mradermacher/gemma-4b-ru-handwrite-GGUF

Same weights, converted (fewer layers) vs DocDuck/gemma-4b-ru-handwrite. mradermacher/gemma-4b-ru-handwrite-GGUF matches part of DocDuck/gemma-4b-ru-handwrite (444 of its 883 tensors; the rest of DocDuck/gemma-4b-ru-handwrite 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

On the Hub · Downloads on Model Pulse