mradermacher/Medra4b-vis-ep0.4-GGUF
Same weights, converted (fewer layers) vs drwlf/Medra4b-vis-ep0.4. mradermacher/Medra4b-vis-ep0.4-GGUF matches part of drwlf/Medra4b-vis-ep0.4 (444 of its 883 tensors; the rest of drwlf/Medra4b-vis-ep0.4 has no counterpart). On the shared part: Same weights stored in another dtype (BF16 → F16/F32).
Tensors changed: Embeddings 0/1, Norms 0/205, MLP 0/102, Attention 0/136