Dataset · Congliu
Chinese DeepSeek R1 distillate at 110k examples
Congliu/Chinese-DeepSeek-R1-Distill-data-110k
110k Chinese examples with answers and R1 reasoning traces for SFT.
What's inside
The dataset contains 110k Chinese examples distilled from full DeepSeek-R1. Inside is not only mathematics, but also plenty of general-purpose data. There is a version for direct SFT: thoughts and the answer are combined into an output field so that most SFT frameworks can load the data. Language is Chinese, modalities are tabular and text.
What you can do with it
- SFT of a small Chinese model on R1 reasoning traces
- Fine-tuning LoRA for general chat and math
- Replicating R1 distillation on Chinese data
- Tasks
- Text generation, Question answering
- Languages
- Chinese
- Size
- 100K–1M rows
- Downloads this month
- 2.6K
- Likes
- 804
Models trained on it
- mradermacher/Zhi-writing-dsr1-14b-GGUF 1.9K
- mradermacher/Zhi-Create-Qwen3-32B-i1-GGUF 1.7K
- mradermacher/Zhi-Create-DSR1-14B-i1-GGUF 1.4K
- mradermacher/Zhi-writing-dsr1-14b-i1-GGUF 978
- mradermacher/Zhi-Create-Qwen3-32B-GGUF 452
- mradermacher/Zhi-Create-DSR1-14B-GGUF 388
- mradermacher/DeepSeek-R1-Psychology-COT-GGUF 179
- mradermacher/NQLSG-Qwen2.5-14B-MegaFusion-v4-reasoning-GGUF 139
How to load it
from datasets import load_dataset
ds = load_dataset("Congliu/Chinese-DeepSeek-R1-Distill-data-110k")