Closed cmungall closed 5 months ago
possibly out of scope for this PR, but it looks GH actions is using python 3.8 here when the project requires python 3.9 or better
With some of these models there may be no response, which causes errors.
I got the following with the example question and collection, but with the phi
model through ollama
:
DEBUG:urllib3.connectionpool:Starting new HTTP connection (1): 0.0.0.0:8000
DEBUG:urllib3.connectionpool:http://0.0.0.0:8000 "POST /chat/completions HTTP/1.1" 200 None
DEBUG:openai:message='OpenAI API response' path=http://0.0.0.0:8000/chat/completions processing_ms=None request_id=None response_code=200
Traceback (most recent call last):
File "/home/harry/curate-gpt/.venv/bin/curategpt", line 6, in <module>
sys.exit(main())
File "/home/harry/curate-gpt/.venv/lib/python3.10/site-packages/click/core.py", line 1157, in __call__
return self.main(*args, **kwargs)
File "/home/harry/curate-gpt/.venv/lib/python3.10/site-packages/click/core.py", line 1078, in main
rv = self.invoke(ctx)
File "/home/harry/curate-gpt/.venv/lib/python3.10/site-packages/click/core.py", line 1688, in invoke
return _process_result(sub_ctx.command.invoke(sub_ctx))
File "/home/harry/curate-gpt/.venv/lib/python3.10/site-packages/click/core.py", line 1434, in invoke
return ctx.invoke(self.callback, **ctx.params)
File "/home/harry/curate-gpt/.venv/lib/python3.10/site-packages/click/core.py", line 783, in invoke
return __callback(*args, **kwargs)
File "/home/harry/curate-gpt/src/curate_gpt/cli.py", line 1250, in ask
response = chatbot.chat(query, collection=collection, conversation=conversation)
File "/home/harry/curate-gpt/src/curate_gpt/agents/chat_agent.py", line 143, in chat
response_text = response.text()
File "/home/harry/curate-gpt/.venv/lib/python3.10/site-packages/llm/models.py", line 112, in text
self._force()
File "/home/harry/curate-gpt/.venv/lib/python3.10/site-packages/llm/models.py", line 106, in _force
list(self)
File "/home/harry/curate-gpt/.venv/lib/python3.10/site-packages/llm/models.py", line 91, in __iter__
for chunk in self.model.execute(
File "/home/harry/curate-gpt/.venv/lib/python3.10/site-packages/llm/default_plugins/openai_models.py", line 292, in execute
response.response_json = combine_chunks(chunks)
File "/home/harry/curate-gpt/.venv/lib/python3.10/site-packages/llm/default_plugins/openai_models.py", line 413, in combine_chunks
content += choice["delta"]["content"]
TypeError: can only concatenate str (not "NoneType") to str
Just need a fix to make that more obvious. Preferences on whether to put that in its own PR?
I get this same issue with mistral-7B
I think the ollama run
command may not be necessary if the ollama server is already running though
See also #15