alpha-ai/gemma-3N-E2B-Just-Chatty
LoRA-style vs unsloth/gemma-3n-E2B-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 30 of 30 layers; norms, biases and embeddings are untouched.
Tensors changed: Attention 120/343, Embeddings 0/25, Norms 0/444, other 0/647, MLP 90/97
1 derivatives checked
- mradermacher/gemma-3N-E2B-Just-Chatty-GGUF: Same weights, converted (fewer layers)