Conversation
|
I tried to fix some bugs in PR 53. |
keli-wen
left a comment
There was a problem hiding this comment.
Great job on this pull request! This is a really well-structured and useful addition.
I have a couple of suggestions for discussion:
-
Built-in vs. Example Flow: I'd like to propose that we discuss whether the
podcast_flowshould be a core, built-in flow or if it would serve a better purpose as a detailed example. My perspective is that it functions as an excellent demonstration of how to customize a personal flow to achieve a specific goal. Positioning it as a showcase in theexamples/directory could be a great way to guide users in creating their own solutions. -
Dedicated
PodcastFlowConfig: To improve flexibility, what are your thoughts on defining a specificPodcastFlowConfig? This would allow us to add more granular controls, such as managing the number of speakers in the podcast (e.g., defaulting to two). This would make the flow more adaptable for different podcast formats.
These are just a couple of points for discussion. Overall, this is excellent work and very close to being ready. Great job
|
Since I sync the latest branch, pull it before dev it. |
@keli-wen Good points of view! |
keli-wen
left a comment
There was a problem hiding this comment.
LGTM, just to confirm, can the podcast flow currently run successfully? (Based on OpenAI API)
Yes. I tested pipeline.py with OpenAI API. It runs successfully. |
Co-authored-by: bridgeqiqi <Bridgeqiqi@163.com> Co-authored-by: LastWhisper <pkuwkl@gmail.com>
Description
Refer to #54
Checklist
Please feel free to remove inapplicable items for your PR.
$CATEGORY(xx): xxx(such asfeat(tool): xxx,fix(source): xxx,docs(README): xxx)