ValiantLabs/DeepSeek-R1-0528-Qwen3-8B-Esper3
LoRA-style vs deepseek-ai/DeepSeek-R1-0528-Qwen3-8B. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of k_proj, o_proj, q_proj, v_proj changed, in 36 of 36 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Attention 144/144, Norms 0/145, MLP 0/108, lm_head 0/1
3 derivatives checked
- mradermacher/DeepSeek-R1-0528-Qwen3-8B-Esper3-i1-GGUF: Quantization
- mradermacher/DeepSeek-R1-0528-Qwen3-8B-Esper3-GGUF: Same weights, converted
- DevQuasar/ValiantLabs.DeepSeek-R1-0528-Qwen3-8B-Esper3-GGUF: Same weights, converted