ANke121/NAS-PO-GRPO-Qwen3-VL-8B
Partial fine-tune vs Qwen/Qwen3-VL-8B-Instruct. Partial change: 86% of the weights changed (attention, embeddings, mlp, norms), in 36 of 36 layers.
Tensors changed: Embeddings 3/4, Norms 175/261, MLP 213/232, Attention 252/252, lm_head 0/1
1 derivatives checked
- mradermacher/NAS-PO-GRPO-Qwen3-VL-8B-GGUF: Same weights, converted (fewer layers)