mradermacher/PAPO-D-Qwen2.5-VL-3B-GGUF

Same weights, converted (fewer layers) vs PAPO-Galaxy/PAPO-D-Qwen2.5-VL-3B. mradermacher/PAPO-D-Qwen2.5-VL-3B-GGUF matches part of PAPO-Galaxy/PAPO-D-Qwen2.5-VL-3B (435 of its 825 tensors; the rest of PAPO-Galaxy/PAPO-D-Qwen2.5-VL-3B has no counterpart). On the shared part: Same weights stored in another dtype (BF16 → F16/F32).

Tensors changed: MLP 0/108, Norms 0/73, Attention 0/252, lm_head 0/1, Embeddings 0/1

On the Hub · Downloads on Model Pulse