CodeGoat24/UnifiedReward-2.0-qwen-7b
Partial fine-tune vs Qwen/Qwen2.5-VL-7B-Instruct. Partial change: 85% of the weights changed (attention, lm_head, mlp, norms), in 28 of 32 layers.
Tensors changed: Embeddings 0/2, Norms 54/122, MLP 84/280, Attention 196/324, lm_head 1/1
3 derivatives checked
- mradermacher/UnifiedReward-2.0-qwen-7b-i1-GGUF: Full fine-tune (fewer layers)
- mradermacher/UnifiedReward-2.0-qwen-7b-GGUF: Full fine-tune (fewer layers)
- CodeGoat24/UnifiedReward-Edit-qwen-7b: Partial fine-tune