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

Full fine-tune vs LLM-OS-Models/LFM2.5-8B-A1B-Terminal-ToolBench-Full-SFT-1Epoch. Fine-tune: 96% of the weights changed, across 24 of 24 layers.

Tensors changed: MLP 2122/2162, Embeddings 0/2, other 54/54, Norms 2/60, Attention 24/24

3 derivatives checked

• mradermacher/Fabliq-8B-Agent-i1-GGUF: Quantization
• mradermacher/Fabliq-8B-Agent-GGUF: Same weights, converted
• LLM-OS-Models/Fabliq-8B-Agent-Reasoning: Full fine-tune

On the Hub · Downloads on Model Pulse