yuxinlu1/gemma-4-12B-coder-fable5-composer2.5-v1
Full fine-tune vs google/gemma-4-12B-it. Fine-tune: 91% of the weights changed, across 48 of 48 layers.
Tensors changed: Embeddings 0/12, Norms 285/289, other 6/48, MLP 144/144, Attention 184/184
13 derivatives checked
- mlx-community/gemma-4-12b-coder-fable5-composer2.5-4bit: Quantization
- mradermacher/gemma-4-12B-coder-fable5-composer2.5-v1-i1-GGUF: Quantization
- mlx-community/gemma-4-12B-coder-fable5-composer2.5-v1-OptiQ-4bit: Quantization
- mlx-community/gemma-4-12b-coder-fable5-composer2.5-8bit: Quantization
- mradermacher/gemma-4-12B-coder-fable5-composer2.5-v1-GGUF: Quantization
- tpls/gemma-4-12B-coder-fable5-composer2.5-v1-abliterated: Abliterated
- huihui-ai/Huihui-gemma-4-12B-coder-fable5-composer2.5-v1-abliterated: Abliterated
- mlx-community/gemma-4-12b-coder-fable5-composer2.5: Same weights, repacked
- tpls/gemma-4-12B-coder-fable5-composer2.5-v1-sft-v3: LoRA-style
- tpls/gemma-4-12B-coder-fable5-composer2.5-v1-steer: Quantization
- mlx-community/Huihui-gemma-4-12B-coder-fable5-composer2.5-v1-abliterated-4bit-msq: Quantization
- tpls/gemma-4-12B-coder-fable5-composer2.5-v1-sft-v5: LoRA-style
- tpls/gemma-4-12B-coder-fable5-composer2.5-v1-sft: LoRA-style
4 other repos hold byte-for-byte identical weight files.