OpenGVLab/ZeroGUI-AndroidLab-7B

Full fine-tune vs ByteDance-Seed/UI-TARS-7B-DPO. Fine-tune: 93% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 1/2, Norms 121/187, MLP 215/216, Attention 322/324, lm_head 1/1

2 derivatives checked

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

On the Hub · Downloads on Model Pulse