nmalinowski/pauper-llama3-8b
LoRA-style vs meta-llama/Meta-Llama-3-8B-Instruct. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of k_proj, o_proj, q_proj, v_proj changed, in 32 of 32 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/65, MLP 0/96, Attention 128/128, lm_head 0/1
2 derivatives checked
- mradermacher/pauper-llama3-8b-i1-GGUF: Quantization
- mradermacher/pauper-llama3-8b-GGUF: Same weights, converted