ManniX-ITA/JackOD-9B-Coder
Full fine-tune vs Qwen/Qwen3.5-9B. Fine-tune: 93% of the weights changed, across 32 of 32 layers.
Tensors changed: lm_head 1/1, Embeddings 1/5, MLP 96/211, Attention 202/336, other 0/1, Norms 105/221
3 derivatives checked
- ManniX-ITA/JackOD-9B-Coder-MTP-GGUF: Same weights, converted
- mradermacher/JackOD-9B-Coder-i1-GGUF: Quantization
- mradermacher/JackOD-9B-Coder-GGUF: Same weights, converted