PyTorch's TorchTitan Gets Major Update With New SOTA Model Support
PyTorch shipped TorchTitan's biggest update in a year, adding support for DeepSeek-V4, Qwen3.5/Qwen3-VL and Kimi K3, along with 5D parallelism, TitanRL, and Muon optimization delivering up to 79% speedup.
2026-10-22 ~ 2026-10-22 · 2 related posts
- TorchTitan adds 5D parallelism for DeepSeek-V4, Qwen3.5, Kimi K3; Muon optimizer 79% faster — PyTorch · 2026-10-22
- PyTorch's TorchTitan adds DeepSeek-V4, Kimi K3 support, RL training and big scale gains — PyTorch · 2026-10-22