stepfun-ai/Step3-VL-10B
Partial fine-tune vs stepfun-ai/Step3-VL-10B-Base. Partial change: 81% of the weights changed (attention, embeddings, lm_head, mlp, norms, other), in 36 of 36 layers.
Tensors changed: Norms 145/335, MLP 108/296, Attention 144/332, Embeddings 1/2, other 1/100, lm_head 1/1
6 derivatives checked
- seanbailey518/Step3-VL-10B-GGUF: Same weights, converted (fewer layers)
- Vastined/Step3-VL-10B-GGUF: Same weights, converted (fewer layers)
- np-deploys/Step3-VL-10B-AWQ-4bit: Quantization
- JamePeng2023/Step3-VL-10B-GGUF: Same weights, converted (fewer layers)
- yujiepan/step3-vl-tiny-random: Different architecture
- huihui-ai/Huihui-Step3-VL-10B-abliterated: Abliterated