Luigi/edge-fall-vlm-2.2b
LoRA-style vs HuggingFaceTB/SmolVLM2-2.2B-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 27 of 27 layers; norms, biases and embeddings are untouched.
Tensors changed: other 0/1, Embeddings 0/4, Norms 0/159, MLP 0/180, Attention 177/312, lm_head 0/1