mradermacher/Youtu-VL-4B-Instruct-GGUF
Same weights, converted vs tencent/Youtu-VL-4B-Instruct. Same weights stored in another dtype (BF16 → F16/F32), plus 80 new tensors.
Tensors changed: Norms 0/161, MLP 0/120, Attention 0/160, Embeddings 0/1
Same weights, converted vs tencent/Youtu-VL-4B-Instruct. Same weights stored in another dtype (BF16 → F16/F32), plus 80 new tensors.
Tensors changed: Norms 0/161, MLP 0/120, Attention 0/160, Embeddings 0/1