SeerRay-Lab/Xiaomi-OCR-0

Full fine-tune vs Qwen/Qwen3.5-0.8B-Base. Fine-tune: 100% of the weights changed, across 24 of 24 layers.

Tensors changed: Attention 220/220, Norms 135/135, Embeddings 5/5, MLP 127/127, other 1/1

2 derivatives checked

• prithivMLmods/Xiaomi-OCR-0-GGUF: Same weights, converted
• Soulfate24/Xiaomi-OCR-0-Paretrix: Not comparable yet

On the Hub · Downloads on Model Pulse