Dongwei/DeepSeek-R1-Distill-Qwen-7B-GRPO_Math

Full fine-tune vs deepseek-ai/DeepSeek-R1-Distill-Qwen-7B. Fine-tune: 93% of the weights changed, across 28 of 28 layers.

Tensors changed: Embeddings 0/1, Attention 196/196, MLP 84/84, Norms 49/57, lm_head 1/1

2 derivatives checked

• mradermacher/DeepSeek-R1-Distill-Qwen-7B-GRPO_Math-GGUF: Same weights, converted
• AlejandroOlmedo/DeepSeek-R1-Distill-Qwen-7B-GRPO_Math-8bit-mlx: Quantization

On the Hub · Downloads on Model Pulse