allenai/olmOCR-7B-0725

Full fine-tune vs Qwen/Qwen2.5-VL-7B-Instruct. Fine-tune: 93% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 1/2, Norms 57/122, MLP 276/280, Attention 324/324, lm_head 1/1

2 derivatives checked

• mradermacher/olmOCR-7B-0725-i1-GGUF: Quantization (fewer layers)
• mradermacher/olmOCR-7B-0725-GGUF: Same weights, converted (fewer layers)

1 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse