PhilSad/Lucie-7B-GRPO-Science-500
LoRA-style vs OpenLLM-France/Lucie-7B-Instruct-v1.1. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of q_proj, v_proj changed, in 32 of 32 layers; norms, biases and embeddings are untouched.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/65, MLP 0/96, Attention 59/128
2 derivatives checked
- mradermacher/Lucie-7B-GRPO-Science-500-i1-GGUF: Quantization
- mradermacher/Lucie-7B-GRPO-Science-500-GGUF: Same weights, converted