Dataset · Jackrong
DeepSeek reasoning SFT distillation
Jackrong/DeepSeek-V4-Distill-8000x
7,716 JSONL question-answer examples for SFT distillation of reasoning models.
What's inside
A dataset for supervised fine-tuning with reasoning distillation. Questions are taken from Jackrong/GLM-5.1-Reasoning-1M-Cleaned, and answers were generated by the DeepSeek-V4-Flash teacher model. After cleaning, the train split contains 7,716 examples in JSONL format; the language is English.
What you can do with it
- SFT reasoning model
- Distilling DeepSeek-V4-Flash into a smaller model
- Evaluating reasoning quality
- Tasks
- Text generation
- Languages
- English
- Size
- 1K–10K rows
- Downloads this month
- 1.9K
- Likes
- 99
Models trained on it
- Jackrong/Qwen3.5-9B-DeepSeek-V4-Flash-GGUFImage text → text 38.1K
- mradermacher/DeepSeek-V4-Pro-Qwen3.5-9B-GGUF 22.2K
- mradermacher/DeepSeek-V4-Pro-Qwen3.5-4B-i1-GGUF 8.1K
- mradermacher/DeepSeek-V4-Pro-Qwen3.5-4B-GGUF 7.6K
- mradermacher/DeepSeek-V4-Pro-Qwen3.5-9B-i1-GGUF 4.8K
- cybertruck32489/LFM3-1.7B-RDText generation 3.9K
- mradermacher/Q3.6-27B-DS-v4-Flash-DA-i1-GGUF 2.3K
- mradermacher/Qwen3-8B-3MPER0RR-BF16-DeepSeek-V4-Flash-i1-GGUF 2.2K
How to load it
from datasets import load_dataset
ds = load_dataset("Jackrong/DeepSeek-V4-Distill-8000x")