mradermacher/Qwen2.5-VL-7B-VRAG-GGUF

Same weights, converted (fewer layers) vs Qiuchen-Wang/Qwen2.5-VL-7B-VRAG. mradermacher/Qwen2.5-VL-7B-VRAG-GGUF matches part of Qiuchen-Wang/Qwen2.5-VL-7B-VRAG (339 of its 729 tensors; the rest of Qiuchen-Wang/Qwen2.5-VL-7B-VRAG has no counterpart). On the shared part: Same weights stored in another dtype (BF16 → F16/F32).

Tensors changed: MLP 0/84, Norms 0/57, Attention 0/196, lm_head 0/1, Embeddings 0/1

On the Hub · Downloads on Model Pulse