SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

Get started docs (#15098)

Browse Source 
* clean commit of changes

* apply review feedback, make edits

* fix backticks, minor formatting

* 🖍 make fixup and minor edits

* 🖍 fix # in header

* 📝 update code sample without from_pt

* 📝 final review
This commit is contained in:
Steven Liu
2022-01-28 19:01:37 -06:00
committed by GitHub
parent cabd6d26a2
commit 16d4acbfdb
3 changed files with 286 additions and 410 deletions
Download Patch File Download Diff File Expand all files Collapse all files

211
docs/source/installation.mdx
View File

@@ -1,5 +1,5 @@
<!---
Copyright 2020 The HuggingFace Team. All rights reserved.
Copyright 2022 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.
@@ -16,183 +16,220 @@ limitations under the License.
# Installation
🤗 Transformers is tested on Python 3.6+, and PyTorch 1.1.0+ or TensorFlow 2.0+.
Install 🤗 Transformers for whichever deep learning library you're working with, setup your cache, and optionally configure 🤗 Transformers to run offline.
You should install 🤗 Transformers in a [virtual environment](https://docs.python.org/3/library/venv.html). If you're
unfamiliar with Python virtual environments, check out the [user guide](https://packaging.python.org/guides/installing-using-pip-and-virtual-environments/). Create a virtual environment with the version of Python you're going
to use and activate it.
🤗 Transformers is tested on Python 3.6+, PyTorch 1.1.0+, TensorFlow 2.0+, and Flax. Follow the installation instructions below for the deep learning library you are using:
Now, if you want to use 🤗 Transformers, you can install it with pip. If you'd like to play with the examples, you
must install it from source.
* [PyTorch](https://pytorch.org/get-started/locally/) installation instructions.
* [TensorFlow 2.0](https://www.tensorflow.org/install/pip) installation instructions.
* [Flax](https://flax.readthedocs.io/en/latest/) installation instructions.
## Installation with pip
## Install with pip
First you need to install one of, or both, TensorFlow 2.0 and PyTorch.
Please refer to [TensorFlow installation page](https://www.tensorflow.org/install/pip#tensorflow-2.0-rc-is-available),
[PyTorch installation page](https://pytorch.org/get-started/locally/#start-locally) and/or
[Flax installation page](https://github.com/google/flax#quick-install)
regarding the specific install command for your platform.
You should install 🤗 Transformers in a [virtual environment](https://docs.python.org/3/library/venv.html). If you're unfamiliar with Python virtual environments, take a look at this [guide](https://packaging.python.org/guides/installing-using-pip-and-virtual-environments/). A virtual environment makes it easier to manage different projects, and avoid compatibility issues between dependencies.
When TensorFlow 2.0 and/or PyTorch has been installed, 🤗 Transformers can be installed using pip as follows:
Start by creating a virtual environment in your project directory:
```bash
python -m venv .env
```
Activate the virtual environment:
```bash
source .env/bin/activate
```
Now you're ready to install 🤗 Transformers with the following command:
```bash
pip install transformers
```
Alternatively, for CPU-support only, you can install 🤗 Transformers and PyTorch in one line with:
For CPU-support only, you can conveniently install 🤗 Transformers and a deep learning library in one line. For example, install 🤗 Transformers and PyTorch with:
```bash
pip install transformers[torch]
```
or 🤗 Transformers and TensorFlow 2.0 in one line with:
🤗 Transformers and TensorFlow 2.0:
```bash
pip install transformers[tf-cpu]
```
or 🤗 Transformers and Flax in one line with:
🤗 Transformers and Flax:
```bash
pip install transformers[flax]
```
To check 🤗 Transformers is properly installed, run the following command:
Finally, check if 🤗 Transformers has been properly installed by running the following command. It will download a pretrained model:
```bash
python -c "from transformers import pipeline; print(pipeline('sentiment-analysis')('we love you'))"
```
It should download a pretrained model then print something like
Then print out the label and score:
```bash
[{'label': 'POSITIVE', 'score': 0.9998704791069031}]
```
(Note that TensorFlow will print additional stuff before that last statement.)
## Install from source
## Installing from source
Here is how to quickly install `transformers` from source:
Install 🤗 Transformers from source with the following command:
```bash
pip install git+https://github.com/huggingface/transformers
```
Note that this will install not the latest released version, but the bleeding edge `master` version, which you may want to use in case a bug has been fixed since the last official release and a new release hasn't been yet rolled out.
This command installs the bleeding edge `master` version rather than the latest `stable` version. The `master` version is useful for staying up-to-date with the latest developments. For instance, if a bug has been fixed since the last official release but a new release hasn't been rolled out yet. However, this means the `master` version may not always be stable. We strive to keep the `master` version operational, and most issues are usually resolved within a few hours or a day. If you run into a problem, please open an [Issue](https://github.com/huggingface/transformers/issues) so we can fix it even sooner!
While we strive to keep `master` operational at all times, if you notice some issues, they usually get fixed within a few hours or a day and you're more than welcome to help us detect any problems by opening an [Issue](https://github.com/huggingface/transformers/issues) and this way, things will get fixed even sooner.
Again, you can run:
Check if 🤗 Transformers has been properly installed by running the following command:
```bash
python -c "from transformers import pipeline; print(pipeline('sentiment-analysis')('I hate you'))"
python -c "from transformers import pipeline; print(pipeline('sentiment-analysis')('I love you'))"
```
to check 🤗 Transformers is properly installed.
## Editable install
If you want to constantly use the bleeding edge `master` version of the source code, or if you want to contribute to the library and need to test the changes in the code you're making, you will need an editable install. This is done by cloning the repository and installing with the following commands:
You will need an editable install if you'd like to:
``` bash
* Use the `master` version of the source code.
* Contribute to 🤗 Transformers and need to test changes in the code.
Clone the repository and install 🤗 Transformers with the following commands:
```bash
git clone https://github.com/huggingface/transformers.git
cd transformers
pip install -e .
```
This command performs a magical link between the folder you cloned the repository to and your python library paths, and it'll look inside this folder in addition to the normal library-wide paths. So if normally your python packages get installed into:
```
~/anaconda3/envs/main/lib/python3.7/site-packages/
```
now this editable install will reside where you clone the folder to, e.g. `~/transformers/` and python will search it too.
These commands will link the folder you cloned the repository to and your Python library paths. Python will now look inside the folder you cloned to in addition to the normal library paths. For example, if your Python packages are typically installed in `~/anaconda3/envs/main/lib/python3.7/site-packages/`, Python will also search the folder you cloned to: `~/transformers/`.
Do note that you have to keep that `transformers` folder around and not delete it to continue using the `transformers` library.
<Tip warning={true}>
Now, let's get to the real benefit of this installation approach. Say, you saw some new feature has been just committed into `master`. If you have already performed all the steps above, to update your transformers to include all the latest commits, all you need to do is to `cd` into that cloned repository folder and update the clone to the latest version:
You must keep the `transformers` folder if you want to keep using the library.
```
</Tip>
Now you can easily update your clone to the latest version of 🤗 Transformers with the following command:
```bash
cd ~/transformers/
git pull
```
There is nothing else to do. Your python environment will find the bleeding edge version of `transformers` on the next run.
Your Python environment will find the `master` version of 🤗 Transformers on the next run.
## Install with conda
## With conda
Install from the conda channel `huggingface`:
Since Transformers version v4.0.0, we now have a conda channel: `huggingface`.
🤗 Transformers can be installed using conda as follows:
```
```bash
conda install -c huggingface transformers
```
Follow the installation pages of TensorFlow, PyTorch or Flax to see how to install them with conda.
## Cache setup
## Caching models
Pretrained models are downloaded and locally cached at: `~/.cache/huggingface/transformers/`. This is the default directory given by the shell environment variable `TRANSFORMERS_CACHE`. On Windows, the default directory is given by `C:\Users\username\.cache\huggingface\transformers`. You can change the shell environment variables shown below - in order of priority - to specify a different cache directory:
This library provides pretrained models that will be downloaded and cached locally. Unless you specify a location with
`cache_dir=...` when you use methods like `from_pretrained`, these models will automatically be downloaded in the
folder given by the shell environment variable ``TRANSFORMERS_CACHE``. The default value for it will be the Hugging
Face cache home followed by ``/transformers/``. This is (by order of priority):
1. Shell environment variable (default): `TRANSFORMERS_CACHE`.
2. Shell environment variable: `HF_HOME` + `transformers/`.
3. Shell environment variable: `XDG_CACHE_HOME` + `/huggingface/transformers`.
* shell environment variable ``HF_HOME``
* shell environment variable ``XDG_CACHE_HOME`` + ``/huggingface/``
* default: ``~/.cache/huggingface/``
<Tip>
So if you don't have any specific environment variable set, the cache directory will be at
``~/.cache/huggingface/transformers/``.
🤗 Transformers will use the shell environment variables `PYTORCH_TRANSFORMERS_CACHE` or `PYTORCH_PRETRAINED_BERT_CACHE` if you are coming from an earlier iteration of this library and have set those environment variables, unless you specify the shell environment variable `TRANSFORMERS_CACHE`.
**Note:** If you have set a shell environment variable for one of the predecessors of this library
(``PYTORCH_TRANSFORMERS_CACHE`` or ``PYTORCH_PRETRAINED_BERT_CACHE``), those will be used if there is no shell
environment variable for ``TRANSFORMERS_CACHE``.
</Tip>
### Offline mode
## Offline mode
It's possible to run 🤗 Transformers in a firewalled or a no-network environment.
🤗 Transformers is able to run in a firewalled or offline environment by only using local files. Set the environment variable `TRANSFORMERS_OFFLINE=1` to enable this behavior.
Setting environment variable `TRANSFORMERS_OFFLINE=1` will tell 🤗 Transformers to use local files only and will not try to look things up.
<Tip>
Most likely you may want to couple this with `HF_DATASETS_OFFLINE=1` that performs the same for 🤗 Datasets if you're using the latter.
Add [🤗 Datasets](https://huggingface.co/docs/datasets/) to your offline training workflow by setting the environment variable `HF_DATASETS_OFFLINE=1`.
Here is an example of how this can be used on a filesystem that is shared between a normally networked and a firewalled to the external world instances.
</Tip>
On the instance with the normal network run your program which will download and cache models (and optionally datasets if you use 🤗 Datasets). For example:
For example, you would typically run a program on a normal network firewalled to external instances with the following command:
```
```bash
python examples/pytorch/translation/run_translation.py --model_name_or_path t5-small --dataset_name wmt16 --dataset_config ro-en ...
```
and then with the same filesystem you can now run the same program on a firewalled instance:
```
Run this same program in an offline instance with:
```bash
HF_DATASETS_OFFLINE=1 TRANSFORMERS_OFFLINE=1 \
python examples/pytorch/translation/run_translation.py --model_name_or_path t5-small --dataset_name wmt16 --dataset_config ro-en ...
```
and it should succeed without any hanging waiting to timeout.
#### Fetching models and tokenizers to use offline
The script should now run without hanging or waiting to timeout because it knows it should only look for local files.
When running a script the first time like mentioned above, the downloaded files will be cached for future reuse.
However, it is also possible to download files and point to their local path instead.
### Fetch models and tokenizers to use offline
Downloading files can be done through the Web Interface by clicking on the "Download" button, but it can also be handled
programmatically using the `huggingface_hub` library that is a dependency to `transformers`:
Another option for using 🤗 Transformers offline is to download the files ahead of time, and then point to their local path when you need to use them offline. There are three ways to do this:
- Using `snapshot_download` to download an entire repository
- Using `hf_hub_download` to download a specific file
* Download a file through the user interface on the [Model Hub](https://huggingface.co/models) by clicking on the ↓ icon.
See the reference for these methods in the huggingface_hub
[documentation](https://github.com/huggingface/huggingface_hub/tree/main/src/huggingface_hub).
![download-icon](https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/download-icon.png)
## Do you want to run a Transformer model on a mobile device?
* Use the [`PreTrainedModel.from_pretrained`] and [`PreTrainedModel.save_pretrained`] workflow:
You should check out our [swift-coreml-transformers](https://github.com/huggingface/swift-coreml-transformers) repo.
1. Download your files ahead of time with [`PreTrainedModel.from_pretrained`]:
It contains a set of tools to convert PyTorch or TensorFlow 2.0 trained Transformer models (currently contains `GPT-2`,
`DistilGPT-2`, `BERT`, and `DistilBERT`) to CoreML models that run on iOS devices.
```py
>>> from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
At some point in the future, you'll be able to seamlessly move from pretraining or fine-tuning models in PyTorch or
TensorFlow 2.0 to productizing them in CoreML, or prototype a model or an app in CoreML then research its
hyperparameters or architecture from PyTorch or TensorFlow 2.0. Super exciting!
>>> tokenizer = AutoTokenizer.from_pretrained("bigscience/T0_3B")
>>> model = AutoModelForSeq2SeqLM.from_pretrained("bigscience/T0_3B")
```
2. Save your files to a specified directory with [`PreTrainedModel.save_pretrained`]:
```py
>>> tokenizer.save_pretrained("./your/path/bigscience_t0")
>>> model.save_pretrained("./your/path/bigscience_t0")
```
3. Now when you're offline, reload your files with [`PreTrainedModel.from_pretrained`] from the specified directory:
```py
>>> tokenizer = AutoTokenizer.from_pretrained("./your/path/bigscience_t0")
>>> model = AutoModel.from_pretrained("./your/path/bigscience_t0")
```
* Programmatically download files with the [huggingface_hub](https://github.com/huggingface/huggingface_hub/tree/main/src/huggingface_hub) library:
1. Install the `huggingface_hub` library in your virtual environment:
```bash
python -m pip install huggingface_hub
```
2. Use the [`hf_hub_download`](https://huggingface.co/docs/hub/adding-a-library#download-files-from-the-hub) function to download a file to a specific path. For example, the following command downloads the `config.json` file from the [T0](https://huggingface.co/bigscience/T0_3B) model to your desired path:
```py
>>> from huggingface_hub import hf_hub_download
>>> hf_hub_download(repo_id="bigscience/T0_3B", filename="config.json", cache_dir="./your/path/bigscience_t0")
```
Once your file is downloaded and locally cached, specify it's local path to load and use it:
```py
>>> from transformers import AutoConfig
>>> config = AutoConfig.from_pretrained("./your/path/bigscience_t0/config.json")
```
<Tip>
See the [How to download files from the Hub](https://huggingface.co/docs/hub/how-to-downstream) section for more details on downloading files stored on the Hub.
</Tip>