mlasli/Muse-Glimmer-30B-Abliterated-FP16-GGUF

Same weights, converted vs mlasli/Muse-Glimmer-30B-Abliterated-BF16. Same weights stored in another dtype (BF16 → F16/F32), plus 104 new tensors.

Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/209, MLP 0/156, Attention 0/260

On the Hub · Downloads on Model Pulse