README.md

<!---
Copyright 2020 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
-->

# Examples

This folder contains actively maintained examples of use of 🤗 Transformers organized along NLP tasks. If you are looking for an example that used to
be in this folder, it may have moved to our [research projects](https://github.com/huggingface/transformers/tree/master/examples/research_projects) subfolder (which contains frozen snapshots of research projects).

## Important note

**Important**

To make sure you can successfully run the latest versions of the example scripts, you have to **install the library from source** and install some example-specific requirements. To do this, execute the following steps in a new virtual environment:
```bash
git clone https://github.com/huggingface/transformers
cd transformers
pip install .
```
Then cd in the example folder of your choice and run
```bash
pip install -r requirements.txt
```

Alternatively, you can run the version of the examples as they were for your current version of Transformers via (for instance with v3.5.1):
```bash
git checkout tags/v3.5.1
```

## The Big Table of Tasks

Here is the list of all our examples:
- with information on whether they are **built on top of `Trainer`/`TFTrainer`** (if not, they still work, they might
  just lack some features),
- whether or not they leverage the [🤗 Datasets](https://github.com/huggingface/datasets) library.
- links to **Colab notebooks** to walk through the scripts and run them easily,
<!--
Coming soon!
- links to **Cloud deployments** to be able to deploy large-scale trainings in the Cloud with little to no setup.
-->

| Task | Example datasets | Trainer support | TFTrainer support | 🤗 Datasets | Colab
|---|---|:---:|:---:|:---:|:---:|
| [**`language-modeling`**](https://github.com/huggingface/transformers/tree/master/examples/language-modeling)       | Raw text        | ✅ | -  | ✅ | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/huggingface/blog/blob/master/notebooks/01_how_to_train.ipynb)
| [**`multiple-choice`**](https://github.com/huggingface/transformers/tree/master/examples/multiple-choice)           | SWAG, RACE, ARC | ✅ | ✅ | ✅ | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/ViktorAlm/notebooks/blob/master/MPC_GPU_Demo_for_TF_and_PT.ipynb)
| [**`question-answering`**](https://github.com/huggingface/transformers/tree/master/examples/question-answering)     | SQuAD           | ✅ | ✅ | ✅ | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/huggingface/notebooks/blob/master/examples/question_answering.ipynb)
| [**`summarization`**](https://github.com/huggingface/transformers/tree/master/examples/seq2seq)                     | CNN/Daily Mail  | ✅  | - | - | -
| [**`text-classification`**](https://github.com/huggingface/transformers/tree/master/examples/text-classification)   | GLUE, XNLI      | ✅ | ✅ | ✅ | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/huggingface/notebooks/blob/master/examples/text_classification.ipynb)
| [**`text-generation`**](https://github.com/huggingface/transformers/tree/master/examples/text-generation)           | -               | n/a | n/a | - | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/huggingface/blog/blob/master/notebooks/02_how_to_generate.ipynb)
| [**`token-classification`**](https://github.com/huggingface/transformers/tree/master/examples/token-classification) | CoNLL NER       | ✅ | ✅ | ✅ | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/huggingface/notebooks/blob/master/examples/token_classification.ipynb)
| [**`translation`**](https://github.com/huggingface/transformers/tree/master/examples/seq2seq)                       | WMT             | ✅  | - | - | -


<!--
## One-click Deploy to Cloud (wip)

**Coming soon!**
-->

## Distributed training and mixed precision

All the PyTorch scripts mentioned above work out of the box with distributed training and mixed precision, thanks to
the [Trainer API](https://huggingface.co/transformers/main_classes/trainer.html). To launch one of them on _n_ GPUS,
use the following command:

```bash
python -m torch.distributed.launch \
    --nproc_per_node number_of_gpu_you_have path_to_script.py \
	--all_arguments_of_the_script 
```

As an example, here is how you would fine-tune the BERT large model (with whole word masking) on the text
classification MNLI task using the `run_glue` script, with 8 GPUs:

```bash
python -m torch.distributed.launch \
    --nproc_per_node 8 text-classification/run_glue.py \
    --model_name_or_path bert-large-uncased-whole-word-masking \
    --task_name mnli \
    --do_train \
    --do_eval \
    --max_seq_length 128 \
    --per_device_train_batch_size 8 \
    --learning_rate 2e-5 \
    --num_train_epochs 3.0 \
    --output_dir /tmp/mnli_output/
```

If you have a GPU with mixed precision capabilities (architecture Pascal or more recent), you can use mixed precision
training with PyTorch 1.6.0 or latest, or by installing the [Apex](https://github.com/NVIDIA/apex) library for previous
versions. Just add the flag `--fp16` to your command launching one of the scripts mentioned above!

Using mixed precision training usually results in 2x-speedup for training with the same final results (as shown in
[this table](https://github.com/huggingface/transformers/tree/master/examples/text-classification#mixed-precision-training)
for text classification).

## Running on TPUs

When using Tensorflow, TPUs are supported out of the box as a `tf.distribute.Strategy`.

When using PyTorch, we support TPUs thanks to `pytorch/xla`. For more context and information on how to setup your TPU environment refer to Google's documentation and to the
very detailed [pytorch/xla README](https://github.com/pytorch/xla/blob/master/README.md).

In this repo, we provide a very simple launcher script named
[xla_spawn.py](https://github.com/huggingface/transformers/tree/master/examples/xla_spawn.py) that lets you run our
example scripts on multiple TPU cores without any boilerplate. Just pass a `--num_cores` flag to this script, then your
regular training script with its arguments (this is similar to the `torch.distributed.launch` helper for
`torch.distributed`):

```bash
python xla_spawn.py --num_cores num_tpu_you_have \
    path_to_script.py \
	--all_arguments_of_the_script 
```

As an example, here is how you would fine-tune the BERT large model (with whole word masking) on the text
classification MNLI task using the `run_glue` script, with 8 TPUs:

```bash
python xla_spawn.py --num_cores 8 \
    text-classification/run_glue.py \
    --model_name_or_path bert-large-uncased-whole-word-masking \
    --task_name mnli \
    --do_train \
    --do_eval \
    --max_seq_length 128 \
    --per_device_train_batch_size 8 \
    --learning_rate 2e-5 \
    --num_train_epochs 3.0 \
    --output_dir /tmp/mnli_output/
```

## Logging & Experiment tracking

You can easily log and monitor your runs code. The following are currently supported:

* [TensorBoard](https://www.tensorflow.org/tensorboard)
* [Weights & Biases](https://docs.wandb.ai/integrations/huggingface)
* [Comet ML](https://www.comet.ml/docs/python-sdk/huggingface/)

### Weights & Biases

To use Weights & Biases, install the wandb package with:

```bash
pip install wandb
```

Then log in the command line:

```bash
wandb login
```

If you are in Jupyter or Colab, you should login with:

```python
import wandb
wandb.login()
```

To enable logging to W&B, include `"wandb"` in the `report_to` of your `TrainingArguments` or script. Or just pass along `--report_to all` if you have `wandb` installed.

Whenever you use `Trainer` or `TFTrainer` classes, your losses, evaluation metrics, model topology and gradients (for `Trainer` only) will automatically be logged.

Advanced configuration is possible by setting environment variables:

<table>
  <thead>
    <tr>
      <th style="text-align:left">Environment Variables</th>
      <th style="text-align:left">Options</th>
    </tr>
  </thead>
  <tbody>
    <tr>
      <td style="text-align:left">WANDB_LOG_MODEL</td>
      <td style="text-align:left">Log the model as artifact at the end of training (<b>false</b> by default)</td>
    </tr>
    <tr>
      <td style="text-align:left">WANDB_WATCH</td>
      <td style="text-align:left">
        <ul>
          <li><b>gradients</b> (default): Log histograms of the gradients</li>
          <li><b>all</b>: Log histograms of gradients and parameters</li>
          <li><b>false</b>: No gradient or parameter logging</li>
        </ul>
      </td>
    </tr>
    <tr>
      <td style="text-align:left">WANDB_PROJECT</td>
      <td style="text-align:left">Organize runs by project</td>
    </tr>
  </tbody>
</table>

Set run names with `run_name` argument present in scripts or as part of `TrainingArguments`.

Additional configuration options are available through generic [wandb environment variables](https://docs.wandb.com/library/environment-variables).

Refer to related [documentation & examples](https://docs.wandb.ai/integrations/huggingface).

### Comet.ml

To use `comet_ml`, install the Python package with:

```bash
pip install comet_ml
```

or if in a Conda environment:

```bash
conda install -c comet_ml -c anaconda -c conda-forge comet_ml
```