CultriX/Qwen2.5-14B-Wernicke-DPO
LoRA-style vs CultriX/Qwen2.5-14B-Wernicke. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, k_proj, o_proj, q_proj, up_proj, v_proj changed, in 48 of 48 layers; norms, biases and embeddings are untouched.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/97, MLP 144/144, Attention 192/336
2 derivatives checked
- mradermacher/Qwen2.5-14B-Wernicke-DPO-i1-GGUF: Quantization
- mradermacher/Qwen2.5-14B-Wernicke-DPO-GGUF: Quantization