tngtech/olmOCR-7B-faithful
Full fine-tune vs allenai/olmOCR-7B-0225-preview. Fine-tune: 93% of the weights changed, across 32 of 32 layers.
Tensors changed: Embeddings 1/2, Norms 131/187, MLP 215/216, Attention 324/324, lm_head 1/1
3 derivatives checked
- mradermacher/olmOCR-7B-faithful-i1-GGUF: Quantization (fewer layers)
- mradermacher/olmOCR-7B-faithful-GGUF: Same weights, converted (fewer layers)
- DevQuasar/tngtech.olmOCR-7B-faithful-GGUF: Same weights, converted (fewer layers)