CodeGoat24/UnifiedReward-2.0-qwen3vl-32b
Full fine-tune vs Qwen/Qwen3-VL-32B-Instruct. Fine-tune: 96% of the weights changed, across 64 of 64 layers.
Tensors changed: Embeddings 1/4, Norms 239/373, MLP 192/316, Attention 256/364, lm_head 1/1
5 derivatives checked
- mradermacher/UnifiedReward-2.0-qwen3vl-32b-i1-GGUF: Quantization (fewer layers)
- mradermacher/UnifiedReward-2.0-qwen3vl-32b-GGUF: Quantization (fewer layers)
- CodeGoat24/UniGenBench-EvalModel-qwen3vl-32b-v1: Full fine-tune
- CodeGoat24/UnifiedReward-Edit-qwen3vl-32b: Full fine-tune
- CodeGoat24/UnifiedReward-Think-qwen3vl-32b: Full fine-tune