Sorawiz/Qwen2.5-14B-Instinct-Talk

Full fine-tune vs Sao10K/14B-Qwen2.5-Freya-x1. Fine-tune: 100% of the weights changed, across 48 of 48 layers.

Tensors changed: Norms 96/97, MLP 144/144, Attention 336/336

2 derivatives checked

• mradermacher/Qwen2.5-14B-Instinct-Talk-i1-GGUF: Quantization
• mradermacher/Qwen2.5-14B-Instinct-Talk-GGUF: Quantization

On the Hub · Downloads on Model Pulse