mradermacher/NQLSG-Qwen2-VL-2B-v2-Base-GGUF

Same weights, converted (fewer layers) vs Lunzima/NQLSG-Qwen2-VL-2B-v2-Base. mradermacher/NQLSG-Qwen2-VL-2B-v2-Base-GGUF matches part of Lunzima/NQLSG-Qwen2-VL-2B-v2-Base (338 of its 729 tensors; the rest of Lunzima/NQLSG-Qwen2-VL-2B-v2-Base has no counterpart). On the shared part: Same weights stored in another dtype (BF16 → F16/F32).

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

On the Hub · Downloads on Model Pulse