duyleekun/Vintern-1B-v3_5-BF16-GGUF

Same weights, converted (fewer layers) vs 5CD-AI/Vintern-1B-v3_5. duyleekun/Vintern-1B-v3_5-BF16-GGUF matches part of 5CD-AI/Vintern-1B-v3_5 (291 of its 637 tensors; the rest of 5CD-AI/Vintern-1B-v3_5 has no counterpart). On the shared part: Same weights stored in another dtype (F32 → BF16/F32).

Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/49, MLP 0/72, Attention 0/168

On the Hub · Downloads on Model Pulse