stvlynn/Qwen-7B-Chat-Cantonese
LoRA-style vs Qwen/Qwen-7B-Chat. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of c_attn changed, in 32 of 32 layers; norms, biases and embeddings are untouched.
Tensors changed: Attention 32/96, Norms 0/65, MLP 0/96, Embeddings 0/1, lm_head 0/1
2 derivatives checked
- mradermacher/Qwen-7B-Chat-Cantonese-i1-GGUF: Quantization
- mradermacher/Qwen-7B-Chat-Cantonese-GGUF: Same weights, converted