larshiakzemil/glm-ocr-farsi
Full fine-tune vs zai-org/GLM-OCR. Fine-tune: 92% of the weights changed, across 24 of 24 layers.
Tensors changed: lm_head 1/1, Embeddings 2/3, Norms 90/164, MLP 179/179, Attention 160/160, other 3/3
Full fine-tune vs zai-org/GLM-OCR. Fine-tune: 92% of the weights changed, across 24 of 24 layers.
Tensors changed: lm_head 1/1, Embeddings 2/3, Norms 90/164, MLP 179/179, Attention 160/160, other 3/3