gpt-pilot
gpt-pilot copied to clipboard
Pythagora-io
[Bug]: Project gets stuck on Implementation stage 5
Version
Command-line (Python) version
Operating System
MacOS
What happened?
First off, let me say that i really like the idea of the tool and it somewhat works really good with enough guidance. Im sure it will only get better.
The execution always stops at Implementing task #5: xxxx, regardless of what the feature is. There is no stack trace, it just gets stuck. I've tried a few different things:
-
When restarting with python pilot/main.py app_id=fde4f6c2-84a6-4662-940b-3b2a4b030ed4 skip_until_dev_step=44, and skip over the steps it always results in getting stuck at the same place
-
I tried with both gpt-4-1106-preview and gpt-4 gpt-4-1106-preview = I've seen the unknown host error more often, the read timeout might have happened due to network interruption. There was a problem with request to openai API: HTTPSConnectionPool(host='api.openai.com', port=443): Read timed out. (read timeout=None) or HTTPSConnectionPool(host='api.openai.com', port=443): Unknown host. (read timeout=None)
gpt-4 = Basically nothing happens for hours.
- I edited the development plan in the DB( development_planning.development_plan) where i removed the step it was getting stuck on. Now it gets stuck on the next step, still at Implementing task #5: xxxx, .
When I interrupt the process, there's only a stack trace for the questionary.py exit.
You can find the SQLLite db & workspace folder attached(hopefully). Let me know if you cant open them, or if you need further assistance in replicating. gpt-pilot.zip workspace.zip
Same, mine keeps stopping on my Implementing task #3 no amount of restarts seems to get it to do anything
Edit: looks like mine is erroring out like in https://github.com/Pythagora-io/gpt-pilot/issues/269
Request too large for gpt-4-1106-preview in organization org-XXXXXXXXXX\
\ on tokens per min (TPM): Limit 150000, Requested 557819. The input or output tokens must be reduced in order to run successfully.
