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
Format unsloth/models/speculative.py (#14997)
Daniel Hanfd924a537792
diff --git a/unsloth/models/speculative.py b/unsloth/models/speculative.pyindex 5c01d91e876..3da8f813376 100644--- a/unsloth/models/speculative.py+++ b/unsloth/models/speculative.py@@ -214,7 +214,6 @@ def _fresh_adapter_dirs(def _drop_stale_head(folder):from ..save import _is_main_rank-stale = os.path.join(folder, "mtp_head.safetensors")if _is_main_rank() and os.path.isfile(stale):os.remove(stale)