matchaaaaa/Testarossa-v1-27B
LoRA-style vs TheDrummer/Gemmasutra-Pro-27B-v1. 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 34 of 46 layers; norms, biases and embeddings are untouched.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/185, MLP 102/138, Attention 136/184
3 derivatives checked
- backyardai/Testarossa-v1-27B-GGUF: Quantization
- mradermacher/Testarossa-v1-27B-i1-GGUF: Quantization
- mradermacher/Testarossa-v1-27B-GGUF: Quantization