VertexAGI/prism-caption-1-micro

Partial fine-tune vs mlx-community/gemma-3-1b-it-4bit. Partial change: 16% of the weights changed (attention, mlp), in 8 of 26 layers.

Tensors changed: lm_head 0/3, Embeddings 0/3, Norms 0/157, MLP 72/234, Attention 96/312

On the Hub · Downloads on Model Pulse