Pinkstack/DistilGPT-OSS-qwen3-4B
LoRA-style vs Qwen/Qwen3-4B-Thinking-2507. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, k_proj, o_proj, q_proj, up_proj, v_proj changed, in 36 of 36 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/145, MLP 108/108, Attention 144/144
4 derivatives checked
- mradermacher/DistilGPT-OSS-qwen3-4B-i1-GGUF: Quantization
- mradermacher/DistilGPT-OSS-qwen3-4B-GGUF: Same weights, converted
- DevQuasar/Pinkstack.DistilGPT-OSS-qwen3-4B-GGUF: Same weights, converted
- bunnycore/Qwen3-4B-Max-Merge: Full fine-tune