PrimeIntellect/INTELLECT-1-Instruct
Full fine-tune vs PrimeIntellect/INTELLECT-1. Fine-tune: 95% of the weights changed, across 42 of 42 layers.
Tensors changed: Embeddings 0/1, Norms 85/85, MLP 126/126, Attention 168/168, lm_head 1/1
4 derivatives checked
- bartowski/INTELLECT-1-Instruct-GGUF: Same weights, converted
- mradermacher/INTELLECT-1-Instruct-i1-GGUF: Quantization
- mradermacher/INTELLECT-1-Instruct-GGUF: Same weights, converted
- DevQuasar-5/PrimeIntellect.INTELLECT-1-Instruct-GGUF: Quantization