inference-optimization/Qwen3-VL-1.0B-A0.4B-Instruct
Full fine-tune (fewer layers) vs Qwen/Qwen3-VL-30B-A3B-Instruct. inference-optimization/Qwen3-VL-1.0B-A0.4B-Instruct matches part of Qwen/Qwen3-VL-30B-A3B-Instruct (87 of its 882 tensors; the rest of Qwen/Qwen3-VL-30B-A3B-Instruct has no counterpart). On the shared part: Fine-tune: 100% of the weights changed, across 4 of 4 layers.
Tensors changed: Embeddings 4/4, Norms 29/29, MLP 33/33, Attention 20/20, lm_head 1/1
1 other repos hold byte-for-byte identical weight files.