mradermacher/ZwZ-8B-GGUF

Same weights, converted (fewer layers) vs inclusionAI/ZwZ-8B. mradermacher/ZwZ-8B-GGUF matches part of inclusionAI/ZwZ-8B (399 of its 750 tensors; the rest of inclusionAI/ZwZ-8B has no counterpart). On the shared part: Same weights stored in another dtype (BF16 → F16/F32).

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

On the Hub · Downloads on Model Pulse