Remove Neptune integration references and deprecate NeptuneCallback (huggingface#42666)

Remove Neptune integration references and deprecate NeptuneCallback

Author: qgallouedec

docs/source/en/main_classes/callback.md

@@ -36,7 +36,6 @@ By default, `TrainingArguments.report_to` is set to `"all"`, so a [`Trainer`] wi
 - [`~integrations.WandbCallback`] if [wandb](https://www.wandb.com/) is installed.
 - [`~integrations.CometCallback`] if [comet_ml](https://www.comet.com/site/) is installed.
 - [`~integrations.MLflowCallback`] if [mlflow](https://www.mlflow.org/) is installed.
-- [`~integrations.NeptuneCallback`] if [neptune](https://neptune.ai/) is installed.
 - [`~integrations.AzureMLCallback`] if [azureml-sdk](https://pypi.org/project/azureml-sdk/) is
   installed.
 - [`~integrations.CodeCarbonCallback`] if [codecarbon](https://pypi.org/project/codecarbon/) is
@@ -84,8 +83,6 @@ Here is the list of the available [`TrainerCallback`] in the library:
 
 [[autodoc]] integrations.CodeCarbonCallback
 
-[[autodoc]] integrations.NeptuneCallback
-
 [[autodoc]] integrations.ClearMLCallback
 
 [[autodoc]] integrations.DagsHubCallback

docs/source/ja/main_classes/callback.md

@@ -37,7 +37,6 @@ rendered properly in your Markdown viewer.
 - [`~integrations.WandbCallback`] [wandb](https://www.wandb.com/) がインストールされている場合。
 - [`~integrations.CometCallback`] [comet_ml](https://www.comet.com/site/) がインストールされている場合。
 - [mlflow](https://www.mlflow.org/) がインストールされている場合は [`~integrations.MLflowCallback`]。
-- [`~integrations.NeptuneCallback`] [neptune](https://neptune.ai/) がインストールされている場合。
 - [`~integrations.AzureMLCallback`] [azureml-sdk](https://pypi.org/project/azureml-sdk/) の場合
   インストールされています。
 - [`~integrations.CodeCarbonCallback`] [codecarbon](https://pypi.org/project/codecarbon/) の場合
@@ -82,8 +81,6 @@ rendered properly in your Markdown viewer.
 
 [[autodoc]] integrations.CodeCarbonCallback
 
-[[autodoc]] integrations.NeptuneCallback
-
 [[autodoc]] integrations.ClearMLCallback
 
 [[autodoc]] integrations.DagsHubCallback

docs/source/ko/main_classes/callback.md

@@ -36,7 +36,6 @@ rendered properly in your Markdown viewer.
  사용됩니다.
 - [`~integrations.CometCallback`]는 [comet_ml](https://www.comet.com/site/)이 설치되어 있으면 사용됩니다.
 - [`~integrations.MLflowCallback`]는 [mlflow](https://www.mlflow.org/)가 설치되어 있으면 사용됩니다.
-- [`~integrations.NeptuneCallback`]는 [neptune](https://neptune.ai/)이 설치되어 있으면 사용됩니다.
 - [`~integrations.AzureMLCallback`]는 [azureml-sdk](https://pypi.org/project/azureml-sdk/)가 설치되어
  있으면 사용됩니다.
 - [`~integrations.CodeCarbonCallback`]는 [codecarbon](https://pypi.org/project/codecarbon/)이 설치되어
@@ -82,8 +81,6 @@ rendered properly in your Markdown viewer.
 
 [[autodoc]] integrations.CodeCarbonCallback
 
-[[autodoc]] integrations.NeptuneCallback
-
 [[autodoc]] integrations.ClearMLCallback
 
 [[autodoc]] integrations.DagsHubCallback

docs/source/zh/main_classes/callback.md

@@ -30,7 +30,6 @@ Callbacks是“只读”的代码片段，除了它们返回的[TrainerControl]
 - [`~integrations.WandbCallback`]，如果安装了[wandb](https://www.wandb.com/)。
 - [`~integrations.CometCallback`]，如果安装了[comet_ml](https://www.comet.com/site/)。
 - [`~integrations.MLflowCallback`]，如果安装了[mlflow](https://www.mlflow.org/)。
-- [`~integrations.NeptuneCallback`]，如果安装了[neptune](https://neptune.ai/)。
 - [`~integrations.AzureMLCallback`]，如果安装了[azureml-sdk](https://pypi.org/project/azureml-sdk/)。
 - [`~integrations.CodeCarbonCallback`]，如果安装了[codecarbon](https://pypi.org/project/codecarbon/)。
 - [`~integrations.ClearMLCallback`]，如果安装了[clearml](https://github.com/allegroai/clearml)。
@@ -71,8 +70,6 @@ Callbacks是“只读”的代码片段，除了它们返回的[TrainerControl]
 
 [[autodoc]] integrations.CodeCarbonCallback
 
-[[autodoc]] integrations.NeptuneCallback
-
 [[autodoc]] integrations.ClearMLCallback
 
 [[autodoc]] integrations.DagsHubCallback

examples/pytorch/README.md

@@ -199,7 +199,6 @@ You can easily log and monitor your runs code. The following are currently suppo
 * [TensorBoard](https://www.tensorflow.org/tensorboard)
 * [Weights & Biases](https://docs.wandb.ai/integrations/huggingface)
 * [Comet ML](https://www.comet.com/docs/v2/integrations/ml-frameworks/transformers/)
-* [Neptune](https://docs.neptune.ai/integrations-and-supported-tools/model-training/hugging-face)
 * [ClearML](https://clear.ml/docs/latest/docs/getting_started/ds/ds_first_steps)
 * [DVCLive](https://dvc.org/doc/dvclive/ml-frameworks/huggingface)
 
@@ -256,91 +255,6 @@ or if in a Conda environment:
 conda install -c comet_ml -c anaconda -c conda-forge comet_ml
 ```
 
-### Neptune
-
-First, install the Neptune client library. You can do it with either `pip` or `conda`:
-
-`pip`:
-
-```bash
-pip install neptune
-```
-
-`conda`:
-
-```bash
-conda install -c conda-forge neptune
-```
-
-Next, in your model training script, import `NeptuneCallback`:
-
-```python
-from transformers.integrations import NeptuneCallback
-```
-
-To enable Neptune logging, in your `TrainingArguments`, set the `report_to` argument to `"neptune"`:
-
-```python
-training_args = TrainingArguments(
-    "quick-training-distilbert-mrpc",
-    eval_strategy="steps",
-    eval_steps=20,
-    report_to="neptune",
-)
-
-trainer = Trainer(
-    model,
-    training_args,
-    ...
-)
-```
-
-**Note:** This method requires saving your Neptune credentials as environment variables (see the bottom of the section).
-
-Alternatively, for more logging options, create a Neptune callback:
-
-```python
-neptune_callback = NeptuneCallback()
-```
-
-To add more detail to the tracked run, you can supply optional arguments to `NeptuneCallback`.
-
-Some examples:
-
-```python
-neptune_callback = NeptuneCallback(
-    name = "DistilBERT",
-    description = "DistilBERT fine-tuned on GLUE/MRPC",
-    tags = ["args-callback", "fine-tune", "MRPC"],  # tags help you manage runs in Neptune
-    base_namespace="callback",  # the default is "finetuning"
-    log_checkpoints = "best",  # other options are "last", "same", and None
-    capture_hardware_metrics = False,  # additional keyword arguments for a Neptune run
-)
-```
-
-Pass the callback to the Trainer:
-
-```python
-training_args = TrainingArguments(..., report_to=None)
-trainer = Trainer(
-    model,
-    training_args,
-    ...
-    callbacks=[neptune_callback],
-)
-```
-
-Now, when you start the training with `trainer.train()`, your metadata will be logged in Neptune.
-
-**Note:** Although you can pass your **Neptune API token** and **project name** as arguments when creating the callback, the recommended way is to save them as environment variables:
-
-| Environment variable | Value                                                |
-| :------------------- | :--------------------------------------------------- |
-| `NEPTUNE_API_TOKEN`  | Your Neptune API token. To find and copy it, click your Neptune avatar and select **Get your API token**. |
-| `NEPTUNE_PROJECT` | The full name of your Neptune project (`workspace-name/project-name`). To find and copy it, head to **project settings** → **Properties**. |
-
-For detailed instructions and examples, see the [Neptune docs](https://docs.neptune.ai/integrations/transformers/).
-
 ### ClearML
 
 To use ClearML, install the clearml package with:

src/transformers/integrations/integration_utils.py

@@ -26,6 +26,7 @@
 import shutil
 import sys
 import tempfile
+import warnings
 from dataclasses import fields
 from enum import Enum
 from pathlib import Path
@@ -1455,6 +1456,10 @@ def __init__(self):
 class NeptuneCallback(TrainerCallback):
     """TrainerCallback that sends the logs to [Neptune](https://app.neptune.ai).
 
+    > [!WARNING]
+    > Neptune integration is deprecated and will be removed in a future version of Transformers. We recommend using
+    > other supported experiment tracking integrations.
+
     Args:
         api_token (`str`, *optional*): Neptune API token obtained upon registration.
             You can leave this argument out if you have saved your token to the `NEPTUNE_API_TOKEN` environment
@@ -1500,6 +1505,11 @@ def __init__(
         log_checkpoints: str | None = None,
         **neptune_run_kwargs,
     ):
+        warnings.warn(
+            "The NeptuneCallback is deprecated and will be removed in a future version of Transformers. We recommend "
+            "using other supported experiment tracking integrations.",
+            FutureWarning,
+        )
         if not is_neptune_available():
             raise ValueError(
                 "NeptuneCallback requires the Neptune client library to be installed. "

src/transformers/training_args.py

@@ -596,9 +596,9 @@ class TrainingArguments:
             instance of `Dataset`.
         report_to (`str` or `list[str]`, *optional*, defaults to `"none"`):
             The list of integrations to report the results and logs to. Supported platforms are `"azure_ml"`,
-            `"clearml"`, `"codecarbon"`, `"comet_ml"`, `"dagshub"`, `"dvclive"`, `"flyte"`, `"mlflow"`, `"neptune"`,
-            `"swanlab"`, `"tensorboard"`, `"trackio"` and `"wandb"`. Use `"all"` to report to all integrations
-            installed, `"none"` for no integrations.
+            `"clearml"`, `"codecarbon"`, `"comet_ml"`, `"dagshub"`, `"dvclive"`, `"flyte"`, `"mlflow"`, `"swanlab"`,
+            `"tensorboard"`, `"trackio"` and `"wandb"`. Use `"all"` to report to all integrations installed, `"none"`
+            for no integrations.
         project (`str`, *optional*, defaults to `"huggingface"`):
             The name of the project to use for logging. Currently, only used by Trackio.
         trackio_space_id (`str` or `None`, *optional*, defaults to `"trackio"`):
@@ -2386,8 +2386,8 @@ def set_logging(
             report_to (`str` or `list[str]`, *optional*, defaults to `"none"`):
                 The list of integrations to report the results and logs to. Supported platforms are `"azure_ml"`,
                 `"clearml"`, `"codecarbon"`, `"comet_ml"`, `"dagshub"`, `"dvclive"`, `"flyte"`, `"mlflow"`,
-                `"neptune"`, `"swanlab"`, `"tensorboard"`, `"trackio"` and `"wandb"`. Use `"all"` to report to all
-                integrations installed, `"none"` for no integrations.
+                `"swanlab"`, `"tensorboard"`, `"trackio"` and `"wandb"`. Use `"all"` to report to all integrations
+                installed, `"none"` for no integrations.
             first_step (`bool`, *optional*, defaults to `False`):
                 Whether to log and evaluate the first `global_step` or not.
             nan_inf_filter (`bool`, *optional*, defaults to `True`):