nkpz/Qwen2.5-Dyanka-7B-Preview-Uncensored-DeLMAT
LoRA-style vs Xiaojian9992024/Qwen2.5-Dyanka-7B-Preview. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of gate_proj changed, in 26 of 28 layers; norms, biases and embeddings are untouched.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/57, MLP 26/84, Attention 0/196
2 derivatives checked
- mradermacher/Qwen2.5-Dyanka-7B-Preview-Uncensored-DeLMAT-i1-GGUF: Quantization
- mradermacher/Qwen2.5-Dyanka-7B-Preview-Uncensored-DeLMAT-GGUF: Same weights, converted