MLAdaptiveIntelligence/LLaVAction-0.5B

Full fine-tune vs lmms-lab/llava-onevision-qwen2-0.5b-ov. Fine-tune: 100% of the weights changed, across 26 of 26 layers.

Tensors changed: Embeddings 4/4, other 5/5, Norms 126/155, MLP 174/176, Attention 376/376

On the Hub · Downloads on Model Pulse