Local UI to run and train LLMs and diffusion models. Supports GGUF, MLX, Qwen3.8, DeepSeek-V4.1, MiniMax-H3, Gemma 4, FLUX and more.
Python agentaichatgptdeepseekfine-tuninggemmaimage-generationllamallmllmsopenaipythonqwenreinforcement-learningself-hostedstable-diffusiontext-to-speechttsuiunsloth
Release
Daniel Han0aaea8b585b0
diff --git a/pyproject.toml b/pyproject.tomlindex ef1cf4b607a..5b6e17108a0 100644--- a/pyproject.toml+++ b/pyproject.toml@@ -229,7 +229,7 @@ huggingfacenotorch = [## 2026.9.9 is the first zoo carrying the ceilings below and the Gemma3 token_type_ids fix# (unsloth-zoo#1505) that transformers >= 5.15 needs.- "unsloth_zoo>=2026.9.9",+ "unsloth_zoo>=2026.10.1","wheel>=0.42.0","packaging","numpy",@@ -297,7 +297,7 @@ audio-torch280 = []huggingface = ["unsloth[huggingfacenotorch]",- "unsloth_zoo>=2026.9.9",+ "unsloth_zoo>=2026.10.1","torchvision","unsloth[triton]",]@@ -969,7 +969,7 @@ colab-ampere-torch220 = ["flash-attn>=2.6.3 ; ('linux' in sys_platform)",]colab-new = [- "unsloth_zoo>=2026.9.9",+ "unsloth_zoo>=2026.10.1","packaging","tyro","transformers>=4.52.4,!=4.53.0,!=4.54.0,!=4.55.0,!=4.55.1,!=4.57.0,!=4.57.4,!=4.57.5,!=5.0.0,!=5.1.0,<=5.17.0",diff --git a/unsloth/_version.py b/unsloth/_version.pyindex 4f1968a5c45..e25d5ff9a8d 100644--- a/unsloth/_version.py+++ b/unsloth/_version.py@@ -15,4 +15,4 @@# and the torch-free MLX path in __init__.py imports it directly, which it cannot do for# models/_utils.py (torch, transformers, trl, peft, Triton). That is why the MLX branch# used to borrow unsloth_zoo's version, reporting a different package's number.-__version__ = "2026.9.14"+__version__ = "2026.10.1"