ggml-org/granite-docling-258M-GGUF

Same weights, converted (fewer layers) vs ibm-granite/granite-docling-258M. ggml-org/granite-docling-258M-GGUF matches part of ibm-granite/granite-docling-258M (272 of its 470 tensors; the rest of ibm-granite/granite-docling-258M has no counterpart). On the shared part: Same weights stored in another dtype (BF16 → F16/F32).

Tensors changed: Embeddings 0/1, Norms 0/61, MLP 0/90, Attention 0/120

On the Hub · Downloads on Model Pulse