mradermacher/MARTHA-LXVII.8b-GGUF

Same weights, converted vs ZERO-POINT-AI/MARTHA-LXVII.8B_QWEN-3.5-3.6_prune_9b-3.6_base. Same weights stored in another dtype (BF16 → F16/F32), plus 21 new tensors.

Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/71, MLP 0/84, Attention 0/28

On the Hub · Downloads on Model Pulse