FractalAIResearch/Fathom-R1-14B
Full fine-tune vs deepseek-ai/DeepSeek-R1-Distill-Qwen-14B. Fine-tune: 100% of the weights changed, across 48 of 48 layers.
Tensors changed: Embeddings 1/1, Attention 336/336, MLP 144/144, Norms 97/97, lm_head 1/1
6 derivatives checked
- mradermacher/Ramanujan-Ganit-R1-14B-i1-GGUF: Quantization
- mradermacher/Fathom-R1-14B-i1-GGUF: Quantization
- mradermacher/Fathom-R1-14B-GGUF: Quantization
- DevQuasar/FractalAIResearch.Fathom-R1-14B-GGUF: Same weights, converted
- mradermacher/Ramanujan-Ganit-R1-14B-GGUF: Quantization
- YOYO-AI/YOYO-O1-14B-V2: Full fine-tune