mradermacher/BGE-VL-Screenshot-GGUF

Same weights, converted (fewer layers) vs BAAI/BGE-VL-Screenshot. mradermacher/BGE-VL-Screenshot-GGUF matches part of BAAI/BGE-VL-Screenshot (434 of its 824 tensors; the rest of BAAI/BGE-VL-Screenshot has no counterpart). On the shared part: Same weights stored in another dtype (BF16 → F16/F32).

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

On the Hub · Downloads on Model Pulse