nvidia/NV-Reason-CXR-3B

Full fine-tune vs Qwen/Qwen2.5-VL-3B-Instruct. Fine-tune: 100% of the weights changed, across 36 of 36 layers.

Tensors changed: Embeddings 2/2, Norms 127/138, MLP 304/304, Attention 380/380

1 derivatives checked

• samwell/NV-Reason-CXR-3B-GGUF: Same weights, converted (fewer layers)

On the Hub · Downloads on Model Pulse