eccv-experiments
Use when designing or auditing the experimental program of an ECCV paper — benchmark selection that survives a September conference, matched-substrate baseline fairness in the foundation-model era, ablations that isolate the claimed mechanism, qualitative failure evidence, and run sequencing toward a March freeze.
Profile →npx skills add brycewang-stanford/Awesome-Journal-Skills --skill eccv-experiments --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
# ECCV Experiments Use this while the experimental plan is still changeable. ECCV's calendar shapes the evidence problem: results freeze in early March, reviews weigh them in May against everything published since, and the field first reads the paper at a **September** conference — the numbers must still look current six months after the freeze. ## The six-month-staleness test For each headline table, ask: *if the strongest lab in this niche publishes their CVPR camera-ready in June, does this table still support the claim in September?* Evidence that passes: mechanism-isolating ablations, efficiency frontiers (accuracy vs compute), and generality sweeps across datasets. Evidence that fails: a raw leaderboard number 0.2 points above a moving SOTA. Build the paper's claim on the first kind and let the leaderboard row be corroboration, not the thesis. ## Matched-substrate fairness The first thing a 2026-era vision reviewer checks is whether wins come from the method or from what it was fed: | Axis to match | Unfair pattern | Fair protocol | |---|---|---| | Backbone / pretraining | Your ViT-L vs their ResNet-50 numbers | Re-run the top baselines on your backbone, or add a matched-back
- The six-month-staleness test
- Matched-substrate fairness
- Ablations that isolate, not decorate
- Qualitative evidence discipline
- Run sequencing toward March 5
- Output format
What does the eccv-experiments skill do?
Use when designing or auditing the experimental program of an ECCV paper — benchmark selection that survives a September conference, matched-substrate baseline fairness in the foundation-model era, ablations that isolate the claimed mechanism, qualitative failure evidence, and run sequencing toward a March freeze.
How do I install it?
Run `npx skills add brycewang-stanford/Awesome-Journal-Skills --skill eccv-experiments --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From brycewang-stanford/Awesome-Journal-Skills, a repository with 909 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.