ibm-granite/granite-3.0-8b-instruct
Full fine-tune vs ibm-granite/granite-3.0-8b-base. Fine-tune: 98% of the weights changed, across 40 of 40 layers.
Tensors changed: Norms 81/81, MLP 120/120, Attention 148/160
7 derivatives checked
- bartowski/granite-3.0-8b-instruct-GGUF: Same weights, converted
- Zynerji/Ektome-granite-3.0-8b-instruct-PristinelyUncensored: Abliterated
- MaziyarPanahi/granite-3.0-8b-instruct-GGUF: Same weights, converted
- mradermacher/granite-3.0-8b-instruct-i1-GGUF: Quantization
- mradermacher/granite-3.0-8b-instruct-GGUF: Same weights, converted
- DevQuasar-8/ibm-granite.granite-3.0-8b-instruct-GGUF: Quantization
- KellanF89/Newton-Insights-V1-cannabis-extraction-science: Full fine-tune