saishf/SOVL-Mega-Mash-L3-8B

Full fine-tune vs saishf/Merge-Mayhem-L3-V2. Fine-tune: 93% of the weights changed, across 32 of 32 layers.

Tensors changed: lm_head 1/1, Embeddings 0/1, Norms 15/65, MLP 96/96, Attention 128/128

3 derivatives checked

• mradermacher/SOVL-Mega-Mash-L3-8B-GGUF: Same weights, converted
• mradermacher/SOVL-Mega-Mash-L3-8B-i1-GGUF: Quantization
• featherless-ai-quants/saishf-SOVL-Mega-Mash-L3-8B-GGUF: Quantization

On the Hub · Downloads on Model Pulse