mradermacher/Fourier-Qwen2-VL-2B-0.67-GGUF

Same weights, converted (fewer layers) vs whyisverysmart/Fourier-Qwen2-VL-2B-0.67. mradermacher/Fourier-Qwen2-VL-2B-0.67-GGUF matches part of whyisverysmart/Fourier-Qwen2-VL-2B-0.67 (338 of its 729 tensors; the rest of whyisverysmart/Fourier-Qwen2-VL-2B-0.67 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