jpacifico/Chocolatine-14B-Instruct-DPO-v1.3
LoRA-style vs microsoft/phi-4. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_up_proj, o_proj, qkv_proj changed, in 40 of 40 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/81, MLP 80/80, Attention 80/80, lm_head 0/1
2 derivatives checked
- mradermacher/Chocolatine-14B-Instruct-DPO-v1.3-i1-GGUF: Quantization
- mradermacher/Chocolatine-14B-Instruct-DPO-v1.3-GGUF: Quantization