Ximeng0831/NaviDrive-Qwen3-VL-2B-SFT

Partial fine-tune vs Qwen/Qwen3-VL-2B-Instruct. Partial change: 85% of the weights changed (attention, embeddings, mlp, norms), in 28 of 28 layers.

Tensors changed: Embeddings 3/4, Norms 142/217, MLP 179/196, Attention 208/208

1 derivatives checked

• mradermacher/NaviDrive-Qwen3-VL-2B-SFT-GGUF: Partial fine-tune (fewer layers)

On the Hub · Downloads on Model Pulse