DeepCybo/PhysBrain1.5-2B

Partial fine-tune vs Qwen/Qwen3-VL-2B-Instruct. Partial change: 78% of the weights changed (attention, mlp, norms), in 28 of 28 layers.

Tensors changed: Norms 108/217, MLP 84/196, Attention 112/208, Embeddings 0/3

2 derivatives checked

• mradermacher/PhysBrain1.5-2B-i1-GGUF: Quantization (fewer layers)
• mradermacher/PhysBrain1.5-2B-GGUF: Same weights, converted (fewer layers)

On the Hub · Downloads on Model Pulse