Trainer push to hub (huggingface#11328)

* Initial support for upload to hub

* push -> upload

* Fixes + examples

* Fix torchhub test

* Torchhub test I hate you

* push_model_to_hub -> push_to_hub

* Apply mixin to other pretrained models

* Remove ABC inheritance

* Add tests

* Typo

* Run tests

* Install git-lfs

* Change approach

* Add push_to_hub to all

* Staging test suite

* Typo

* Maybe like this?

* More deps

* Cache

* Adapt name

* Quality

* MOAR tests

* Put it in testing_utils

* Docs + torchhub last hope

* Styling

* Wrong method

* Typos

* Update src/transformers/file_utils.py

Co-authored-by: Julien Chaumond <julien@huggingface.co>

* Address review comments

* Apply suggestions from code review

Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>

Co-authored-by: Julien Chaumond <julien@huggingface.co>
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>

.circleci/config.yml

@@ -317,24 +317,33 @@ jobs:
             - store_artifacts:
                   path: ~/transformers/reports
 
-    run_tests_git_lfs:
+    run_tests_hub:
         working_directory: ~/transformers
         docker:
             - image: circleci/python:3.7
         environment:
+            HUGGINGFACE_CO_STAGING: yes
             RUN_GIT_LFS_TESTS: yes
             TRANSFORMERS_IS_CI: yes
         resource_class: xlarge
         parallelism: 1
         steps:
             - checkout
+            - restore_cache:
+                  keys:
+                      - v0.4-hub-{{ checksum "setup.py" }}
+                      - v0.4-{{ checksum "setup.py" }}
             - run: sudo apt-get install git-lfs
             - run: |
                 git config --global user.email "ci@dummy.com"
                 git config --global user.name "ci"
             - run: pip install --upgrade pip
-            - run: pip install .[testing]
-            - run: python -m pytest -sv ./tests/test_hf_api.py -k "HfLargefilesTest"
+            - run: pip install .[torch,sentencepiece,testing]
+            - save_cache:
+                  key: v0.4-hub-{{ checksum "setup.py" }}
+                  paths:
+                      - '~/.cache/pip'
+            - run: python -m pytest -sv ./tests/ -m is_staging_test
 
     build_doc:
         working_directory: ~/transformers
@@ -469,7 +478,7 @@ workflows:
             - run_tests_flax
             - run_tests_pipelines_torch
             - run_tests_pipelines_tf
-            - run_tests_git_lfs
+            - run_tests_hub
             - build_doc
             - deploy_doc: *workflow_filters
 #    tpu_testing_jobs:

docs/source/main_classes/model.rst

@@ -73,3 +73,10 @@ Generation
 
 .. autoclass:: transformers.generation_tf_utils.TFGenerationMixin
     :members:
+
+
+Pushing to the Hub
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+.. autoclass:: transformers.file_utils.PushToHubMixin
+    :members:

docs/source/model_sharing.rst

@@ -22,16 +22,14 @@ the `model hub <https://huggingface.co/models>`__.
 
     Optionally, you can join an existing organization or create a new one.
 
-Prepare your model for uploading
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 We have seen in the :doc:`training tutorial <training>`: how to fine-tune a model on a given task. You have probably
 done something similar on your task, either using the model directly in your own training loop or using the
 :class:`~.transformers.Trainer`/:class:`~.transformers.TFTrainer` class. Let's see how you can share the result on the
 `model hub <https://huggingface.co/models>`__.
 
 Model versioning
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 Since version v3.5.0, the model hub has built-in model versioning based on git and git-lfs. It is based on the paradigm
 that one model *is* one repo.
@@ -54,6 +52,106 @@ For instance:
     >>>   revision="v2.0.1" # tag name, or branch name, or commit hash
     >>> )
 
+
+Push your model from Python
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Preparation
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The first step is to make sure your credentials to the hub are stored somewhere. This can be done in two ways. If you
+have access to a terminal, you cam just run the following command in the virtual environment where you installed 🤗
+Transformers:
+
+.. code-block:: bash
+
+    transformers-cli login
+
+It will store your access token in the Hugging Face cache folder (by default :obj:`~/.cache/`).
+
+If you don't have an easy access to a terminal (for instance in a Colab session), you can find a token linked to your
+acount by going on `huggingface.co <https://huggingface.co/>`, click on your avatar on the top left corner, then on
+`Edit profile` on the left, just beneath your profile picture. In the submenu `API Tokens`, you will find your API
+token that you can just copy.
+
+Directly push your model to the hub
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Once you have an API token (either stored in the cache or copied and pasted in your notebook), you can directly push a
+finetuned model you saved in :obj:`save_drectory` by calling:
+
+.. code-block:: python
+
+    finetuned_model.push_to_hub("my-awesome-model")
+
+If you have your API token not stored in the cache, you will need to pass it with :obj:`use_auth_token=your_token`.
+This is also be the case for all the examples below, so we won't mention it again.
+
+This will create a repository in your namespace name :obj:`my-awesome-model`, so anyone can now run:
+
+.. code-block:: python
+
+    from transformers import AutoModel
+
+    model = AutoModel.from_pretrained("your_username/my-awesome-model")
+
+Even better, you can combine this push to the hub with the call to :obj:`save_pretrained`:
+
+.. code-block:: python
+
+    finetuned_model.save_pretrained(save_directory, push_to_hub=True, repo_name="my-awesome-model")
+
+If you are a premium user and want your model to be private, just add :obj:`private=True` to this call.
+
+If you are a member of an organization and want to push it inside the namespace of the organization instead of yours,
+just add :obj:`organization=my_amazing_org`.
+
+Add new files to your model repo
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Once you have pushed your model to the hub, you might want to add the tokenizer, or a version of your model for another
+framework (TensorFlow, PyTorch, Flax). This is super easy to do! Let's begin with the tokenizer. You can add it to the
+repo you created before like this
+
+.. code-block:: python
+
+    tokenizer.push_to_hub("my-awesome-model")
+
+If you know its URL (it should be :obj:`https://huggingface.co/username/repo_name`), you can also do:
+
+.. code-block:: python
+
+    tokenizer.push_to_hub(repo_url=my_repo_url)
+
+And that's all there is to it! It's also a very easy way to fix a mistake if one of the files online had a bug.
+
+To add a model for another backend, it's also super easy. Let's say you have fine-tuned a TensorFlow model and want to
+add the pytorch model files to your model repo, so that anyone in the community can use it. The following allows you to
+directly create a PyTorch version of your TensorFlow model:
+
+.. code-block:: python
+
+    from transfomers import AutoModel
+
+    model = AutoModel.from_pretrained(save_directory, from_tf=True)
+
+You can also replace :obj:`save_directory` by the identifier of your model (:obj:`username/repo_name`) if you don't
+have a local save of it anymore. Then, just do the same as before:
+
+.. code-block:: python
+
+    model.push_to_hub("my-awesome-model")
+
+or
+
+.. code-block:: python
+
+    model.push_to_hub(repo_url=my_repo_url)
+
+
+Use your terminal and git
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
 Basic steps
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 

examples/pytorch/language-modeling/run_clm.py

@@ -447,6 +447,9 @@ def group_texts(examples):
         trainer.log_metrics("eval", metrics)
         trainer.save_metrics("eval", metrics)
 
+    if training_args.push_to_hub:
+        trainer.push_to_hub()
+
 
 def _mp_fn(index):
     # For xla_spawn (TPUs)

examples/pytorch/language-modeling/run_mlm.py

@@ -476,6 +476,9 @@ def group_texts(examples):
         trainer.log_metrics("eval", metrics)
         trainer.save_metrics("eval", metrics)
 
+    if training_args.push_to_hub:
+        trainer.push_to_hub()
+
 
 def _mp_fn(index):
     # For xla_spawn (TPUs)

examples/pytorch/language-modeling/run_plm.py

@@ -452,6 +452,9 @@ def group_texts(examples):
         trainer.log_metrics("eval", metrics)
         trainer.save_metrics("eval", metrics)
 
+    if training_args.push_to_hub:
+        trainer.push_to_hub()
+
 
 def _mp_fn(index):
     # For xla_spawn (TPUs)

examples/pytorch/multiple-choice/run_swag.py

@@ -428,6 +428,9 @@ def compute_metrics(eval_predictions):
         trainer.log_metrics("eval", metrics)
         trainer.save_metrics("eval", metrics)
 
+    if training_args.push_to_hub:
+        trainer.push_to_hub()
+
 
 def _mp_fn(index):
     # For xla_spawn (TPUs)

examples/pytorch/question-answering/run_qa.py

@@ -599,6 +599,9 @@ def compute_metrics(p: EvalPrediction):
         trainer.log_metrics("test", metrics)
         trainer.save_metrics("test", metrics)
 
+    if training_args.push_to_hub:
+        trainer.push_to_hub()
+
 
 def _mp_fn(index):
     # For xla_spawn (TPUs)

examples/pytorch/question-answering/run_qa_beam_search.py

@@ -638,6 +638,9 @@ def compute_metrics(p: EvalPrediction):
         trainer.log_metrics("test", metrics)
         trainer.save_metrics("test", metrics)
 
+    if training_args.push_to_hub:
+        trainer.push_to_hub()
+
 
 def _mp_fn(index):
     # For xla_spawn (TPUs)

examples/pytorch/summarization/run_summarization.py

@@ -579,6 +579,9 @@ def compute_metrics(eval_preds):
                 with open(output_test_preds_file, "w") as writer:
                     writer.write("\n".join(test_preds))
 
+    if training_args.push_to_hub:
+        trainer.push_to_hub()
+
     return results
 
 

examples/pytorch/text-classification/run_glue.py

@@ -517,6 +517,9 @@ def compute_metrics(p: EvalPrediction):
                             item = label_list[item]
                             writer.write(f"{index}\t{item}\n")
 
+    if training_args.push_to_hub:
+        trainer.push_to_hub()
+
 
 def _mp_fn(index):
     # For xla_spawn (TPUs)

examples/pytorch/token-classification/run_ner.py

@@ -491,6 +491,9 @@ def compute_metrics(p):
                 for prediction in true_predictions:
                     writer.write(" ".join(prediction) + "\n")
 
+    if training_args.push_to_hub:
+        trainer.push_to_hub()
+
 
 def _mp_fn(index):
     # For xla_spawn (TPUs)

examples/pytorch/translation/run_translation.py

@@ -571,6 +571,9 @@ def compute_metrics(eval_preds):
                 with open(output_test_preds_file, "w") as writer:
                     writer.write("\n".join(test_preds))
 
+    if training_args.push_to_hub:
+        trainer.push_to_hub()
+
     return results
 
 

hubconf.py

@@ -31,7 +31,7 @@
 )
 
 
-dependencies = ["torch", "numpy", "tokenizers", "filelock", "requests", "tqdm", "regex", "sentencepiece", "sacremoses", "importlib_metadata"]
+dependencies = ["torch", "numpy", "tokenizers", "filelock", "requests", "tqdm", "regex", "sentencepiece", "sacremoses", "importlib_metadata", "huggingface_hub"]
 
 
 @add_start_docstrings(AutoConfig.__doc__)

src/transformers/configuration_utils.py

@@ -22,14 +22,14 @@
 from typing import Any, Dict, Tuple, Union
 
 from . import __version__
-from .file_utils import CONFIG_NAME, cached_path, hf_bucket_url, is_offline_mode, is_remote_url
+from .file_utils import CONFIG_NAME, PushToHubMixin, cached_path, hf_bucket_url, is_offline_mode, is_remote_url
 from .utils import logging
 
 
 logger = logging.get_logger(__name__)
 
 
-class PretrainedConfig(object):
+class PretrainedConfig(PushToHubMixin):
     r"""
     Base class for all configuration classes. Handles a few parameters common to all models' configurations as well as
     methods for loading/downloading/saving configurations.
@@ -310,14 +310,19 @@ def num_labels(self, num_labels: int):
             self.id2label = {i: f"LABEL_{i}" for i in range(num_labels)}
             self.label2id = dict(zip(self.id2label.values(), self.id2label.keys()))
 
-    def save_pretrained(self, save_directory: Union[str, os.PathLike]):
+    def save_pretrained(self, save_directory: Union[str, os.PathLike], push_to_hub: bool = False, **kwargs):
         """
         Save a configuration object to the directory ``save_directory``, so that it can be re-loaded using the
         :func:`~transformers.PretrainedConfig.from_pretrained` class method.
 
         Args:
             save_directory (:obj:`str` or :obj:`os.PathLike`):
                 Directory where the configuration JSON file will be saved (will be created if it does not exist).
+            push_to_hub (:obj:`bool`, `optional`, defaults to :obj:`False`):
+                Whether or not to push your model to the Hugging Face model hub after saving it.
+            kwargs:
+                Additional key word arguments passed along to the
+                :meth:`~transformers.file_utils.PushToHubMixin.push_to_hub` method.
         """
         if os.path.isfile(save_directory):
             raise AssertionError(f"Provided path ({save_directory}) should be a directory, not a file")
@@ -328,6 +333,10 @@ def save_pretrained(self, save_directory: Union[str, os.PathLike]):
         self.to_json_file(output_config_file, use_diff=True)
         logger.info(f"Configuration saved in {output_config_file}")
 
+        if push_to_hub:
+            url = self._push_to_hub(save_files=[output_config_file], **kwargs)
+            logger.info(f"Configuration pushed to the hub in this commit: {url}")
+
     @classmethod
     def from_pretrained(cls, pretrained_model_name_or_path: Union[str, os.PathLike], **kwargs) -> "PretrainedConfig":
         r"""