flammenai/Llama3.1-Flammades-70B
LoRA-style vs nbeerbower/Llama3.1-Gutenberg-Doppel-70B. 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 80 of 80 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/161, MLP 240/240, Attention 320/320, lm_head 0/1
6 derivatives checked
- mradermacher/Llama3.1-Flammades-70B-i1-GGUF: Quantization
- mradermacher/Llama3.1-Flammades-70B-GGUF: Quantization
- Jolly-Q/Llama-3.3-70B-Inst-Ablit-Flammades-SLERP: Full fine-tune
- flammenai/Mahou-1.5-llama3.1-70B: LoRA-style
- nbeerbower/Llama-3.1-Saoirse-70B: Full fine-tune
- nbeerbower/Llama3-Asobi-70B: Full fine-tune