hamishivi/Qwen3.5-2B
Partial fine-tune vs Qwen/Qwen3.5-2B-Base. Partial change: 87% of the weights changed (attention, embeddings, mlp, norms, other), in 24 of 24 layers.
Tensors changed: Attention 146/268, Norms 87/183, Embeddings 3/5, MLP 79/175, other 1/1
35 other repos hold byte-for-byte identical weight files.