Boggy666/EVisRAG-3B
Partial fine-tune vs Qwen/Qwen2.5-VL-3B-Instruct. Partial change: 82% of the weights changed (attention, embeddings, mlp, norms), in 36 of 36 layers.
Tensors changed: Embeddings 1/2, Norms 70/138, MLP 108/304, Attention 252/380
2 derivatives checked
- mradermacher/EVisRAG-3B-i1-GGUF: Quantization (fewer layers)
- mradermacher/EVisRAG-3B-GGUF: Same weights, converted (fewer layers)
2 other repos hold byte-for-byte identical weight files.