karakuri-ai/karakuri-vl-32b-instruct-2507
Full fine-tune vs Qwen/Qwen2.5-VL-32B-Instruct. Fine-tune: 100% of the weights changed, across 64 of 64 layers.
Tensors changed: Attention 576/576, MLP 388/388, Norms 194/194, Embeddings 2/2, lm_head 1/1
4 derivatives checked
- mradermacher/karakuri-vl-32b-instruct-2507-i1-GGUF: Quantization (fewer layers)
- mmnga/karakuri-vl-32b-instruct-2507-gguf: Quantization (fewer layers)
- mradermacher/karakuri-vl-32b-instruct-2507-GGUF: Quantization (fewer layers)
- karakuri-ai/karakuri-vl-32b-thinking-2507-exp: Full fine-tune