hf: warn of deprecating internal callback #740
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,333 @@ | ||
| { | ||
| "cells": [ | ||
| { | ||
| "cell_type": "markdown", | ||
| "metadata": { | ||
| "id": "3SJ8SY6ldmsS" | ||
| }, | ||
| "source": [ | ||
| "### How to do Experiment tracking with DVCLive\n", | ||
| "\n", | ||
| "What you will learn?\n", | ||
| "\n", | ||
| "- Fine-tuning a model on a binary text classification task\n", | ||
| "- Track machine learning experiments with DVCLive\n", | ||
| "- Visualize results and create a report\n" | ||
| ] | ||
| }, | ||
| { | ||
| "cell_type": "markdown", | ||
| "metadata": { | ||
| "id": "nxiSBytidmsU" | ||
| }, | ||
| "source": [ | ||
| "#### Setup (Install Dependencies & Setup Git)\n", | ||
| "\n", | ||
| "- Install accelerate , Datasets , evaluate , transformers and dvclive\n", | ||
| "- Start a Git repo. Your experiments will be saved in a commit but hidden in\n", | ||
| " order to not clutter your repo.\n", | ||
| "- Initialize DVC\n" | ||
| ] | ||
| }, | ||
| { | ||
| "cell_type": "code", | ||
| "execution_count": null, | ||
| "metadata": { | ||
| "id": "CLRgy2W4dmsU" | ||
| }, | ||
| "outputs": [], | ||
| "source": [ | ||
| "!pip install datasets dvclive evaluate pandas 'transformers[torch]' --upgrade" | ||
| ] | ||
| }, | ||
| { | ||
| "cell_type": "code", | ||
| "execution_count": null, | ||
| "metadata": { | ||
| "id": "fo0sq84UdmsV" | ||
| }, | ||
| "outputs": [], | ||
| "source": [ | ||
| "!git init -q\n", | ||
| "!git config --local user.email \"you@example.com\"\n", | ||
| "!git config --local user.name \"Your Name\"\n", | ||
| "!dvc init -q\n", | ||
| "!git commit -m \"DVC init\"" | ||
| ] | ||
| }, | ||
| { | ||
| "cell_type": "markdown", | ||
| "metadata": { | ||
| "id": "T5WYJ31UdmsV" | ||
| }, | ||
| "source": [ | ||
| "### Fine-tuning a model on a text classification task\n", | ||
| "\n", | ||
| "#### Loading the dataset\n", | ||
| "\n", | ||
| "We will use the [imdb](https://huggingface.co/datasets/imdb) Large Movie Review Dataset. This is a dataset for binary\n", | ||
| "sentiment classification containing a set of 25K movie reviews for training and\n", | ||
| "25K for testing.\n" | ||
| ] | ||
| }, | ||
| { | ||
| "cell_type": "code", | ||
| "execution_count": null, | ||
| "metadata": { | ||
| "id": "41fP0WCbdmsV" | ||
| }, | ||
| "outputs": [], | ||
| "source": [ | ||
| "from datasets import load_dataset\n", | ||
| "from transformers import AutoTokenizer\n", | ||
| "\n", | ||
| "dataset = load_dataset(\"imdb\")" | ||
| ] | ||
| }, | ||
| { | ||
| "cell_type": "markdown", | ||
| "metadata": { | ||
| "id": "V3gDKbbSdmsV" | ||
| }, | ||
| "source": [ | ||
| "#### Preprocessing the data\n", | ||
| "\n", | ||
| "We use `transformers.AutoTokenizer` which transforms the inputs and put them in a format\n", | ||
| "the model expects.\n" | ||
| ] | ||
| }, | ||
| { | ||
| "cell_type": "code", | ||
| "execution_count": null, | ||
| "metadata": { | ||
| "id": "uVr5lufodmsV" | ||
| }, | ||
| "outputs": [], | ||
| "source": [ | ||
| "tokenizer = AutoTokenizer.from_pretrained(\"distilbert-base-cased\")\n", | ||
| "\n", | ||
| "def tokenize_function(examples):\n", | ||
| " return tokenizer(examples[\"text\"], padding=\"max_length\", truncation=True)\n", | ||
| "\n", | ||
| "small_train_dataset = dataset[\"train\"].shuffle(seed=42).select(range(2000)).map(tokenize_function, batched=True)\n", | ||
| "small_eval_dataset = dataset[\"test\"].shuffle(seed=42).select(range(200)).map(tokenize_function, batched=True)" | ||
| ] | ||
| }, | ||
| { | ||
| "cell_type": "markdown", | ||
| "metadata": { | ||
| "id": "g9sELYMHdmsV" | ||
| }, | ||
| "source": [ | ||
| "#### Define evaluation metrics\n", | ||
| "\n", | ||
| "f1 is a metric for combining precision and recall metrics in one unique value, so\n", | ||
| "we take this criteria for evaluating the models.\n" | ||
| ] | ||
| }, | ||
| { | ||
| "cell_type": "code", | ||
| "execution_count": null, | ||
| "metadata": { | ||
| "id": "wmJoy5V-dmsW" | ||
| }, | ||
| "outputs": [], | ||
| "source": [ | ||
| "import numpy as np\n", | ||
| "import evaluate\n", | ||
| "\n", | ||
| "metric = evaluate.load(\"f1\")\n", | ||
| "\n", | ||
| "def compute_metrics(eval_pred):\n", | ||
| " logits, labels = eval_pred\n", | ||
| " predictions = np.argmax(logits, axis=-1)\n", | ||
| " return metric.compute(predictions=predictions, references=labels)" | ||
| ] | ||
| }, | ||
| { | ||
| "cell_type": "markdown", | ||
| "metadata": { | ||
| "id": "NwFntrIKdmsW" | ||
| }, | ||
| "source": [ | ||
| "### Training and Tracking experiments with DVCLive\n", | ||
| "\n", | ||
| "Track experiments in DVC by changing a few lines of your Python code.\n", | ||
| "Save model artifacts using `HF_DVCLIVE_LOG_MODEL=true`." | ||
| ] | ||
| }, | ||
| { | ||
| "cell_type": "code", | ||
| "source": [ | ||
| "%env HF_DVCLIVE_LOG_MODEL=true" | ||
| ], | ||
| "metadata": { | ||
| "id": "-A1oXCxE4zGi" | ||
| }, | ||
| "execution_count": null, | ||
| "outputs": [] | ||
| }, | ||
| { | ||
| "cell_type": "code", | ||
| "execution_count": null, | ||
| "metadata": { | ||
| "id": "gKKSTh0ZdmsW" | ||
| }, | ||
| "outputs": [], | ||
| "source": [ | ||
| "from dvclive.huggingface import DVCLiveCallback\n", | ||
| "from transformers import AutoModelForSequenceClassification, TrainingArguments, Trainer\n", | ||
| "\n", | ||
| "model = AutoModelForSequenceClassification.from_pretrained(\"distilbert-base-cased\", num_labels=2)\n", | ||
| "for param in model.base_model.parameters():\n", | ||
| " param.requires_grad = False\n", | ||
| "\n", | ||
| "lr = 3e-4\n", | ||
| "\n", | ||
| "training_args = TrainingArguments(\n", | ||
| " evaluation_strategy=\"epoch\",\n", | ||
| " learning_rate=lr,\n", | ||
| " logging_strategy=\"epoch\",\n", | ||
| " num_train_epochs=5,\n", | ||
| " output_dir=\"output\",\n", | ||
| " overwrite_output_dir=True,\n", | ||
| " load_best_model_at_end=True,\n", | ||
| " save_strategy=\"epoch\",\n", | ||
| " weight_decay=0.01,\n", | ||
| ")\n", | ||
| "\n", | ||
| "trainer = Trainer(\n", | ||
| " model=model,\n", | ||
| " args=training_args,\n", | ||
| " train_dataset=small_train_dataset,\n", | ||
| " eval_dataset=small_eval_dataset,\n", | ||
| " compute_metrics=compute_metrics,\n", | ||
| ")\n", | ||
| "trainer.train()" | ||
| ] | ||
| }, | ||
| { | ||
| "cell_type": "markdown", | ||
| "source": [ | ||
| "To customize tracking, include `transformers.integrations.DVCLiveCallback` in the `Trainer` callbacks and pass additional keyword arguments to `dvclive.Live`." | ||
| ], | ||
| "metadata": { | ||
| "id": "KKJCw0Vj6UTw" | ||
| } | ||
| }, | ||
| { | ||
| "cell_type": "code", | ||
| "source": [ | ||
| "from dvclive import Live\n", | ||
| "from transformers.integrations import DVCLiveCallback\n", | ||
| "\n", | ||
| "lr = 1e-4\n", | ||
| "\n", | ||
| "training_args = TrainingArguments(\n", | ||
| " evaluation_strategy=\"epoch\",\n", | ||
| " learning_rate=lr,\n", | ||
| " logging_strategy=\"epoch\",\n", | ||
| " num_train_epochs=5,\n", | ||
| " output_dir=\"output\",\n", | ||
| " overwrite_output_dir=True,\n", | ||
| " load_best_model_at_end=True,\n", | ||
| " save_strategy=\"epoch\",\n", | ||
| " weight_decay=0.01,\n", | ||
| ")\n", | ||
| "\n", | ||
| "trainer = Trainer(\n", | ||
| " model=model,\n", | ||
| " args=training_args,\n", | ||
| " train_dataset=small_train_dataset,\n", | ||
| " eval_dataset=small_eval_dataset,\n", | ||
| " compute_metrics=compute_metrics,\n", | ||
| " callbacks=[DVCLiveCallback(live=Live(report=\"notebook\"), log_model=True)],\n", | ||
| ")\n", | ||
| "trainer.train()" | ||
| ], | ||
| "metadata": { | ||
| "id": "M4FKUYTi5zYQ" | ||
| }, | ||
| "execution_count": null, | ||
| "outputs": [] | ||
| }, | ||
| { | ||
| "cell_type": "markdown", | ||
| "metadata": { | ||
| "id": "l29wqAaDdmsW" | ||
| }, | ||
| "source": [ | ||
| "### Comparing Experiments\n", | ||
| "\n", | ||
| "We create a dataframe with the experiments in order to visualize it.\n" | ||
| ] | ||
| }, | ||
| { | ||
| "cell_type": "code", | ||
| "execution_count": null, | ||
| "metadata": { | ||
| "id": "wwMwHvVtdmsW" | ||
| }, | ||
| "outputs": [], | ||
| "source": [ | ||
| "import dvc.api\n", | ||
| "import pandas as pd\n", | ||
| "\n", | ||
| "columns = [\"Experiment\", \"epoch\", \"eval.f1\"]\n", | ||
| "\n", | ||
| "df = pd.DataFrame(dvc.api.exp_show(), columns=columns)\n", | ||
| "\n", | ||
| "df.dropna(inplace=True)\n", | ||
| "df.reset_index(drop=True, inplace=True)\n", | ||
| "df\n" | ||
| ] | ||
| }, | ||
| { | ||
| "cell_type": "code", | ||
| "execution_count": null, | ||
| "metadata": { | ||
| "id": "TNBGUqoCdmsW" | ||
| }, | ||
| "outputs": [], | ||
| "source": [ | ||
| "!dvc plots diff $(dvc exp list --names-only)" | ||
| ] | ||
| }, | ||
| { | ||
| "cell_type": "code", | ||
| "execution_count": null, | ||
| "metadata": { | ||
| "id": "sL5pH4X5dmsW" | ||
| }, | ||
| "outputs": [], | ||
| "source": [ | ||
| "from IPython.display import HTML\n", | ||
| "HTML(filename='./dvc_plots/index.html')" | ||
| ] | ||
| } | ||
| ], | ||
| "metadata": { | ||
| "colab": { | ||
| "provenance": [] | ||
| }, | ||
| "kernelspec": { | ||
| "display_name": "Python 3 (ipykernel)", | ||
| "language": "python", | ||
| "name": "python3" | ||
| }, | ||
| "language_info": { | ||
| "codemirror_mode": { | ||
| "name": "ipython", | ||
| "version": 3 | ||
| }, | ||
| "file_extension": ".py", | ||
| "mimetype": "text/x-python", | ||
| "name": "python", | ||
| "nbconvert_exporter": "python", | ||
| "pygments_lexer": "ipython3", | ||
| "version": "3.9.13" | ||
| } | ||
| }, | ||
| "nbformat": 4, | ||
| "nbformat_minor": 0 | ||
| } | ||
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
q: is it expected that we use our own callback in this example?