compute_metrics,
    callbacks=[aim_callback],
)

trainer.train()