Simple and efficient training framework for long-context models
-
Updated
Jan 12, 2026 - Python
Simple and efficient training framework for long-context models
Repo of CACL framework for bot detection
Serious calibration and benchmarking for the `state_collapser` HRL package
A PyTorch framework that handles object detection across 6 different architectures (RetinaNet, Faster R-CNN, SSD, FCOS, and more). Takes care of the optimization setup and training quirks for each model.
Intelligent training framework that automatically skips mastered samples and gives 5× more compute to hard ones. Up to 80% compute savings on LLM fine-tuning.
Tensorless is an open-source Python ML framework that automates the complex parts of model training — from data preparation and model selection to hardware detection, checkpointing, and inference. Just provide your data and let Tensorless handle the rest.
A PyTorch framework for image classification covering 11 CNN architectures (ResNet, EfficientNet, MobileNet, etc.). Handles the optimization setup and training specifics for each model.
A Toy Framework for Model Training
Zero-RAM, JAX-Centric Dataloading, Streaming, and Asynchronous Checkpointing Toolkit
A lightweight PyTorch framework designed to train large deep learning models on resource-constrained hardware (laptops/desktops) using dataset streaming, dynamic batch size auto-scaling, and resumable checkpoints to prevent Out-Of-Memory (OOM) crashes.
Library for config based Neural Network Training
openPangu-2.0 style MoE LLM full-pipeline framework: pretrain -> SFT -> RL (GRPO/PPO) -> serving, Ascend-native with a deterministic trainable mock backend (zero-dependency core).
Fault-tolerant distributed training framework with async checkpointing for LLM's
A comprehensive framework for developing YOLO family models, featuring streamlined workflows for training, validation, testing, and deployment through easy-to-use config files, enabling flexible customization to suit various object detection tasks.
PyTorch training framework with AMP, checkpointing, TensorBoard, profiling, and modular experiment pipelines.
To associate your repository with the training-framework topic, visit your repo's landing page and select "manage topics."