LLM-OS-Models/Fabliq-8B-Agent-Reasoning

Full fine-tune vs LLM-OS-Models/Fabliq-8B-Agent. Fine-tune: 92% of the weights changed, across 24 of 24 layers.

Tensors changed: MLP 2015/2162, Embeddings 0/2, other 54/54, Norms 1/60, Attention 23/24

2 derivatives checked

• mradermacher/Fabliq-8B-Agent-Reasoning-i1-GGUF: Quantization
• mradermacher/Fabliq-8B-Agent-Reasoning-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse