gpt-oss-120b RYS 0..19,17..35

This repository is a layer-routed RYS variant of openai/gpt-oss-120b.

  • Base model revision: b5c939de8f754692c1647ca79fbf85e8c1e70f8a
  • Requested path: 0..19,17..35
  • Resolved path: 0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35
  • Original layers: 36
  • Output layers: 39
  • Repeated source layers: 17,18,19

No additional quantization was applied while building this repo. The tensor bytes are copied directly from the source checkpoint and only re-indexed into a new layer execution path.

The model config has been updated so model.layers follows the path above. Tokenizer and chat template files are copied from the base repository unchanged.

Downloads last month
119
Safetensors
Model size
130B params
Tensor type
BF16
·
U8
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for KotshinZ/gpt-oss-120b-rys-0_19-17_35

Quantized
(98)
this model