# open-mmlab/mmagic

**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/open-mmlab-mmagic).**

7,434 stars · 1,099 forks · Jupyter Notebook · Apache-2.0

## Links

- GitHub: https://github.com/open-mmlab/mmagic
- Homepage: https://mmagic.readthedocs.io/en/latest/
- awesome-repositories: https://awesome-repositories.com/repository/open-mmlab-mmagic.md

## Topics

`aigc` `computer-vision` `deep-learning` `diffusion` `diffusion-models` `generative-adversarial-network` `generative-ai` `image-editing` `image-generation` `image-processing` `image-synthesis` `inpainting` `matting` `pytorch` `super-resolution` `text2image` `video-frame-interpolation` `video-interpolation` `video-super-resolution`

## Description

mmagic is a multimodal training pipeline and framework for generative AI, focusing on visual synthesis and restoration. It provides the infrastructure to build and train models for tasks such as text-to-image and text-to-video generation, 3D-aware content synthesis, and high-fidelity image translation using diffusion models and generative adversarial networks.

The project distinguishes itself through specialized capabilities for generative model personalization, including techniques for fine-tuning subjects and styles. It also supports advanced visual manipulations such as latent space interpolation, point-based image editing, and stable animation generation to reduce flickering in video sequences.

The framework covers a broad range of image processing and restoration domains, including denoising, super-resolution, inpainting, and foreground matting. Its architectural surface includes a registry-based component system for custom model and loss definitions, as well as comprehensive data pipelines for multimodal asset loading and augmentation.

Training and operationalization are supported through distributed-data-parallel execution, mixed-precision optimization, and inference backend conversion for hardware accelerators.

## Tags

### Graphics & Multimedia

- [Image and Video Restoration Suites](https://awesome-repositories.com/f/graphics-multimedia/image-and-video-restoration-suites.md) — Implements a comprehensive framework for improving visual fidelity through upscaling, denoising, and deblurring.
- [Image Denoising](https://awesome-repositories.com/f/graphics-multimedia/image-denoising.md) — Removes noise and artifacts from both grayscale and color images to improve visual quality. ([source](https://mmagic.readthedocs.io/en/latest/model_zoo/image_denoising.html))
- [Dataset Composition](https://awesome-repositories.com/f/graphics-multimedia/image-editing-processing/image-processing/image-sequence-processors/animation-frame-sequencers/generative-animation-sequences/image-to-video-animators/image-composition/dataset-composition.md) — Implements workflows to merge foreground objects with backgrounds and generate annotation files for matting datasets. ([source](https://mmagic.readthedocs.io/en/latest/dataset_zoo/comp1k.html))
- [Restoration Quality Metrics](https://awesome-repositories.com/f/graphics-multimedia/restoration-quality-metrics.md) — Calculates quantitative fidelity metrics including PSNR, SSIM, and NIQE to evaluate image and video restoration quality. ([source](https://mmagic.readthedocs.io/en/latest/user_guides/metrics.html))
- [AI Foreground Isolation](https://awesome-repositories.com/f/graphics-multimedia/video-post-production-effects/ai-foreground-isolation.md) — Separates foreground objects from backgrounds by estimating precise alpha mattes for transparency. ([source](https://mmagic.readthedocs.io/en/latest/model_zoo/matting.html))
- [Video Restoration Tools](https://awesome-repositories.com/f/graphics-multimedia/video-restoration-tools.md) — Provides specialized directory organization for low-quality video frames to train restoration models. ([source](https://mmagic.readthedocs.io/en/latest/dataset_zoo/videolq.html))
- [AI Image Masking](https://awesome-repositories.com/f/graphics-multimedia/ai-image-masking.md) — Creates bounding boxes and irregular masks to isolate specific image regions for processing. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/utils/index.html))
- [Temporal Frame Interpolation](https://awesome-repositories.com/f/graphics-multimedia/frame-buffer-snapshots/sequential-frame-buffers/temporal-frame-interpolation.md) — Generates intermediate frames between existing ones to increase video playback smoothness and frame rate. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/models/editors/index.html))
- [Temporal Consistency Optimization](https://awesome-repositories.com/f/graphics-multimedia/frame-by-frame-stream-processing/frame-by-frame-generative-synthesis/temporal-consistency-optimization.md) — Produces consistent video by applying multi-frame rendering with diffusion models to reduce flickering. ([source](https://mmagic.readthedocs.io/en/latest/model_zoo/controlnet_animation.html))
- [Image Enhancement Tools](https://awesome-repositories.com/f/graphics-multimedia/image-editing-processing/image-enhancement-tools.md) — Improves image clarity and detail by removing artifacts and applying adaptive lookup tables. ([source](https://mmagic.readthedocs.io/en/latest/community/projects.html))
- [Image Noise Reduction](https://awesome-repositories.com/f/graphics-multimedia/image-noise-reduction.md) — Cleans digital images by eliminating Gaussian noise to improve overall visual clarity. ([source](https://mmagic.readthedocs.io/en/latest/model_zoo/deraining.html))
- [Degradation Simulation Pipelines](https://awesome-repositories.com/f/graphics-multimedia/image-noise-simulations/degradation-simulation-pipelines.md) — Simulates real-world image quality loss by applying random blur and compression to create restoration training pairs. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/datasets/transforms/index.html))
- [Pixel Value Normalization](https://awesome-repositories.com/f/graphics-multimedia/image-to-tensor-conversions/pixel-value-normalization.md) — Standardizes image inputs by adjusting pixel values using mean and standard deviation constants. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/datasets/transforms/index.html))
- [JPEG Artifact Reduction](https://awesome-repositories.com/f/graphics-multimedia/jpeg-artifact-reduction.md) — Eliminates visual distortions caused by JPEG compression in grayscale and color images. ([source](https://mmagic.readthedocs.io/en/latest/model_zoo/image_denoising.html))
- [Alpha Matting](https://awesome-repositories.com/f/graphics-multimedia/media-processing-analysis/face-portrait-manipulation/image-masking/binary-mask-extraction/alpha-matting.md) — Provides alpha matting capabilities to separate foreground objects from backgrounds with smooth transparency transitions. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/models/editors/index.html))
- [Matting Accuracy Metrics](https://awesome-repositories.com/f/graphics-multimedia/media-processing-analysis/face-portrait-manipulation/image-masking/binary-mask-extraction/alpha-matting/matting-accuracy-metrics.md) — Measures alpha matte prediction accuracy using specialized metrics such as SAD and MattingMSE. ([source](https://mmagic.readthedocs.io/en/latest/user_guides/metrics.html))
- [High-Fidelity Synthesis](https://awesome-repositories.com/f/graphics-multimedia/media-processing-analysis/media-manipulation/image-processing/libraries/high-performance-image/high-fidelity-synthesis.md) — Generates natural, high-resolution images from conditional inputs using generative adversarial networks. ([source](https://mmagic.readthedocs.io/en/latest/model_zoo/conditional_gans.html))
- [Rain Streak Removal](https://awesome-repositories.com/f/graphics-multimedia/rain-streak-removal.md) — Eliminates rain streaks from images to restore original clarity and visibility. ([source](https://mmagic.readthedocs.io/en/latest/model_zoo/deraining.html))
- [Real-World Noise Suppression](https://awesome-repositories.com/f/graphics-multimedia/real-world-noise-suppression.md) — Suppresses complex noise found in authentic digital photographs to improve image quality. ([source](https://mmagic.readthedocs.io/en/latest/model_zoo/deraining.html))
- [Video Frame Loading](https://awesome-repositories.com/f/graphics-multimedia/video-frame-loading.md) — Loads sequences of video frames and annotations for super-resolution and frame interpolation tasks. ([source](https://mmagic.readthedocs.io/en/latest/howto/dataset.html))
- [Video Upscaling Pipelines](https://awesome-repositories.com/f/graphics-multimedia/video-upscaling-pipelines.md) — Increases video resolution by reconstructing high-frequency details using temporal alignment. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/models/editors/index.html))

### Artificial Intelligence & ML

- [Data Processing Pipelines](https://awesome-repositories.com/f/artificial-intelligence-ml/data-processing-pipelines.md) — Provides composable data processing pipelines for loading, rescaling, and augmenting multimodal assets for machine learning.
- [Image Inpainting](https://awesome-repositories.com/f/artificial-intelligence-ml/deep-learning-architectures/image-inpainting.md) — Restores corrupted image regions by synthesizing new pixels that blend with the surrounding context. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/apis/inferencers/index.html))
- [Data-Parallel Training](https://awesome-repositories.com/f/artificial-intelligence-ml/distributed-training-frameworks/data-parallel-training.md) — Implements distributed-data-parallel training to scale generative model workloads across multiple GPUs and compute nodes.
- [Distributed Training](https://awesome-repositories.com/f/artificial-intelligence-ml/distributed-training-frameworks/distributed-training.md) — Implements distributed data parallel patterns to execute training for dynamic architectures across multiple compute nodes. ([source](https://cdn.jsdelivr.net/gh/open-mmlab/mmagic@main/README.md))
- [Resolution Upscalers](https://awesome-repositories.com/f/artificial-intelligence-ml/example-based-image-generation/resolution-upscalers.md) — Increases the resolution of images by synthesizing high-frequency details for classical and real-world scenarios. ([source](https://mmagic.readthedocs.io/en/latest/model_zoo/image_denoising.html))
- [Text-to-Image Generators](https://awesome-repositories.com/f/artificial-intelligence-ml/generative-ai-resources/diffusion-visual-models/generative-ai-pipelines/text-to-image-generators.md) — Generates high-resolution imagery from natural language text descriptions using diffusion-based models. ([source](https://cdn.jsdelivr.net/gh/open-mmlab/mmagic@main/README.md))
- [Text-to-Video Generators](https://awesome-repositories.com/f/artificial-intelligence-ml/generative-ai-resources/diffusion-visual-models/generative-ai-pipelines/text-to-video-generators.md) — Synthesizes animated video sequences from text prompts using personalized diffusion models. ([source](https://mmagic.readthedocs.io/en/latest/model_zoo/text2video.html))
- [Generative Model Fine-Tuning](https://awesome-repositories.com/f/artificial-intelligence-ml/generative-model-fine-tuning.md) — Provides a modular framework for fine-tuning generative models using techniques like DreamBooth and Textual Inversion. ([source](https://mmagic.readthedocs.io/en/latest/model_zoo/text2image.html))
- [Generative Model Training Tools](https://awesome-repositories.com/f/artificial-intelligence-ml/generative-model-training-tools.md) — Provides a modular framework for training and fine-tuning generative architectures on custom datasets. ([source](https://mmagic.readthedocs.io/en/latest/get_started/overview.html))
- [Multi-Node Training Scaling](https://awesome-repositories.com/f/artificial-intelligence-ml/gpu-model-deployments/multi-node-training-scaling.md) — Supports scaling distributed training across multiple compute nodes and GPUs using IP-based communication or Slurm. ([source](https://mmagic.readthedocs.io/en/latest/user_guides/train_test.html))
- [Image Data Preprocessing](https://awesome-repositories.com/f/artificial-intelligence-ml/image-data-preprocessing.md) — Provides essential preprocessing utilities to normalize pixel values and apply padding to image tensors for model inputs. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/models/data_preprocessors/index.html))
- [Image Editing](https://awesome-repositories.com/f/artificial-intelligence-ml/image-generation/image-editing.md) — Provides tools for visual content modification, including inpainting, colorization, and foreground matting.
- [Generative Objective 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/generative-objective-functions.md) — Computes adversarial, perceptual, and pixel-wise objective functions to optimize generative AI models. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/models/losses/index.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) — Includes a mixed-precision training wrapper that optimizes memory and speed by using half-precision tensors and gradient scaling.
- [Distributed Training](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/machine-learning-training/distributed-training.md) — Distributes complex generative model training workloads across multiple GPUs or compute nodes. ([source](https://mmagic.readthedocs.io/en/latest/migration/distributed_train.html))
- [Model Training and Inference Engines](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/model-management/model-training-and-inference-engines.md) — Provides a unified implementation for forward passes, loss calculations, and training/validation/testing cycles. ([source](https://mmagic.readthedocs.io/en/latest/howto/models.html))
- [Multimodal Data Preprocessing Utilities](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/model-training-and-tuning/data-and-checkpointing/data-preprocessing/multimodal-data-preprocessing-utilities.md) — Provides utilities to transform and batch raw multimodal datasets before moving them to target hardware. ([source](https://mmagic.readthedocs.io/en/latest/advanced_guides/data_flow.html))
- [Model Architectures](https://awesome-repositories.com/f/artificial-intelligence-ml/model-architectures.md) — Creates custom neural network structures by inheriting from base module classes and registering them. ([source](https://mmagic.readthedocs.io/en/latest/howto/models.html))
- [Multimodal Content Generation](https://awesome-repositories.com/f/artificial-intelligence-ml/multimodal-content-generation.md) — Synthesizes high-resolution visual content from mixed-modality prompts using diffusion models and GANs. ([source](https://mmagic.readthedocs.io/en/latest/get_started/overview.html))
- [Multimodal Training](https://awesome-repositories.com/f/artificial-intelligence-ml/multimodal-training.md) — Builds data processing workflows and distributed training pipelines specifically for multimodal generative AI.
- [Multimodal Tensor Formatting](https://awesome-repositories.com/f/artificial-intelligence-ml/text-dataset-preparation/input-tensor-formatting/multimodal-tensor-formatting.md) — Converts processed data dictionaries into packed tensors for efficient model forward method execution. ([source](https://mmagic.readthedocs.io/en/latest/howto/transforms.html))
- [Mask and Trimap Synthesis](https://awesome-repositories.com/f/artificial-intelligence-ml/computer-vision-systems/image-segmentation/object-mask-generators/point-based-mask-generators/binary-mask-generators/mask-and-trimap-synthesis.md) — Generates binary or soft masks and trimaps from alpha mattes for foreground matting tasks. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/datasets/transforms/index.html))
- [Image Blur Removal](https://awesome-repositories.com/f/artificial-intelligence-ml/convolutional-neural-networks/image-restorers/image-blur-removal.md) — Fixes blurring caused by out-of-focus elements to sharpen edges and restore visual structures. ([source](https://mmagic.readthedocs.io/en/latest/model_zoo/deraining.html))
- [Execution Hooks](https://awesome-repositories.com/f/artificial-intelligence-ml/custom-training-loops/execution-hooks.md) — Allows attaching custom operations to training loops for managing tasks like checkpointing and logging. ([source](https://mmagic.readthedocs.io/en/latest/user_guides/config.html))
- [Preprocessing Optimizations](https://awesome-repositories.com/f/artificial-intelligence-ml/dataset-management/evaluation-datasets/dataset-curation/video-dataset-curators/preprocessing-optimizations.md) — Generates downsampled images and LMDB databases from video sequences to accelerate training data access. ([source](https://mmagic.readthedocs.io/en/latest/dataset_zoo/vid4.html))
- [Denoising Model Trainers](https://awesome-repositories.com/f/artificial-intelligence-ml/denoising-model-trainers.md) — Organizes image restoration and denoising data into structured directories to facilitate model training and evaluation. ([source](https://mmagic.readthedocs.io/en/latest/dataset_zoo/denoising.html))
- [Distributed Model Execution](https://awesome-repositories.com/f/artificial-intelligence-ml/distributed-model-execution.md) — Executes model testing and inference across single or multiple GPUs to reduce overall evaluation time. ([source](https://mmagic.readthedocs.io/en/latest/user_guides/train_test.html))
- [Encoder-Decoder Architectures](https://awesome-repositories.com/f/artificial-intelligence-ml/encoder-decoder-architectures.md) — Constructs neural network architectures using modular blocks like ResNet and gated convolutions. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/models/archs/index.html))
- [GAN Implementations](https://awesome-repositories.com/f/artificial-intelligence-ml/gan-implementations.md) — Implements data preparation pipelines for unconditional GANs, including tensor conversion and image flipping. ([source](https://mmagic.readthedocs.io/en/latest/dataset_zoo/unconditional_gans.html))
- [GAN Training Loops](https://awesome-repositories.com/f/artificial-intelligence-ml/gan-training-loops.md) — Manages independent optimizers and alternating schedules for generators and discriminators. ([source](https://mmagic.readthedocs.io/en/latest/migration/optimizers.html))
- [Inference Acceleration](https://awesome-repositories.com/f/artificial-intelligence-ml/generative-ai-resources/diffusion-visual-models/generative-ai-models/diffusion-models/inference-acceleration.md) — Accelerates diffusion model sampling by merging redundant tokens in the vision transformer. ([source](https://mmagic.readthedocs.io/en/latest/model_zoo/inpainting.html))
- [Image-to-Image Translation](https://awesome-repositories.com/f/artificial-intelligence-ml/generative-ai-resources/diffusion-visual-models/generative-ai-pipelines/text-to-image-generators/image-inpainting/image-to-image-translation.md) — Implements algorithms to transform images from one visual domain to another based on target styles or prompts. ([source](https://mmagic.readthedocs.io/en/latest/model_zoo/text2image.html))
- [Generative Distribution Assessments](https://awesome-repositories.com/f/artificial-intelligence-ml/generative-distribution-assessments.md) — Implements generative quality evaluation using Fréchet Inception Distance (FID) and Inception Score to compare real and synthetic datasets. ([source](https://mmagic.readthedocs.io/en/latest/user_guides/metrics.html))
- [Generative Diversity Measurements](https://awesome-repositories.com/f/artificial-intelligence-ml/generative-diversity-measurements.md) — Analyzes the variety and smoothness of generative outputs using Perceptual Path Length and MS-SSIM. ([source](https://mmagic.readthedocs.io/en/latest/user_guides/metrics.html))
- [Noise-to-Image Generation](https://awesome-repositories.com/f/artificial-intelligence-ml/generative-image-models/noise-to-image-generation.md) — Samples new images from random noise using both unconditional and conditional generative models. ([source](https://mmagic.readthedocs.io/en/latest/user_guides/inference.html))
- [Generative Model Accuracy Metrics](https://awesome-repositories.com/f/artificial-intelligence-ml/generative-model-accuracy-metrics.md) — Quantifies model performance and accuracy using generative-specific metrics like FID and Precision & Recall. ([source](https://mmagic.readthedocs.io/en/latest/migration/eval_test.html))
- [Automatic Precision Casting](https://awesome-repositories.com/f/artificial-intelligence-ml/half-precision-inference/automatic-precision-casting.md) — Optimizes memory and training speed by automatically casting the forward process to half-precision. ([source](https://mmagic.readthedocs.io/en/latest/migration/amp.html))
- [Video Sequence Preprocessing](https://awesome-repositories.com/f/artificial-intelligence-ml/image-data-preprocessing/video-sequence-preprocessing.md) — Performs temporal mirroring and frame reversal to prepare video sequences for generative model training. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/datasets/transforms/index.html))
- [Conditional Image Generation](https://awesome-repositories.com/f/artificial-intelligence-ml/image-generation-models/conditional-image-generation.md) — Produces synthetic images guided by discrete labels and specific input conditions. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/apis/inferencers/index.html))
- [Perspective Simulation](https://awesome-repositories.com/f/artificial-intelligence-ml/image-generation-models/conditional-image-generation/image-conditioned-3d-generation/perspective-simulation.md) — Produces high-resolution images that simulate 3D perspectives by interpolating camera positions and style codes. ([source](https://mmagic.readthedocs.io/en/latest/model_zoo/3d_aware_generation.html))
- [Structural Conditioning](https://awesome-repositories.com/f/artificial-intelligence-ml/image-generation-models/conditional-image-generation/structural-conditioning.md) — Adds conditional constraints to the generation process to dictate specific structural and visual layouts. ([source](https://mmagic.readthedocs.io/en/latest/model_zoo/text2image.html))
- [Unconditional Generation](https://awesome-repositories.com/f/artificial-intelligence-ml/image-generation/unconditional-generation.md) — Creates realistic images from random noise using various Generative Adversarial Network architectures. ([source](https://mmagic.readthedocs.io/en/latest/model_zoo/unconditional_gans.html))
- [Dataset Preparation](https://awesome-repositories.com/f/artificial-intelligence-ml/image-super-resolution-models/dataset-preparation.md) — Organizes raw image data into directory structures to enable training for blind image super-resolution. ([source](https://mmagic.readthedocs.io/en/latest/dataset_zoo/realsrset.html))
- [Inference Model Deployment](https://awesome-repositories.com/f/artificial-intelligence-ml/inference-model-deployment.md) — Transforms trained generative models into optimized formats like ONNX and TensorRT for hardware accelerators. ([source](https://mmagic.readthedocs.io/en/latest/user_guides/deploy.html))
- [Learning Rate Schedulers](https://awesome-repositories.com/f/artificial-intelligence-ml/learning-rate-schedulers.md) — Dynamically adjusts learning rates and decay factors based on performance metrics to improve convergence. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/engine/schedulers/index.html))
- [Loss Function Implementations](https://awesome-repositories.com/f/artificial-intelligence-ml/loss-function-implementations.md) — Defines new loss modules by wrapping functional implementations in classes and registering them via configuration. ([source](https://mmagic.readthedocs.io/en/latest/howto/losses.html))
- [Hybrid Loss and Architecture Integration](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/model-training-and-tuning/architecture-and-operations/model-architecture/hybrid-loss-and-architecture-integration.md) — Implements the functional merging of primary and auxiliary loss modules to improve training stability. ([source](https://mmagic.readthedocs.io/en/latest/howto/losses.html))
- [Low-Rank Adaptation](https://awesome-repositories.com/f/artificial-intelligence-ml/machine-learning/infrastructure/model-training-and-tuning/fine-tuning-and-customization/model-fine-tuning/low-rank-adaptation.md) — Integrates LoRA layers into modules to enable efficient parameter fine-tuning of generative models. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/models/archs/index.html))
- [Model Complexity Analysis](https://awesome-repositories.com/f/artificial-intelligence-ml/model-complexity-analysis.md) — Evaluates the computational cost of neural networks by calculating parameter counts and floating-point operations (FLOPs). ([source](https://mmagic.readthedocs.io/en/latest/user_guides/useful_tools.html))
- [Model Output Visualizers](https://awesome-repositories.com/f/artificial-intelligence-ml/model-output-visualizers.md) — Generates and saves sample images or GIFs during training to visualize model progress. ([source](https://mmagic.readthedocs.io/en/latest/user_guides/visualization.html))
- [Model Training Optimizers](https://awesome-repositories.com/f/artificial-intelligence-ml/model-training-optimizers.md) — Manages parameter updates, gradient zeroing, and backward passes through a specialized optimization wrapper. ([source](https://mmagic.readthedocs.io/en/latest/migration/optimizers.html))
- [Training Resumption](https://awesome-repositories.com/f/artificial-intelligence-ml/model-training/training-resumption.md) — Restores model states from specific checkpoints to continue training from previous iterations. ([source](https://mmagic.readthedocs.io/en/latest/migration/runtime.html))
- [Weight Smoothing](https://awesome-repositories.com/f/artificial-intelligence-ml/model-weight-reconstruction/weight-smoothing.md) — Provides exponential moving average weight smoothing to improve convergence and stability during model training.
- [Parameter Weight Smoothing](https://awesome-repositories.com/f/artificial-intelligence-ml/model-weight-reconstruction/weight-smoothing/exponential-moving-average-weight-updates/parameter-weight-smoothing.md) — Implements exponential moving average weight maintenance to ensure training stability and better convergence. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/models/base_models/index.html))
- [Multi-Dataset Performance Benchmarking](https://awesome-repositories.com/f/artificial-intelligence-ml/multi-dataset-performance-benchmarking.md) — Calculates multiple quality metrics across several datasets simultaneously to evaluate model performance. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/engine/runner/index.html))
- [NPU Accelerators](https://awesome-repositories.com/f/artificial-intelligence-ml/npu-accelerators.md) — Supports training execution on Huawei Ascend NPU hardware using single or multi-device configurations. ([source](https://mmagic.readthedocs.io/en/latest/device/npu.html))
- [Submodule](https://awesome-repositories.com/f/artificial-intelligence-ml/optimizer-configurations/submodule.md) — Constructs separate optimizers for different model submodules to enable independent learning rates. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/engine/optimizers/index.html))
- [Parameter Optimization Strategies](https://awesome-repositories.com/f/artificial-intelligence-ml/parameter-optimization-strategies.md) — Manages hyper-parameter adjustments via defined optimizers, gradient clipping, and mixed precision strategies. ([source](https://mmagic.readthedocs.io/en/latest/user_guides/config.html))
- [Ground Truth Comparisons](https://awesome-repositories.com/f/artificial-intelligence-ml/segmentation-visualizations/ground-truth-comparisons.md) — Generates side-by-side visual comparisons of input images, ground truth, and model predictions. ([source](https://mmagic.readthedocs.io/en/latest/user_guides/visualization.html))
- [Gradient Accumulation Strategies](https://awesome-repositories.com/f/artificial-intelligence-ml/training-convergence-optimization/batch-size-scaling/gradient-accumulation-strategies.md) — Simulates larger batch sizes by aggregating gradients over multiple iterations before parameter updates. ([source](https://mmagic.readthedocs.io/en/latest/migration/optimizers.html))
- [Image Augmentations](https://awesome-repositories.com/f/artificial-intelligence-ml/training-data-transformations/image-augmentations.md) — Implements random geometric and color transformations to increase dataset variety and model robustness. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/datasets/transforms/index.html))
- [Restoration Dataset Preparation](https://awesome-repositories.com/f/artificial-intelligence-ml/training-dataset-preparation/restoration-dataset-preparation.md) — Organizes raw image pairs into structured directory formats for training image and video restoration models. ([source](https://mmagic.readthedocs.io/en/latest/dataset_zoo/gopro.html))
- [Training Loop Schedulers](https://awesome-repositories.com/f/artificial-intelligence-ml/training-loop-schedulers.md) — Defines optimization processes, learning rate schedules, and iteration limits for model training loops. ([source](https://mmagic.readthedocs.io/en/latest/migration/schedule.html))
- [Unpaired Image Translation](https://awesome-repositories.com/f/artificial-intelligence-ml/unpaired-image-translation.md) — Implements data loading logic to sample random image pairs from separate domains for unsupervised translation. ([source](https://mmagic.readthedocs.io/en/latest/howto/dataset.html))
- [Dataset Preparations](https://awesome-repositories.com/f/artificial-intelligence-ml/unpaired-image-translation/dataset-preparations.md) — Organizes image collections into folder structures required for unpaired image-to-image translation training. ([source](https://mmagic.readthedocs.io/en/latest/dataset_zoo/unpaired_cyclegan.html))
- [Vision Data Loaders](https://awesome-repositories.com/f/artificial-intelligence-ml/vision-data-loaders.md) — Provides configurable vision data loaders to manage dataset sampling, batch sizes, and worker counts. ([source](https://mmagic.readthedocs.io/en/latest/migration/data.html))

### Data & Databases

- [Data Standardization](https://awesome-repositories.com/f/data-databases/data-governance-modeling/data-standardization.md) — Unifies disparate data and metadata into a standardized interface to simplify information flow between multimodal models. ([source](https://mmagic.readthedocs.io/en/latest/advanced_guides/structures.html))
- [Batched Data Loading](https://awesome-repositories.com/f/data-databases/data-pipeline-orchestration/data-engineering-pipelines/batched-data-loading.md) — Stacks multiple multimodal data samples into single batches using tensor-like operations for efficient training. ([source](https://mmagic.readthedocs.io/en/latest/advanced_guides/structures.html))
- [Data Pipelines](https://awesome-repositories.com/f/data-databases/data-pipelines.md) — Constructs composable data pipelines that transform raw images through sequential loading, rescaling, and cropping. ([source](https://mmagic.readthedocs.io/en/latest/migration/data.html))
- [Vision](https://awesome-repositories.com/f/data-databases/dataset-loading/vision.md) — Loads image data and optional annotations required for high-fidelity restoration and inpainting tasks. ([source](https://mmagic.readthedocs.io/en/latest/howto/dataset.html))
- [Multimodal Data Loading](https://awesome-repositories.com/f/data-databases/multimodal-data-storage/multimodal-data-loading.md) — Implements the ingestion of images and video frames from files into structured samples for multimodal training. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/datasets/transforms/index.html))
- [Training Data Pipelines](https://awesome-repositories.com/f/data-databases/data-processing-pipelines/data-processing/ml-data-pipelines/training-data-pipelines.md) — Provides pipelines that load, normalize, and format multimodal data for training on GPU hardware. ([source](https://mmagic.readthedocs.io/en/latest/advanced_guides/data_preprocessor.html))
- [Data Transformation Registrations](https://awesome-repositories.com/f/data-databases/data-processing-pipelines/processing-pipelines/pipeline-customizers/data-transformation-registrations.md) — Supports registering user-defined data transformation functions into a pipeline registry for modular processing sequences. ([source](https://mmagic.readthedocs.io/en/latest/howto/transforms.html))
- [Paired Image Dataset Preparation](https://awesome-repositories.com/f/data-databases/dataset-preparation-tools/image-text-pair-pipelines/paired-image-dataset-preparation.md) — Organizes image pairs into specific directory structures to facilitate supervised image-to-image translation. ([source](https://mmagic.readthedocs.io/en/latest/dataset_zoo/paired_pix2pix.html))
- [Paired Image Loaders](https://awesome-repositories.com/f/data-databases/lazy-image-dataset-loaders/paired-image-loaders.md) — Loads concatenated image pairs from directories for use in image-to-image translation models. ([source](https://mmagic.readthedocs.io/en/latest/howto/dataset.html))
- [LMDB Dataset Converters](https://awesome-repositories.com/f/data-databases/lmdb-dataset-converters.md) — Converts raw image datasets into LMDB memory-mapped databases to accelerate I/O performance during high-throughput training.
- [Image Dataset Format Converters](https://awesome-repositories.com/f/data-databases/structured-data-schemas/format-conversions/segmentation-dataset-conversions/image-dataset-format-converters.md) — Provides preprocessing scripts to crop, resize, and reformat raw image data for training and testing. ([source](https://mmagic.readthedocs.io/en/latest/dataset_zoo/sidd.html))

### Part of an Awesome List

- [3D and Spatial Synthesis](https://awesome-repositories.com/f/awesome-lists/ai/3d-and-spatial-synthesis.md) — Creates 3D-aware generative visuals to produce spatial representations from 2D inputs. ([source](https://cdn.jsdelivr.net/gh/open-mmlab/mmagic@main/README.md))

### DevOps & Infrastructure

- [Multi-Backend Inference Executions](https://awesome-repositories.com/f/devops-infrastructure/model-conversion/tensorflow-lite/webassembly-inference-executions/multi-backend-inference-executions.md) — Executes generated models across various backends or via SDKs to produce visual synthesis results. ([source](https://mmagic.readthedocs.io/en/latest/user_guides/deploy.html))

### Software Engineering & Architecture

- [Registry-Based Extensibility](https://awesome-repositories.com/f/software-engineering-architecture/application-frameworks/application-framework-extensions/plugin-development-kits/source-customization/registry-based-extensibility.md) — Implements a registry-based system allowing custom data transforms and model modules to be loaded via configuration strings.
- [Model Architecture Configurations](https://awesome-repositories.com/f/software-engineering-architecture/application-lifecycle-management/configuration-management/automation-and-templating-frameworks/configuration-modularization/model-architecture-configurations.md) — Specifies network architectures and loss functions using configuration files to instantiate models without code changes. ([source](https://mmagic.readthedocs.io/en/latest/howto/models.html))

### System Administration & Monitoring

- [ML Experiment Logging](https://awesome-repositories.com/f/system-administration-monitoring/ml-experiment-logging.md) — Records training scalars, images, and configuration data to Tensorboard and WandB for real-time monitoring. ([source](https://mmagic.readthedocs.io/en/latest/autoapi/mmagic/visualization/index.html))