wingchiuloong/OpenPerfectBlend-Qwen3-4B-Regen
OpenPerfectBlend-Qwen3-4B-Regen Overview This dataset is a regenerated version of mlabonne/open-perfectblend in which every assistant turn was re-sampled by Qwen3-4B in no-think (non-thinking) mode (greedy decoding, temperature 0, thinking disabled) with the original user prompts kept intact. It was produced as instruction-following / SFT training data for the DCore project on speculative decoding with draft models, where a target model is fine-tuned to act as the… See the full description on the dataset page: https://huggingface.co/datasets/wingchiuloong/OpenPerfectBlend-Qwen3-4B-Regen.
| Author / organization | wingchiuloong |
|---|---|
| License (as declared) | apache-2.0 |
| Downloads | 236 (as reported by Hugging Face) |
| Likes | 0 |
| Created | 22 October 2026 |
| Last updated at source | 22 October 2026 |
Popularity, last 20 days
236 now (+236 since 2026-10-22). Counts as reported by the source, snapshotted daily by Silytics.
Connected in the knowledge graph
Recent changes
- 22 Oct 2026 License: none → apache-2.0
- 22 Oct 2026 Description: none → OpenPerfectBlend-Qwen3-4B-Regen Overview This dataset is a regenerated version of mlabonne/open-perfectblend in which every assistant turn was re-sampled by Qwen3-4B in no-think (non-thinking) mode (greedy decoding, temperature 0, thinking
- 22 Oct 2026 Topics added task_categories:text-generation, synthetic-data, instruction-tuning, sft, speculative-decoding, qwen3
Changes detected by Silytics between refreshes of the source.
Related papers
More from wingchiuloong
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 11 November 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.