tatsuyaaaaaaa/translategemma-12b-it-gguf

Same weights, converted (fewer layers) vs google/translategemma-12b-it. tatsuyaaaaaaa/translategemma-12b-it-gguf matches part of google/translategemma-12b-it (626 of its 1065 tensors; the rest of google/translategemma-12b-it has no counterpart). On the shared part: Same weights stored in another dtype (BF16 → BF16/F32).

Tensors changed: Embeddings 0/1, Norms 0/289, MLP 0/144, Attention 0/192

On the Hub · Downloads on Model Pulse