flypg/DeepSeek-R1-Distill-Qwen-14B-Japanese-chat
LoRA-style vs cyberagent/DeepSeek-R1-Distill-Qwen-14B-Japanese. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of q_proj, v_proj changed, in 48 of 48 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/97, MLP 0/144, Attention 96/336, lm_head 0/1
1 derivatives checked
- flypg/DeepSeek-R1-Distill-Qwen-14B-Japanese-chat-quant: Quantization