pytorch/torchtitan

★5,926PythonBSD-3-Clause⑂ 1,092Issue 113Updated Nov 20, 2026

A PyTorch native platform for training generative AI models

PyTorch-native stack that spans pretraining through RL — torchtitan is a clean-room, PyTorch-native codebase that takes a model from pretraining through SFT to reinforcement learning, composing FSDP2, tensor, pipeline, context, and expert parallelism without rewriting model code. TitanRL now shares the same model definitions and kernels between the trainer and vLLM generation, with a bitwise-reproducible mode, and it already supports DeepSeek, Qwen, Kimi, gpt-oss, Llama, and FLUX.

LLMs & ML

Metrics

Relative growth (7d)+0%Growth over the last 7 days relative to size
Acceleration+13%Change in growth pace from the week before
Attention37The three metrics scaled to 0–100 and combined

Retention can be computed 30 days after a repository first trends (it isn't shown until then). See About for how each metric is defined.

Star growth and Trending appearances (last 30 days)

First listed+12★/dayDaily
Oct 22, 2026Nov 20, 2026

Bands mark days on the trending list (darker = higher rank)

Daily trending ranks as a table (1 days)
DateDaily
Nov 20, 2026
python #13 0

Created Dec 13, 2023 · First on trending Nov 20, 2026

Other notable repositories in the same category (LLMs & ML)

See all LLMs & ML repositories →

Get the weekly digest by email

Every Monday: the repositories that grew most last week, and how the buzz spread between the global and Japanese-language web. Unsubscribe any time. See the privacy and ad policy for how we handle your address.