Rorical/qode-30b

Full fine-tune vs Qwen/Qwen3-30B-A3B-Base. Fine-tune: 100% of the weights changed, across 48 of 48 layers.

Tensors changed: Embeddings 1/1, Norms 193/193, MLP 2613/2613, Attention 192/192, lm_head 1/1

2 derivatives checked

• mradermacher/qode-30b-i1-GGUF: Quantization
• mradermacher/qode-30b-GGUF: Quantization

On the Hub · Downloads on Model Pulse