DevQuasar/cerebras.GLM-4.5-Air-REAP-82B-A12B-GGUF

Same weights, converted vs cerebras/GLM-4.5-Air-REAP-82B-A12B. Same weights stored in another dtype (BF16/F32 → F16/F32), plus 184 new tensors.

Tensors changed: Embeddings 0/2, Norms 0/98, MLP 0/187, Attention 0/329, lm_head 0/1, other 0/2

On the Hub · Downloads on Model Pulse