valiolla/Qwen3.8-4B-Distill-heretic

Full fine-tune vs Qwen/Qwen3.5-4B. Fine-tune: 93% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 1/4, MLP 96/196, Attention 177/320, Norms 51/203

3 derivatives checked

• mradermacher/Qwen3.8-4B-Distill-heretic-GGUF: Same weights, converted
• mradermacher/Qwen3.8-4B-Distill-heretic-i1-GGUF: Quantization
• PJRM/Qwen3.8-4B-Distill-heretic-Q4_K_M-GGUF: Quantization

On the Hub · Downloads on Model Pulse