Sakalti/Calme-3.2-fp16

Same weights, converted vs MaziyarPanahi/calme-3.2-instruct-78b. Same weights stored in another dtype (F32 → F16).

Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/173, MLP 0/258, Attention 0/602

3 derivatives checked

• mradermacher/Calme-3.2-fp16-i1-GGUF: Quantization
• mradermacher/Calme-3.2-fp16-GGUF: Quantization
• mradermacher/ultima-1-GGUF: Quantization

On the Hub · Downloads on Model Pulse