prometheus04/vibethinker-1.5b-rstar-coder
LoRA-style vs WeiboAI/VibeThinker-1.5B. 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 28 of 28 layers; norms, biases and embeddings are untouched.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/57, MLP 0/84, Attention 112/196
1 derivatives checked
- mradermacher/vibethinker-1.5b-rstar-coder-GGUF: Same weights, converted