Modelpublic

TiGa-RCE/Scalpel-VL-1.6B-oQ8

sourceHugging Faceupdated 1mo agoView on Hugging Face

0likes7downloads

Model Card

[!IMPORTANT] This quantization was uploaded on 2026-08-26 and replaces a previous version. If you downloaded this model before this date, please re-download for the updated weights.

Scalpel-VL-1.6B-oQ8

This model was quantized using oQ (oMLX v0.6.3rc3) mixed-precision quantization.

Quantization details

—Model type: qwen3_vl
—Bits: 8
—Group size: 64
—Format: MLX safetensors