mradermacher/Silicon-Masha-7B-GGUF

Same weights, converted vs LakoMoor/Silicon-Masha-7B. Same weights stored in another dtype (BF16 → F16/F32).

Tensors changed: Norms 0/65, Attention 0/128, MLP 0/96, Embeddings 0/1, lm_head 0/1

On the Hub · Downloads on Model Pulse