mradermacher/drkernel-8b-GGUF

Same weights, converted vs hkust-nlp/drkernel-8b. Same weights stored in another dtype (BF16 → F16/F32).

Tensors changed: Norms 0/145, Attention 0/144, MLP 0/108, lm_head 0/1, Embeddings 0/1

On the Hub · Downloads on Model Pulse