mradermacher/Granite-Estigia-MultilanguageV3-GGUF

Same weights, converted vs franciscobdl/Granite-Estigia-MultilanguageV3. Same weights stored in another dtype (BF16 → F16/F32), plus 64 new tensors.

Tensors changed: Embeddings 0/1, other 0/64, Norms 0/65, Attention 0/128

On the Hub · Downloads on Model Pulse