# fastai/course22

**Attribution required: if you use, quote, or summarise this content, you must credit and link back to [awesome-repositories.com](https://awesome-repositories.com/repository/fastai-course22).**

3,398 stars · 1,282 forks · Jupyter Notebook

## Links

- GitHub: https://github.com/fastai/course22
- Homepage: https://course.fast.ai
- awesome-repositories: https://awesome-repositories.com/repository/fastai-course22.md

## Topics

`deep-learning` `fastai` `jupyter-notebooks` `machine-learning` `notebooks` `python`

## Description

This is a structured deep learning curriculum for programmers, delivered as a collection of Jupyter notebooks. It teaches the fundamentals of training neural networks for computer vision, natural language processing, tabular data analysis, and collaborative filtering using PyTorch and the fastai library. The course is designed to be hands-on, guiding learners from building a training loop from scratch to fine-tuning pretrained models for a variety of practical tasks.

The curriculum distinguishes itself by covering the full lifecycle of a deep learning project, from data preparation and augmentation to model deployment and interpretation. It includes dedicated material on medical imaging with DICOM files, generative adversarial networks, and distributed training across multiple GPUs. The course also provides practical guidance on using cloud environments for execution and on sharing models through the Hugging Face Hub.

Beyond training, the course material covers model evaluation with custom metrics, uncertainty estimation through Monte Carlo dropout, and model interpretation through feature attribution and embedding visualization. It also addresses reproducibility with random seed management and offers a structured path for migrating existing PyTorch workflows into the fastai training loop.

## Tags

### Artificial Intelligence & ML

- [Multi-Domain Training](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/model-training-and-tuning/training-frameworks/model-training-frameworks/vision-model-training/multi-domain-training.md) — Trains models for computer vision, NLP, tabular analysis, and collaborative filtering using PyTorch and fastai. ([source](https://course.fast.ai/))
- [Augmentation Pipelines](https://awesome-repositories.com/f/artificial-intelligence-ml/augmentation-pipelines.md) — Creates augmentation pipelines with flips, rotations, zooms, warps, and lighting changes at configurable probabilities. ([source](https://docs.fast.ai/vision.augment.html))
- [Collaborative Filtering Models](https://awesome-repositories.com/f/artificial-intelligence-ml/collaborative-filtering-models.md) — Creates dot-product or neural-network models that predict user-item ratings from a ratings table. ([source](https://docs.fast.ai/collab.html))
- [Segmentation Model Training](https://awesome-repositories.com/f/artificial-intelligence-ml/computer-vision-systems/image-segmentation/segmentation-model-training.md) — Provides a training workflow for pixel-level image segmentation models. ([source](https://docs.fast.ai/quick_start.html))
- [High-Level Loop Replacements](https://awesome-repositories.com/f/artificial-intelligence-ml/custom-training-loops/high-level-loop-replacements.md) — Replaces handwritten training loops with a single high-level call that handles batching, loss, and metrics. ([source](https://docs.fast.ai/examples/migrating_pytorch.html))
- [Data Block APIs](https://awesome-repositories.com/f/artificial-intelligence-ml/data-block-apis.md) — Defines new data blocks with transforms and types for modular ML data pipeline construction. ([source](https://docs.fast.ai/tutorial.siamese.html))
- [Type-Dispatch Pipelines](https://awesome-repositories.com/f/artificial-intelligence-ml/data-block-apis/type-dispatch-pipelines.md) — Uses a new data block API and type dispatch system to express common data processing patterns concisely. ([source](https://docs.fast.ai/))
- [Tabular Metadata Wrappers](https://awesome-repositories.com/f/artificial-intelligence-ml/data-preparation/tabular-metadata-wrappers.md) — Wraps DataFrames with column metadata for automated tabular data preprocessing. ([source](https://docs.fast.ai/tabular.core.html))
- [High-Level Training APIs](https://awesome-repositories.com/f/artificial-intelligence-ml/deep-learning-training-pipelines/high-level-training-apis.md) — Provides high-level components that deliver state-of-the-art results in standard deep learning domains with minimal code. ([source](https://docs.fast.ai/))
- [Classification Training](https://awesome-repositories.com/f/artificial-intelligence-ml/image-classification/classification-training.md) — Provides a high-level training workflow for image classification models with automatic architecture selection and learning rate finding. ([source](https://docs.fast.ai/tutorial.imagenette.html))
- [Generative Model Sampling](https://awesome-repositories.com/f/artificial-intelligence-ml/inference-sampling-strategies/reproducible-sampling/probabilistic-generative-sampling/distribution-based-sampling/generative-model-sampling.md) — Provides techniques for selecting tokens from a language model's output distribution to generate text. ([source](https://docs.fast.ai/text.learner.html))
- [Language Model Builders](https://awesome-repositories.com/f/artificial-intelligence-ml/language-model-architectures/language-model-builders.md) — Teaches building language models from scratch with dropout scaling for regularization. ([source](https://docs.fast.ai/text.models.core.html))
- [AWD-LSTM Language Model Builders](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/frameworks/model-construction/neural-network-layers/recurrent-layers/recurrent-model-definitions/awd-lstm-language-model-builders.md) — Teaches building AWD-LSTM language models with configurable dropout for regularization. ([source](https://docs.fast.ai/text.models.awdlstm.html))
- [Batched DataLoader Wrappers](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/machine-learning-training/dataloader-integrations/batched-dataloader-wrappers.md) — Wraps transformed datasets into DataLoaders that apply transforms before, during, and after batching. ([source](https://docs.fast.ai/data.core.html))
- [Model Persistence](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/model-training-and-tuning/data-and-checkpointing/model-loading/model-persistence.md) — Provides utilities for saving and loading model weights and optimizer states to disk. ([source](https://docs.fast.ai/learner.html))
- [Classification Fine-Tuning](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/model-training-and-tuning/fine-tuning-and-customization/model-fine-tuning/fine-tuned-model-deployment/classification-fine-tuning.md) — Provides a fine-tuning workflow for single-label image classification using pretrained models. ([source](https://docs.fast.ai/tutorial.vision.html))
- [Vision Model Training](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/model-training-and-tuning/training-frameworks/model-training-frameworks/vision-model-training.md) — Provides a framework for building and training vision models from data loaders with optional pretrained weights. ([source](https://docs.fast.ai/tutorial.html))
- [Language Model Training](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/model-fine-tuning-adaptation/language-model-training.md) — Provides a training workflow for language models on tokenized text with loss and perplexity measurement. ([source](https://docs.fast.ai/tutorial.wikitext.html))
- [Discriminative Layer Unfreezing](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/model-fine-tuning-adaptation/learning-rate-schedulers/discriminative-layer-unfreezing.md) — Teaches progressive unfreezing of layer groups with discriminative learning rates for fine-tuning pretrained models.
- [One-Cycle Schedulers](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/model-fine-tuning-adaptation/learning-rate-schedulers/one-cycle-schedulers.md) — Teaches the one-cycle learning rate schedule for faster convergence and better generalization. ([source](https://docs.fast.ai/examples/migrating_pytorch.html))
- [Model Performance Evaluators](https://awesome-repositories.com/f/artificial-intelligence-ml/model-performance-evaluators.md) — Provides tools for quantifying model accuracy by comparing predictions against ground truth labels. ([source](https://docs.fast.ai/learner.html))
- [Model Training Pipelines](https://awesome-repositories.com/f/artificial-intelligence-ml/model-training-pipelines.md) — Offers low-level, composable building blocks and a callback system to modify any part of training. ([source](https://docs.fast.ai/))
- [U-Net Architectures](https://awesome-repositories.com/f/artificial-intelligence-ml/neural-network-architectures/u-net-architectures.md) — Teaches building dynamic U-Net segmentation models that adapt decoders to encoder feature map sizes. ([source](https://docs.fast.ai/vision.models.unet.html))
- [Parameter Freezing](https://awesome-repositories.com/f/artificial-intelligence-ml/parameter-freezing.md) — Teaches parameter freezing and unfreezing for transfer learning and fine-tuning. ([source](https://docs.fast.ai/learner.html))
- [Pooling Classifiers](https://awesome-repositories.com/f/artificial-intelligence-ml/pooling-layers/pooling-classifiers.md) — Teaches pooling-based text classification by feeding pooled features through a linear classifier. ([source](https://docs.fast.ai/text.models.core.html))
- [Prompt-Based Text Generation](https://awesome-repositories.com/f/artificial-intelligence-ml/prompt-based-text-generation.md) — Generates text sequences conditioned on input prompts using a pretrained transformer model. ([source](https://docs.fast.ai/tutorial.transformers.html))
- [Text Classifier Fine-Tuning](https://awesome-repositories.com/f/artificial-intelligence-ml/text-classifiers/text-classifier-fine-tuning.md) — Provides a fine-tuning workflow for text classification using pretrained language models on labeled datasets. ([source](https://docs.fast.ai/tutorial.text.html))
- [Training Callbacks](https://awesome-repositories.com/f/artificial-intelligence-ml/training-callbacks.md) — Provides a configurable callback system that hooks into training loop events like batch start and loss computation.
- [Parameter Group Unfreezers](https://awesome-repositories.com/f/artificial-intelligence-ml/training-configurations/training-parameter-configurations/parameter-group-unfreezers.md) — Teaches parameter group unfreezing to resume training frozen layers during fine-tuning. ([source](https://docs.fast.ai/optimizer.html))
- [Batched DataLoader Wrapping](https://awesome-repositories.com/f/artificial-intelligence-ml/training-data-transformations/batched-dataloader-wrapping.md) — Converts transformed datasets into batched data loaders for model training and inference. ([source](https://docs.fast.ai/tutorial.pets.html))
- [Training Loop Controllers](https://awesome-repositories.com/f/artificial-intelligence-ml/training-loop-controllers.md) — Provides a training loop controller that groups model, data loaders, and loss function. ([source](https://docs.fast.ai/learner.html))
- [Type-Dispatching Pipelines](https://awesome-repositories.com/f/artificial-intelligence-ml/type-based-dispatchers/type-dispatching-pipelines.md) — Implements type-dispatch transform pipelines that route data items to type-specific encode and decode methods.
- [Sentence Embeddings](https://awesome-repositories.com/f/artificial-intelligence-ml/vector-embeddings/sentence-embeddings.md) — Teaches sentence encoding using modules that process padded input in chunked format. ([source](https://docs.fast.ai/text.models.core.html))
- [Normalization Freezers](https://awesome-repositories.com/f/artificial-intelligence-ml/batch-normalization/normalization-freezers.md) — Freezes batch normalization statistics during transfer learning to preserve pretrained feature extractors. ([source](https://docs.fast.ai/callback.training.html))
- [DataFrame-Based Loaders](https://awesome-repositories.com/f/artificial-intelligence-ml/collaborative-filtering-utilities/dataframe-based-loaders.md) — Constructs data loaders for collaborative filtering directly from in-memory ratings tables. ([source](https://docs.fast.ai/collab.html))
- [Lighting Adjusters](https://awesome-repositories.com/f/artificial-intelligence-ml/computer-vision-systems/computer-vision/image-augmentation/lighting-adjusters.md) — Modifies brightness, contrast, saturation, and hue of images to simulate different lighting conditions. ([source](https://docs.fast.ai/vision.augment.html))
- [Center Cropping and Padding](https://awesome-repositories.com/f/artificial-intelligence-ml/convolutional-operations/input-padding-utilities/padding-maskers/spatial-image-padding/center-cropping-and-padding.md) — Center crops or pads images to a specified size for consistent model input dimensions. ([source](https://docs.fast.ai/vision.augment.html))
- [Custom Data Transform Extensions](https://awesome-repositories.com/f/artificial-intelligence-ml/custom-data-transform-extensions.md) — Adds normalization, data augmentation, and other transforms to custom datasets via after_item and after_batch hooks. ([source](https://docs.fast.ai/tutorial.siamese.html))
- [Custom Model Training](https://awesome-repositories.com/f/artificial-intelligence-ml/custom-model-training.md) — Guides experienced users through collecting data and training specialized models like Siamese networks. ([source](https://docs.fast.ai/tutorial.html))
- [Data-Parallel Training](https://awesome-repositories.com/f/artificial-intelligence-ml/distributed-training-frameworks/data-parallel-training.md) — Implements data-parallel training across multiple GPUs by synchronizing gradients and splitting batches using PyTorch's distributed backend.
- [Custom Performance Metrics](https://awesome-repositories.com/f/artificial-intelligence-ml/evaluation-metrics/custom-performance-metrics.md) — Teaches creating custom metrics that accumulate state across batches and compute final values. ([source](https://docs.fast.ai/learner.html))
- [Feature Attribution Methods](https://awesome-repositories.com/f/artificial-intelligence-ml/feature-attribution-methods.md) — Uses attribution methods to explain which input features most influence a model's output. ([source](https://docs.fast.ai/callback.captum.html))
- [Tensor Flattening](https://awesome-repositories.com/f/artificial-intelligence-ml/feature-extraction/convolutional-feature-extractors/feature-map-aggregators/tensor-flattening.md) — Demonstrates flattening multi-dimensional tensors into one-dimensional vectors for linear layers. ([source](https://docs.fast.ai/layers.html))
- [GAN Training Loops](https://awesome-repositories.com/f/artificial-intelligence-ml/gan-training-loops.md) — Alternates training iterations between generator and critic for adversarial training. ([source](https://docs.fast.ai/vision.gan.html))
- [Focal Loss Calculators](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/architectures/neural-network-components/loss-functions/perceptual-loss/content-loss-calculators/focal-loss-calculators.md) — Down-weights easy-to-classify observations in cross-entropy loss to focus training on hard examples. ([source](https://docs.fast.ai/losses.html))
- [Regression Loss Functions](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/architectures/neural-network-components/loss-functions/perceptual-loss/content-loss-calculators/focal-loss-calculators/detection-loss-calculators/smooth-l1-loss-calculators/regression-loss-functions.md) — Applies MSE or L1 loss with automatic tensor flattening for regression tasks. ([source](https://docs.fast.ai/losses.html))
- [Dice Loss Calculators](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/architectures/neural-network-components/loss-functions/perceptual-loss/dice-loss-calculators.md) — Measures overlap between predicted and ground-truth segmentation masks using Dice loss. ([source](https://docs.fast.ai/losses.html))
- [Combined Loss Functions](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/architectures/neural-network-components/loss-functions/perceptual-loss/medical-segmentation-loss-functions/combined-loss-functions.md) — Mixes Dice loss with focal loss to balance global and local feature learning during segmentation training. ([source](https://docs.fast.ai/losses.html))
- [Sequential Layer Containers](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/frameworks/model-construction/neural-network-layers/normalization-layers/sequential-layer-containers.md) — Demonstrates chaining neural network layers sequentially using PyTorch's nn.Sequential container. ([source](https://docs.fast.ai/vision.models.xresnet.html))
- [Sequential Containers](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/frameworks/model-construction/sequential-containers.md) — Shows how to create sequential containers that automatically wrap functions as layers. ([source](https://docs.fast.ai/layers.html))
- [Per-Split Augmentations](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/data-ingestion-preparation/data-augmentation/per-split-augmentations.md) — Runs different augmentation pipelines on training and validation sets by specifying a split index. ([source](https://docs.fast.ai/tutorial.albumentations.html))
- [Mixed Precision Training](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/machine-learning-training/distributed-and-accelerated-compute/training-acceleration-tools/mixed-precision-training.md) — Teaches mixed precision training to accelerate training and reduce memory usage. ([source](https://docs.fast.ai/callback.fp16.html))
- [Custom Loss Functions](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/machine-learning-training/pipelines-and-orchestration/training-orchestration-systems/training-lifecycle-management/custom-loss-functions.md) — Implements custom loss functions with activation and decoding methods for integration with prediction tools. ([source](https://docs.fast.ai/tutorial.imagenette.html))
- [Encoder Persisters](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/model-training-and-tuning/data-and-checkpointing/model-loading/model-persistence/encoder-persisters.md) — Persists the encoder portion of a text model to disk and restores it for reuse or transfer learning. ([source](https://docs.fast.ai/text.learner.html))
- [Multi-Label](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/model-training-and-tuning/fine-tuning-and-customization/model-fine-tuning/fine-tuned-model-deployment/classification-fine-tuning/multi-label.md) — Ships a fine-tuning workflow for multi-label image classification tasks using pretrained models. ([source](https://docs.fast.ai/tutorial.vision.html))
- [Coordinate Prediction Fine-Tuning](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/model-training-and-tuning/fine-tuning-and-customization/model-fine-tuning/fine-tuned-model-deployment/regression-fine-tuning/coordinate-prediction-fine-tuning.md) — Provides a fine-tuning workflow for predicting spatial coordinates from images with consistent augmentation. ([source](https://docs.fast.ai/tutorial.vision.html))
- [Classification Interpreters](https://awesome-repositories.com/f/artificial-intelligence-ml/model-interpretability/classification-interpreters.md) — Explores predictions from classification models to identify errors and understand model behavior. ([source](https://docs.fast.ai/interpret.html))
- [Segmentation Interpreters](https://awesome-repositories.com/f/artificial-intelligence-ml/model-interpretability/segmentation-interpreters.md) — Explores predictions from segmentation models to identify errors and understand model behavior. ([source](https://docs.fast.ai/interpret.html))
- [Batch Prediction Analyzers](https://awesome-repositories.com/f/artificial-intelligence-ml/model-prediction-evaluation/batch-prediction-analyzers.md) — Ships batch processing utilities that generate predictions and losses for interpreting model behavior. ([source](https://docs.fast.ai/interpret.html))
- [Dataset Prediction Generators](https://awesome-repositories.com/f/artificial-intelligence-ml/model-predictions/dataset-prediction-generators.md) — Computes predictions and targets for entire datasets with options for decoding and saving to disk. ([source](https://docs.fast.ai/learner.html))
- [Single Item Predictors](https://awesome-repositories.com/f/artificial-intelligence-ml/model-predictions/single-item-predictors.md) — Runs single items through the model and returns decoded predictions with probabilities. ([source](https://docs.fast.ai/learner.html))
- [Optimizer Refactors](https://awesome-repositories.com/f/artificial-intelligence-ml/model-training-optimizers/optimizer-refactors.md) — Refactors optimization algorithms into composable pieces for writing new optimizers concisely. ([source](https://docs.fast.ai/))
- [Adversarial Training Procedures](https://awesome-repositories.com/f/artificial-intelligence-ml/model-training/adversarial-training-procedures.md) — Wraps generator and critic together, switching between them during adversarial training. ([source](https://docs.fast.ai/vision.gan.html))
- [Learner Exports](https://awesome-repositories.com/f/artificial-intelligence-ml/model-training/model-exporting/production-inference-exports/learner-exports.md) — Saves trained learner objects without data or optimizer state for deployment. ([source](https://docs.fast.ai/learner.html))
- [Weights and Biases Training Logging](https://awesome-repositories.com/f/artificial-intelligence-ml/model-weight-management/dynamic-weight-updates/training-weight-adjustments/weights-and-biases-training-logging.md) — Logs model topology, losses, metrics, gradients, and sample predictions to Weights & Biases during training. ([source](https://docs.fast.ai/callback.wandb.html))
- [Multi-GPU Training Utilities](https://awesome-repositories.com/f/artificial-intelligence-ml/multi-gpu-training-utilities.md) — Distributes training across multiple GPUs using PyTorch's distributed backend and Accelerate. ([source](https://docs.fast.ai/distributed.html))
- [Pretrained Model Loading](https://awesome-repositories.com/f/artificial-intelligence-ml/natural-language-processing/language-model-pretraining/pretrained-model-loading.md) — Provides utilities for loading pretrained language models and adapting their vocabularies for downstream tasks. ([source](https://docs.fast.ai/text.learner.html))
- [DataFrame Column Tokenization](https://awesome-repositories.com/f/artificial-intelligence-ml/natural-language-processing/text-tokenization/dataframe-column-tokenization.md) — Tokenizes specified text columns of a DataFrame in parallel for NLP preprocessing. ([source](https://docs.fast.ai/text.core.html))
- [Convolutional Network Builders](https://awesome-repositories.com/f/artificial-intelligence-ml/neural-network-construction/convolutional-network-builders.md) — Teaches building convolutional neural networks from filter size lists with configurable strides. ([source](https://docs.fast.ai/layers.html))
- [Adam Optimizers](https://awesome-repositories.com/f/artificial-intelligence-ml/optimization-algorithms/adam-optimizers.md) — Implements Adam and AdamW optimizers as part of the training loop for deep learning models. ([source](https://docs.fast.ai/optimizer.html))
- [Prediction Visualization](https://awesome-repositories.com/f/artificial-intelligence-ml/prediction-visualization.md) — Provides tools for displaying and filtering model prediction results alongside inputs for visual inspection. ([source](https://docs.fast.ai/learner.html))
- [Binary Cross-Entropy Calculators](https://awesome-repositories.com/f/artificial-intelligence-ml/prediction-visualization/loss-function-calculators/binary-cross-entropy-calculators.md) — Applies binary cross-entropy loss with automatic tensor flattening for binary classification tasks. ([source](https://docs.fast.ai/losses.html))
- [Medical Imaging Training Frameworks](https://awesome-repositories.com/f/artificial-intelligence-ml/pytorch-training-frameworks/medical-imaging-training-frameworks.md) — Provides a training workflow for medical image classifiers using DICOM files to detect conditions like pneumothorax. ([source](https://docs.fast.ai/tutorial.medical_imaging.html))
- [Self-Attention Mechanisms](https://awesome-repositories.com/f/artificial-intelligence-ml/self-attention-mechanisms.md) — Teaches adding self-attention layers to convolutional networks for capturing long-range spatial dependencies. ([source](https://docs.fast.ai/layers.html))
- [Temporal Sequence Processors](https://awesome-repositories.com/f/artificial-intelligence-ml/sequence-modeling/temporal-sequence-processors.md) — Teaches applying the same module to each temporal slice of a sequence for consistent processing. ([source](https://docs.fast.ai/layers.html))
- [Test Time Augmentation](https://awesome-repositories.com/f/artificial-intelligence-ml/test-time-augmentation.md) — Applies data transformations during inference and averages predictions to improve accuracy. ([source](https://docs.fast.ai/learner.html))
- [Text Classifier Construction](https://awesome-repositories.com/f/artificial-intelligence-ml/text-classifiers/text-classifier-construction.md) — Provides a workflow for constructing text classification models from pretrained encoders with sentence-level processing. ([source](https://docs.fast.ai/text.learner.html))
- [CSV](https://awesome-repositories.com/f/artificial-intelligence-ml/text-tokenizers/csv.md) — Tokenizes text columns of CSV files in parallel, writing tokenized results with added length columns. ([source](https://docs.fast.ai/text.core.html))
- [Best Checkpoint Selectors](https://awesome-repositories.com/f/artificial-intelligence-ml/training-checkpointers/best-checkpoint-selectors.md) — Saves the best-performing model checkpoint based on validation metrics during training. ([source](https://docs.fast.ai/callback.tracker.html))
- [Gradient Accumulation Strategies](https://awesome-repositories.com/f/artificial-intelligence-ml/training-convergence-optimization/batch-size-scaling/gradient-accumulation-strategies.md) — Accumulates gradients across several batches before updating weights for larger effective batch sizes. ([source](https://docs.fast.ai/callback.training.html))
- [Pretrained Weight Initializers](https://awesome-repositories.com/f/artificial-intelligence-ml/weight-initialization/pretrained-weight-initializers.md) — Downloads and loads pretrained model weights for language models and transformers from remote sources. ([source](https://docs.fast.ai/data.external.html))

### Education & Learning Resources

- [Deep Learning Education](https://awesome-repositories.com/f/education-learning-resources/deep-learning-education.md) — Teaches deep learning fundamentals through hands-on Jupyter notebooks for programmers.
- [Deep Learning Courses](https://awesome-repositories.com/f/education-learning-resources/deep-learning-courses.md) — A structured curriculum of Jupyter notebooks teaching deep learning with PyTorch and fastai.
- [Deep Learning Curriculum](https://awesome-repositories.com/f/education-learning-resources/deep-learning-curriculum.md) — Walks through a complete course of Jupyter notebooks, slides, and spreadsheets for deep learning. ([source](https://cdn.jsdelivr.net/gh/fastai/course22@master/README.md))
- [Sentiment Classification Fine-Tuning](https://awesome-repositories.com/f/education-learning-resources/educational-resources/systems-applied-computing/machine-learning-education/sentiment-analysis-models/sentiment-classifiers/sentiment-classification-fine-tuning.md) — Provides a fine-tuning workflow for sentiment classification using pretrained language models with transfer learning. ([source](https://docs.fast.ai/quick_start.html))

### Part of an Awesome List

- [Vision Model Fine-Tuning](https://awesome-repositories.com/f/awesome-lists/ai/model-training-and-fine-tuning/vision-model-fine-tuning.md) — Fine-tunes pretrained models to recognize objects or breeds in custom image datasets. ([source](https://docs.fast.ai/quick_start.html))
- [Learner Restorers](https://awesome-repositories.com/f/awesome-lists/ai/model-training-and-fine-tuning/checkpoint-saving-and-restoration/learner-restorers.md) — Restores exported learner objects from disk for inference using pickle. ([source](https://docs.fast.ai/learner.html))
- [Domain-Specific Fine-Tuning](https://awesome-repositories.com/f/awesome-lists/ai/model-training-and-fine-tuning/domain-specific-fine-tuning.md) — Provides a workflow for adapting pretrained language models to specific text corpora through continued pretraining. ([source](https://docs.fast.ai/tutorial.text.html))
- [Pixel-Wise Segmentation Fine-Tuning](https://awesome-repositories.com/f/awesome-lists/ai/model-training-and-fine-tuning/model-fine-tuning/pixel-wise-segmentation-fine-tuning.md) — Provides a fine-tuning workflow for pixel-wise image segmentation using pretrained models. ([source](https://docs.fast.ai/tutorial.vision.html))
- [Transformer Fine-Tuning](https://awesome-repositories.com/f/awesome-lists/ai/model-training-and-fine-tuning/pretrained-checkpoint-fine-tuning/transformer-fine-tuning.md) — Provides a fine-tuning workflow for HuggingFace transformer models using fastai's learning rate finder and 1cycle policy. ([source](https://docs.fast.ai/tutorial.transformers.html))
- [Transformer Training Walkthroughs](https://awesome-repositories.com/f/awesome-lists/ai/model-training-and-fine-tuning/transformer-training-walkthroughs.md) — Provides step-by-step walkthroughs for building and fine-tuning NLP models using a high-level API. ([source](https://docs.fast.ai/tutorial.html))
- [Optimizer Composers](https://awesome-repositories.com/f/awesome-lists/ai/robotics-foundation-models/training-optimizations/optimizer-composers.md) — Teaches a compositional optimizer design that lets new algorithms be written in a few lines. ([source](https://docs.fast.ai/.md))

### Data & Databases

- [Declarative Pipeline Blueprints](https://awesome-repositories.com/f/data-databases/data-pipeline-transformers/declarative-pipeline-blueprints.md) — Provides a declarative API for assembling complete data processing pipelines from scratch. ([source](https://docs.fast.ai/tutorial.datablock.html))
- [Declarative Pipeline Construction](https://awesome-repositories.com/f/data-databases/data-processing-pipelines/data-processing-frameworks/declarative-pipeline-construction.md) — Constructs custom data processing pipelines using a declarative block API. ([source](https://docs.fast.ai/tutorial.imagenette.html))
- [Data Block Blueprints](https://awesome-repositories.com/f/data-databases/data-processing-pipelines/data-processing-frameworks/declarative-pipeline-construction/data-block-blueprints.md) — Ships a declarative API for constructing data processing pipelines by specifying types, sources, splits, and transforms in a blueprint.
- [ML Data Pipeline Declarations](https://awesome-repositories.com/f/data-databases/declarative-data-reconciliation-pipelines/ml-data-pipeline-declarations.md) — Defines types, sources, splits, and transforms to produce Datasets and DataLoaders declaratively. ([source](https://docs.fast.ai/data.block.html))
- [CSV Data Loaders](https://awesome-repositories.com/f/data-databases/tabular-data-frameworks/csv-data-loaders.md) — Creates data loaders from CSV files, applying preprocessing and splitting into training and validation sets. ([source](https://docs.fast.ai/tabular.data.html))
- [Tabular Predictive Models](https://awesome-repositories.com/f/data-databases/tabular-data-frameworks/tabular-predictive-models.md) — Provides a workflow for building predictive models on structured tabular data to forecast target values. ([source](https://docs.fast.ai/quick_start.html))
- [Tabular Data Loaders](https://awesome-repositories.com/f/data-databases/tabular-data-frameworks/tabular-predictive-models/tabular-data-loaders.md) — Loads and preprocesses tabular data with encoding, imputation, and normalization for model training. ([source](https://docs.fast.ai/tutorial.datablock.html))
- [Configurable Preprocessor Pipelines](https://awesome-repositories.com/f/data-databases/tabular-data-loaders/configurable-preprocessor-pipelines.md) — Transforms raw tabular data through configurable pipelines of categorification, imputation, and normalization. ([source](https://docs.fast.ai/tutorial.tabular.html))
- [Categorical Encoders](https://awesome-repositories.com/f/data-databases/data-categorization/categorical-encoders.md) — Converts categorical columns into numeric representations, storing unique values per column for ML models. ([source](https://docs.fast.ai/tabular.core.html))
- [Label Extraction Functions](https://awesome-repositories.com/f/data-databases/data-loading-optimizations/custom-dataset-loading/label-extraction-functions.md) — Ships utilities for extracting labels from image filenames using custom functions or regex. ([source](https://docs.fast.ai/vision.data.html))
- [Custom Data Pipeline Guidance](https://awesome-repositories.com/f/data-databases/data-pipelines/custom-data-pipeline-guidance.md) — Guides users through constructing custom data pipelines for vision and text tasks using a mid-level API. ([source](https://docs.fast.ai/tutorial.html))
- [Label Reading from Columns](https://awesome-repositories.com/f/data-databases/dataframe-integration/label-reading-from-columns.md) — Reads column values from DataFrame rows as labels for supervised learning tasks. ([source](https://docs.fast.ai/data.transforms.html))
- [Transform Wrapping for Datasets](https://awesome-repositories.com/f/data-databases/dataset-transformations/transform-wrapping-for-datasets.md) — Wraps custom dataset logic into Transform objects so they integrate with the data pipeline system. ([source](https://docs.fast.ai/tutorial.siamese.html))
- [Label Smoothing Techniques](https://awesome-repositories.com/f/data-databases/label-based-data-selection/metadata-labelers/label-smoothing-utilities/label-smoothing-techniques.md) — Replaces hard one-hot targets with smoothed labels in cross-entropy loss to reduce overfitting. ([source](https://docs.fast.ai/losses.html))
- [Single Row Predictors](https://awesome-repositories.com/f/data-databases/tabular-data-frameworks/tabular-predictive-models/single-row-predictors.md) — Returns predictions for single rows of tabular data along with the original input values. ([source](https://docs.fast.ai/tabular.learner.html))
- [Tabular Data Preprocessing](https://awesome-repositories.com/f/data-databases/tabular-data-frameworks/tabular-predictive-models/tabular-data-preprocessing.md) — Provides data preprocessing pipelines for tabular data including categorical encoding, missing value imputation, and normalization. ([source](https://docs.fast.ai/tutorial.html))

### Development Tools & Productivity

- [Data Transformation Pipelines](https://awesome-repositories.com/f/development-tools-productivity/data-transformation-pipelines.md) — Applies pipelines of transforms to collections of items, producing transformed datasets for model training. ([source](https://docs.fast.ai/data.core.html))
- [Reusable Transform Pipelines](https://awesome-repositories.com/f/development-tools-productivity/data-transformation-pipelines/reusable-transform-pipelines.md) — Provides reusable data processing pipelines with encode, decode, and display capabilities for ML workflows. ([source](https://docs.fast.ai/tutorial.pets.html))
- [Sequential Pipelines](https://awesome-repositories.com/f/development-tools-productivity/task-pipeline-managers/sequential-pipelines.md) — Combines multiple data transformations into sequential pipelines that process raw items end-to-end. ([source](https://docs.fast.ai/tutorial.pets.html))

### Graphics & Multimedia

- [Labeled](https://awesome-repositories.com/f/graphics-multimedia/image-file-loading/labeled.md) — Reads image files from folders or dataframes and pairs them with category labels for model training. ([source](https://docs.fast.ai/tutorial.datablock.html))
- [Labeled Image Loading from DataFrames](https://awesome-repositories.com/f/graphics-multimedia/image-file-loading/labeled-image-loading-from-dataframes.md) — Loads images and their labels from DataFrames or CSV files for training. ([source](https://docs.fast.ai/vision.data.html))
- [DICOM Visualization Tools](https://awesome-repositories.com/f/graphics-multimedia/dicom-visualization-tools.md) — Displays X-ray images from DICOM pixel arrays using standard plotting functions. ([source](https://docs.fast.ai/tutorial.medical_imaging.html))
- [DICOM Parsing](https://awesome-repositories.com/f/graphics-multimedia/dicom-visualization-tools/dicom-parsing.md) — Reads medical DICOM files, extracting pixel data and metadata tags for analysis. ([source](https://docs.fast.ai/tutorial.medical_imaging.html))
- [Affine Transformation Engines](https://awesome-repositories.com/f/graphics-multimedia/graphics-engines-rendering/rendering/coordinate-viewport-transformations/matrix-transformation-engines/affine-transformation-engines.md) — Applies affine transformations like rotation, zoom, and warp to images and associated masks or bounding boxes. ([source](https://docs.fast.ai/vision.augment.html))
- [Random Crop and Flip Augmentations](https://awesome-repositories.com/f/graphics-multimedia/image-editing-processing/image-editors/image-cropping-tools/random-crop-and-flip-augmentations.md) — Randomly flips images horizontally or applies dihedral transformations with configurable probability during training. ([source](https://docs.fast.ai/vision.augment.html))
- [Multi-Strategy Resizing](https://awesome-repositories.com/f/graphics-multimedia/image-editing-processing/image-editors/image-cropping-tools/random-crop-and-flip-augmentations/multi-strategy-resizing.md) — Resizes images using squish, pad, or crop methods with different strategies for training and validation sets. ([source](https://docs.fast.ai/vision.augment.html))
- [Dimension Resizing](https://awesome-repositories.com/f/graphics-multimedia/image-editing-processing/image-processing/dimension-resizing.md) — Changes pixel dimensions of images to a specified size for model input consistency. ([source](https://docs.fast.ai/medical.imaging.html))
- [DICOM File Loading](https://awesome-repositories.com/f/graphics-multimedia/image-file-loading/dicom-file-loading.md) — Recursively finds and loads DICOM medical image files from a specified path. ([source](https://docs.fast.ai/medical.imaging.html))
- [Training](https://awesome-repositories.com/f/graphics-multimedia/media-processing-analysis/face-portrait-manipulation/image-blending/training.md) — Blends pairs of training images with configurable ratios to create augmented samples for better generalization. ([source](https://docs.fast.ai/callback.mixup.html))

### Scientific & Mathematical Computing

- [Neural Network Training from Scratch](https://awesome-repositories.com/f/scientific-mathematical-computing/from-scratch-implementations/neural-network-training-from-scratch.md) — Teaches building SGD and a complete training loop from scratch to understand deep learning foundations. ([source](https://course.fast.ai/))
- [Optimizer Stepper Composition](https://awesome-repositories.com/f/scientific-mathematical-computing/monte-carlo-sampling/game-tree-search/state-steppers/optimizer-stepper-composition.md) — Provides a modular system for composing custom optimizers from small stepper functions that define individual update rules.

### User Interface & Experience

- [Collaborative Filtering Systems](https://awesome-repositories.com/f/user-interface-experience/form-builders/builder-item-collapsers/builder-item-managers/list-item-markers/item-to-item-similarity/collaborative-filtering-systems.md) — Trains collaborative filtering models to predict user preferences for items like movies. ([source](https://docs.fast.ai/quick_start.html))
- [Augmentation-Aware Mask Handlers](https://awesome-repositories.com/f/user-interface-experience/animation-and-motion-systems/image-masking/segmentation-mask-handlers/augmentation-aware-mask-handlers.md) — Applies identical spatial augmentations to images and segmentation masks for consistent training data. ([source](https://docs.fast.ai/tutorial.albumentations.html))

### Web Development

- [Tabular Data Loaders](https://awesome-repositories.com/f/web-development/web-infrastructure-deployment/asset-management-build-tools/asset-loading-resolution/asset-loaders/data-file-loaders/tabular-data-loaders.md) — Creates data loaders from pandas DataFrames for tabular model training. ([source](https://docs.fast.ai/tabular.data.html))

### Networking & Communication

- [Residual Shortcut Mergers](https://awesome-repositories.com/f/networking-communication/connection-shortcut-generators/residual-shortcut-mergers.md) — Teaches merging residual shortcut connections with main path outputs in convolutional networks. ([source](https://docs.fast.ai/layers.html))

### System Administration & Monitoring

- [Training Metrics](https://awesome-repositories.com/f/system-administration-monitoring/logging/training-metrics.md) — Logs learning rate, loss, and metrics during training and provides loss curve plotting. ([source](https://docs.fast.ai/learner.html))