Epiculous/NovaSpark

Full fine-tune vs grimjim/Llama-3.1-SuperNova-Lite-lorabilterated-8B. Fine-tune: 93% of the weights changed, across 32 of 32 layers.

Tensors changed: lm_head 1/1, Embeddings 0/1, Norms 14/65, MLP 96/96, Attention 128/128

3 derivatives checked

• featherless-ai-quants/Epiculous-NovaSpark-GGUF: Quantization
• mradermacher/NovaSpark-i1-GGUF: Quantization
• mradermacher/NovaSpark-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse