Qwen/Qwen2.5-Coder-32B
Full fine-tune vs Qwen/Qwen2.5-32B. Fine-tune: 100% of the weights changed, across 64 of 64 layers.
Tensors changed: Embeddings 1/1, Norms 129/129, MLP 192/192, Attention 448/448, lm_head 1/1
18 derivatives checked
- Qwen/Qwen2.5-Coder-32B-Instruct: Full fine-tune
- unsloth/Qwen2.5-Coder-32B: Identical copy
- lmstudio-community/Qwen2.5-Coder-32B-GGUF: Quantization
- bartowski/Qwen2.5-Coder-32B-GGUF: Quantization
- ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-GGUF: Full fine-tune
- mradermacher/Qwen2.5-Coder-32B-i1-GGUF: Quantization
- mbetixz/Qwen2.5-Coder-32B-Instruct: Full fine-tune
- mradermacher/Qwen2.5-Coder-32B-GGUF: Quantization
- raychan0713/Qwen2.5-Coder-32B-Instruct: Full fine-tune
- blestv85/Qwen2.5-Coder-32B-Instruct: Full fine-tune
- tensorblock/Qwen2.5-Coder-32B-GGUF: Quantization
- mlx-community/Qwen2.5-Coder-32B-8bit: Quantization
- mlx-community/Qwen2.5-Coder-32B-4bit: Quantization
- unsloth/Qwen2.5-Coder-32B-bnb-4bit: Quantization
- Snowflake/Qwen-2.5-coder-Arctic-ExCoT-32B: Full fine-tune
- YOYO-AI/Qwen2.5-Coder-32B-YOYO: Full fine-tune
- YOYO-AI/YOYO-O1-32B-V4-preview4: Full fine-tune
- lurf21/Qwen2.5-Coder-32B-NEP: LoRA-style
4 other repos hold byte-for-byte identical weight files.