katanemo/Arch-Function-Chat-1.5B
Full fine-tune vs Qwen/Qwen2.5-Coder-1.5B-Instruct. Fine-tune: 100% of the weights changed, across 28 of 28 layers.
Tensors changed: Embeddings 1/1, Norms 56/57, MLP 84/84, Attention 196/196
3 derivatives checked
- bartowski/katanemo_Arch-Function-Chat-1.5B-GGUF: Same weights, converted
- mradermacher/Arch-Function-Chat-1.5B-i1-GGUF: Quantization
- mradermacher/Arch-Function-Chat-1.5B-GGUF: Same weights, converted