mirror of
https://github.com/ollama/ollama-python.git
synced 2026-07-21 09:05:23 -04:00
Moondream:v2 issue 500 error #212
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @MichielBontenbal on GitHub (Feb 3, 2025).
I get the error message as seen below: ResponseError: POST predict: Post "http://127.0.0.1:61868/completion": EOF (status code: 500)
My settings:
Python 3.12.2
MacOS intel 8Gb RAM with limited free disk space (± 15 Gb)
VS Code with Jupyter Notebook
Latest version of ollama, latest version of ollama-python package
ollama is working correctly from the command line (ollama run moondream:v2 describe this image /CV images datasets/solved/elephant_swimming.jpg gives expected result)
I only get this when running moondream:v2. When I run text models there is no issue.
ResponseError Traceback (most recent call last)
Cell In[7], line 5
1 #source: https://github.com/ollama/ollama-python
2 # WARNING: THIS MIGHT TAKE LONGER THAN A MINUTE DEPENDING ON YOUR DEVICE
3 import ollama
----> 5 res = ollama.chat(
6 model="moondream:v2",
7 messages=[
8 {
9 'role': 'user',
10 'content': 'What is strange about this image?:',
11 'images': [image_path]
12 }
13 ]
14 )
16 print(res['message']['content'])
File /opt/anaconda3/lib/python3.12/site-packages/ollama/_client.py:333, in Client.chat(self, model, messages, tools, stream, format, options, keep_alive)
289 def chat(
290 self,
291 model: str = '',
(...)
298 keep_alive: Optional[Union[float, str]] = None,
299 ) -> Union[ChatResponse, Iterator[ChatResponse]]:
300 """
301 Create a chat response using the requested model.
302
(...)
331 Returns
ChatResponseifstreamisFalse, otherwise returns aChatResponsegenerator.332 """
--> 333 return self._request(
334 ChatResponse,
335 'POST',
336 '/api/chat',
337 json=ChatRequest(
338 model=model,
339 messages=[message for message in _copy_messages(messages)],
340 tools=[tool for tool in _copy_tools(tools)],
341 stream=stream,
342 format=format,
343 options=options,
344 keep_alive=keep_alive,
345 ).model_dump(exclude_none=True),
346 stream=stream,
347 )
File /opt/anaconda3/lib/python3.12/site-packages/ollama/_client.py:178, in Client._request(self, cls, stream, *args, **kwargs)
174 yield cls(**part)
176 return inner()
--> 178 return cls(**self._request_raw(*args, **kwargs).json())
File /opt/anaconda3/lib/python3.12/site-packages/ollama/_client.py:122, in Client._request_raw(self, *args, **kwargs)
120 return r
121 except httpx.HTTPStatusError as e:
--> 122 raise ResponseError(e.response.text, e.response.status_code) from None
123 except httpx.ConnectError:
124 raise ConnectionError(CONNECTION_ERROR_MESSAGE) from None
ResponseError: POST predict: Post "http://127.0.0.1:61868/completion": EOF (status code: 500)
@ParthSareen commented on GitHub (Feb 10, 2025):
Are you port forwarding to get port
61868?Can you also drop the server logs from Ollama, if it's a 500 won't be able to help without that
@jessegross commented on GitHub (Feb 11, 2025):
This is triggered by https://github.com/ollama/ollama/issues/7441