nicoboss/MedraN-E4B-Uncensored-EP7
LoRA-style vs google/gemma-3n-E4B-it. 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 35 of 35 layers; norms, biases and embeddings are untouched.
Tensors changed: Attention 140/363, Embeddings 0/25, Norms 0/489, other 0/687, MLP 105/112
4 derivatives checked
- mradermacher/MedraN-E4B-Uncensored-EP7-i1-GGUF: Quantization (fewer layers)
- mradermacher/MedraN-E4B-Uncensored-EP7-GGUF: Same weights, converted (fewer layers)
- ijohn07/MedraN-E4B-Uncensored-EP7-Q4_K_M-GGUF: Quantization (fewer layers)
- ijohn07/MedraN-E4B-Uncensored-EP7-Q8_0-GGUF: Quantization (fewer layers)