anushreeberlia/fashion-florence
LoRA-style vs microsoft/Florence-2-large. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of fc1, fc2, k_proj, out_proj, proj, q_proj, qkv, v_proj changed, in 12 of 12 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/9, Norms 0/226, other 0/105, Attention 192/384, MLP 96/192