miike-ai/DeepSeek-R1-0528-Qwen3-11B
Partial fine-tune vs deepseek-ai/DeepSeek-R1-0528-Qwen3-8B. Partial change: 28% of the weights changed (attention, mlp, norms), in 12 of 36 layers.
Tensors changed: Embeddings 0/1, Attention 48/144, Norms 48/145, MLP 36/108, lm_head 0/1
2 derivatives checked
- mradermacher/DeepSeek-R1-0528-Qwen3-11B-i1-GGUF: Quantization
- mradermacher/DeepSeek-R1-0528-Qwen3-11B-GGUF: Same weights, converted