ML Frameworks

It split off from `verl` to give diffusion, video, and omni-modality models an RL post-training framework that doesn't treat them like chatbots.

Pits PatchTST, TFT, and N-HiTS against CatBoost to help energy markets pick a forecasting model.

TabFM exists so you can run classification and regression on messy, mixed-type tables without retraining a model on your data.

NeMo AutoModel automates the busywork of wiring HuggingFace LLMs and VLMs into PyTorch-native distributed training so you can fine-tune or pretrain without hand-rolling parallelism code.

KTransformers makes CPU-GPU heterogeneous inference and fine-tuning for massive MoE models almost practical on consumer hardware.

Curated technical deep-dives covering everything from NVLink signal integrity to Kubernetes GPU scheduling and Huawei NPU porting.

Google's edge ML runtime grows up, adds async NPU support, and finally admits PyTorch exists.

This engine decouples RL training and inference into independent GPU services, letting text, vision, and audio models post-train asynchronously at scale.

This book teaches modern GPU kernel programming as a hardware-first progression, using the Blackwell architecture and a Python IR DSL called TIRx to move from concepts to production-grade kernels.

A Chinese-language notebook curriculum that teaches LLM inference engineering by rebuilding vLLM and SGLang internals in Python.

Chinese open-source community Datawhale built a from-zero embodied-AI course that gets you from `print('hello')` to fine-tuning SmolVLA and Pi0.

MiniMind-O packs listen-see-speak intelligence into a 0.1B-parameter model you can retrain from the first line of code on a single desktop GPU.

One framework to run RL post-training on diffusion, autoregressive, and hybrid models without rewriting the orchestration layer each time.

A PyTorch implementation of "Attention Is All You Need" that scales from 13M to multi-billion parameter models.

A tiny deep-learning model that guesses what a file actually contains, not just what its extension claims.

A library of 98 structured skill packs that turns coding agents into end-to-end AI researchers, from distributed training to LaTeX submission.

It corrals the latest subquadratic sequence-model research into hardware-efficient, training-ready PyTorch layers verified across NVIDIA, AMD, and Intel GPUs.

RLinf exists because fine-tuning policies for robots and agents still requires rewriting your training stack for every new simulator, world model, or hardware rig.

MLX-VLM crams speculative decoding, continuous batching, and KV cache quantization into a Mac-native toolkit for running multimodal models locally.

A production-hardened fork of slime that keeps massive MoE models from collapsing by obsessing over bit-wise alignment between rollout and training.