RLWRLD/RLDX-1-VLM

Full fine-tune vs Qwen/Qwen3-VL-8B-Instruct. Fine-tune: 100% of the weights changed, across 36 of 36 layers.

Tensors changed: Norms 234/261, MLP 231/232, Attention 252/252, Embeddings 2/3

1 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse