stepfun-ai/GELab-Zero-4B-preview

Full fine-tune vs Qwen/Qwen3-VL-4B-Instruct. Fine-tune: 91% of the weights changed, across 36 of 36 layers.

Tensors changed: Embeddings 1/4, Norms 130/249, MLP 108/220, Attention 144/240

3 derivatives checked

• bartowski/stepfun-ai_GELab-Zero-4B-preview-GGUF: Same weights, converted (fewer layers)
• prithivMLmods/GELab-Zero-4B-preview-GGUF: Same weights, converted (fewer layers)
• noctrex/GELab-Zero-4B-preview-GGUF: Same weights, converted (fewer layers)

On the Hub · Downloads on Model Pulse