CodeGoat24/UnifiedReward-Think-qwen3vl-2b
Partial fine-tune vs CodeGoat24/UnifiedReward-2.0-qwen3vl-2b. Partial change: 81% of the weights changed (attention, embeddings, mlp, norms), in 28 of 28 layers.
Tensors changed: Embeddings 2/4, Norms 108/217, MLP 84/196, Attention 112/208
1 derivatives checked
- mradermacher/UnifiedReward-Think-qwen3vl-2b-GGUF: Same weights, converted (fewer layers)