Reza2kn/surya-ocr-2-persian-handwriting-369k
Full fine-tune vs datalab-to/surya-ocr-2. Fine-tune: 100% of the weights changed, across 24 of 24 layers.
Tensors changed: Embeddings 4/4, Norms 52/129, Attention 180/216, MLP 115/124
Full fine-tune vs datalab-to/surya-ocr-2. Fine-tune: 100% of the weights changed, across 24 of 24 layers.
Tensors changed: Embeddings 4/4, Norms 52/129, Attention 180/216, MLP 115/124