Metadata-Version: 2.5
Name: megh
Version: 0.0.2
Summary: MeghCloud metrics tracking SDK for ML training runs
Requires-Python: >=3.13
Requires-Dist: influxdb-client>=1.36.0
Provides-Extra: huggingface
Requires-Dist: transformers>=4.0.0; extra == 'huggingface'
Description-Content-Type: text/markdown

# megh

MeghCloud metrics tracking SDK for ML training runs.

## Installation

```bash
pip install megh[huggingface]
```

## Usage with Hugging Face Trainer

```python
import megh
from transformers import Trainer, TrainingArguments

# Register megh_metrics with HuggingFace Trainer
megh.init(framework="hf_trainer")

# Use Trainer as usual — megh_metrics is automatically enabled
training_args = TrainingArguments(
    output_dir="./results",
    num_train_epochs=3,
    per_device_train_batch_size=8,
    logging_steps=10,
)

trainer = Trainer(
    model=model,
    args=training_args,
    train_dataset=train_dataset,
)

trainer.train()
```

Metrics, hyperparameters, and model config are automatically tracked.
