Video Cold-Start CoT Dataset (OMNEX-VL format)
Generated 2026-05-31 for the video-hallucination project's Stage-1 cold-start SFT.
Files
coldstart_cot_all.json— 12,262 valid samples (master, post-filter)coldstart_cot_10k.json— 10,000-sample training slice (seed=0 shuffle of master)coldstart_cot.raw.jsonl.gz— every generation attempt + reject reasons (audit)build_coldstart_cot.py/sft_coldstart.py/coldstart_sft.sh— exact code used
Schema (per sample, matches OMNEX-VL sft.py)
{ "problem", "data_type":"video", "path":, "process_and_answer": "...\n...\n...\n...", "meta": {"idx","source","GT"} }
How it was built
- Source: TemporalBench short_qa (9,740) + long_qa (2,522), MCQ with letter GT.
- Teacher: Qwen2.5-VL-72B-Instruct (vLLM, TP=8), video sampled fps=2 / max 64 frames.
- GT given to teacher at gen-time (LLaVA-CoT style), stripped from stored prompt.
- Rejection sampling: 4 tags present AND ==GT; quality filters (answer-leak phrases, min stage length, n-gram repetition). Pass rate 98.2%.