janhq/Jan-v2-VL-high
Partial fine-tune vs Qwen/Qwen3-VL-8B-Thinking. Partial change: 79% of the weights changed (attention, mlp, norms), in 36 of 36 layers.
Tensors changed: Embeddings 0/4, Norms 91/261, MLP 108/232, Attention 144/252, lm_head 0/1
4 derivatives checked
- janhq/Jan-v2-VL-high-gguf: Quantization (fewer layers)
- prithivMLmods/Jan-v2-VL-f32-AIO-GGUF: Same weights, converted (fewer layers)
- mradermacher/Jan-v2-VL-high-i1-GGUF: Quantization (fewer layers)
- mradermacher/Jan-v2-VL-high-GGUF: Same weights, converted (fewer layers)
2 other repos hold byte-for-byte identical weight files.