Skip to content

docs: correct OpenAI vision request count #117

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
kstonekuan merged 1 commit into from
Aug 23, 2026
Merged

docs: correct OpenAI vision request count #117

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 5 additions & 5 deletions docs/how-to/call-openai-vision.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -64,7 +64,7 @@ same file can point at a different endpoint without editing it. Inside the
step, the endpoint remains an ordinary OpenAI SDK option:

```python
@app.check(uses="vision")
@app.check(uses="vision", version="responses-contact-sheet-v1")
def describe_activity(episode: hflow.Episode) -> hflow.CheckResult:
client = OpenAI(
api_key=os.environ["OPENAI_API_KEY"],
Expand Down Expand Up @@ -92,10 +92,10 @@ uv run --extra openai python examples/openai_vision/pipeline.py
```

Pass an MCAP path to use your own episode. The example samples frames at 0.5
FPS, caps the request at 12 tiles, and makes one model request per episode. Those
choices are part of the measurement definition: change them deliberately and
version the step explicitly when external model configuration cannot be hashed
from the function.
FPS, caps each request at 12 tiles, and makes one model request per check: two
requests per episode for the two checks it ships. Those choices are part of the
measurement definition: change them deliberately and version the step explicitly
when external model configuration cannot be hashed from the function.

## Keep model output as evidence

Expand Down
Loading