omlab/VLM-R1-Qwen2.5VL-3B-OVD-0321
Full fine-tune vs Qwen/Qwen2.5-VL-3B-Instruct. Fine-tune: 92% of the weights changed, across 36 of 36 layers.
Tensors changed: Embeddings 1/2, Norms 73/138, MLP 277/304, Attention 380/380
2 derivatives checked
- mradermacher/VLM-R1-Qwen2.5VL-3B-OVD-0321-i1-GGUF: Quantization (fewer layers)
- mradermacher/VLM-R1-Qwen2.5VL-3B-OVD-0321-GGUF: Same weights, converted (fewer layers)