DireDreadlord/Dragon-1-0.5B

Full fine-tune vs Qwen/Qwen2-0.5B-Instruct. Fine-tune: 100% of the weights changed, across 24 of 24 layers.

Tensors changed: Norms 13/49, MLP 72/72, Attention 155/168

2 derivatives checked

• mradermacher/Dragon-1-0.5B-i1-GGUF: Quantization
• mradermacher/Dragon-1-0.5B-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse