gokaygokay/Florence-2-Flux-Large
Full fine-tune vs microsoft/Florence-2-large. Fine-tune: 100% of the weights changed, across 12 of 12 layers.
Tensors changed: Embeddings 6/7, Norms 226/226, other 106/107, Attention 384/384, MLP 192/192
Full fine-tune vs microsoft/Florence-2-large. Fine-tune: 100% of the weights changed, across 12 of 12 layers.
Tensors changed: Embeddings 6/7, Norms 226/226, other 106/107, Attention 384/384, MLP 192/192