GPT response Issue

[marcelklopper]

I am Having the following issue:

Got 7 frames from video
Process-90: prompting GPT-4v
Process-89: GPT response {'id': 'chatcmpl-8VF2rf3HoYVZZfy3Wvwk8OJbcyOMi', 'object': 'chat.completion', 'created': 1702456693, 'model': 'gpt-4-1106-vision-preview', 'usage': {'prompt_tokens': 283, 'completion_tokens': 135, 'total_tokens': 418}, 'choices': [{'message': {'role': 'assistant', 'content': "It appears that you've shared a series of images from a security camera or monitoring system showing a view of what looks like a residential area. The images are time-stamped, indicating that they were taken within seconds of each other. There's a patterned paved driveway or courtyard with trees and a wooden fence at the far end. \n\nIn the second image, a person carrying a bag or bucket enters the frame from the right side. In the third image, the same person continues to walk across the area, still carrying the item, and a small dog is visible following them in the background. It looks like a regular day-to-day moment captured on a home security system."}, 'finish_details': {'type': 'stop', 'stop': '<|fim_suffix|>'}, 'index': 0}]}
Process-89: Error processing video for event 1702456682.852495-onmarj
Traceback (most recent call last):
File "/app/./mqtt_client.py", line 304, in process_message
result = json.loads(json_str)
^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.11/json/init.py", line 346, in loads
return _default_decoder.decode(s)
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.11/json/decoder.py", line 337, in decode
obj, end = self.raw_decode(s, idx=_w(s, 0).end())
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.11/json/decoder.py", line 355, in raw_decode
raise JSONDecodeError("Expecting value", s, err.value) from None
json.decoder.JSONDecodeError: Expecting value: line 1 column 1 (char 0)

I think this is probably because I am using gpt-4-1106-vision-preview and not gpt-4-vision-preview

I can't seem to find the gpt-4-vision-preview model on my openAI account or at least I don't have access to it.

I will also try see what I can do from my end but any help is much appreciated.

[mhaowork]

@marcelklopper Sorry for the late reply. I have been on a vacation recently. The program expects a valid json in OpenAI API response. Did you customize the prompt? If not, this should be a transient error. Otherwise, please make sure your prompt has instructions for the model to return json.

[marcelklopper]

No worries, hope you have a cheerful festive season. I am also away atm. I have not changed the prompt from the default. I will try play with a custom prompt to see if it helps to get a valid json response. will keep you posted.

[ttopholm]

@marcelklopper if you have used the default config, it contains a prompt, try delete that.