prithivMLmods/Codepy-Deepthink-3B
LoRA-style vs meta-llama/Llama-3.2-3B-Instruct. 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 28 of 28 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/57, MLP 84/84, Attention 112/112
8 derivatives checked
- mradermacher/Codepy-Deepthink-3B-i1-GGUF: Quantization
- QuantFactory/Llama-Deepsync-3B-GGUF: Full fine-tune
- mradermacher/Codepy-Deepthink-3B-GGUF: Same weights, converted
- MaziyarPanahi/Codepy-Deepthink-3B-GGUF: Same weights, converted
- prithivMLmods/Codepy-Deepthink-3B-GGUF: Same weights, converted
- prithivMLmods/Llama-Deepsync-3B: LoRA-style
- prithivMLmods/Omni-Reasoner3-Merged: LoRA-style
- FBogaerts/CodePy-3B-Finetuned: LoRA-style