mradermacher/Merge-Gemma-2b-it-with-a-Fine-Tuned-one-for-Arabic-GGUF

Same weights, converted vs Ruqiya/Merge-Gemma-2b-it-with-a-Fine-Tuned-one-for-Arabic. Same weights stored in another dtype (BF16 → F16/F32).

Tensors changed: Norms 0/37, MLP 0/54, Attention 0/72, Embeddings 0/1

On the Hub · Downloads on Model Pulse