github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 16:40:08 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #9813] Properly Interrupting Inference in Ollama #32182

New Issue
Closed
opened 2026-04-22 13:13:26 -05:00 by GiteaMirror · 4 comments
GiteaMirror commented 2026-04-22 13:13:26 -05:00
Owner
Copy Link

Originally created by @20246688 on GitHub (Mar 17, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/9813

How can I properly interrupt the inference in Ollama when I notice the output doesn't match the expected content during streaming?

Originally created by @20246688 on GitHub (Mar 17, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/9813 How can I properly interrupt the inference in Ollama when I notice the output doesn't match the expected content during streaming?
GiteaMirror added the feature request label 2026-04-22 13:13:26 -05:00
GiteaMirror commented 2026-04-22 13:13:28 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Mar 17, 2025):

Close the connection.

<!-- gh-comment-id:2729021203 --> @rick-github commented on GitHub (Mar 17, 2025): Close the connection.
GiteaMirror commented 2026-04-22 13:13:28 -05:00
Author
Owner
Copy Link

@20246688 commented on GitHub (Mar 18, 2025):

Close the connection.

Do you mean client._client.aclose()?

<!-- gh-comment-id:2731310179 --> @20246688 commented on GitHub (Mar 18, 2025): > Close the connection. Do you mean client._client.aclose()?
GiteaMirror commented 2026-04-22 13:13:30 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Mar 18, 2025):

Depends on your client. If you are using curl, then ^C and stop the program, thereby closing the connection. If you are using the ollama python client, close the connection:

#!/usr/bin/env python3

import ollama

response = ollama.Client().chat(model='deepseek-r1:7b', messages=[{"role":"user","content":"2+2=?"}], stream=True)
i = 0
for r in response:
  c = r['message']['content']
  if i == 0 and c != '<think>':
    response.close()
    print("reasoning model didn't start with <think>")
    break
  print(c, end='', flush=True)
  i += 1
print()

Whatever your client uses, closing the connection will tell the ollama server that the client no longer needs the output and the server will stop the inference.

<!-- gh-comment-id:2731353450 --> @rick-github commented on GitHub (Mar 18, 2025): Depends on your client. If you are using curl, then ^C and stop the program, thereby closing the connection. If you are using the ollama python client, close the connection: ```python #!/usr/bin/env python3 import ollama response = ollama.Client().chat(model='deepseek-r1:7b', messages=[{"role":"user","content":"2+2=?"}], stream=True) i = 0 for r in response: c = r['message']['content'] if i == 0 and c != '<think>': response.close() print("reasoning model didn't start with <think>") break print(c, end='', flush=True) i += 1 print() ``` Whatever your client uses, closing the connection will tell the ollama server that the client no longer needs the output and the server will stop the inference.
GiteaMirror commented 2026-04-22 13:13:31 -05:00
Author
Owner
Copy Link

@20246688 commented on GitHub (Mar 18, 2025):

For example：

async def async_chat(message, stop_event):
    client = None
    try:
        client = AsyncClient(host=HOST)
        response = await client.chat(model=MODEL, messages=[{'role': 'user', 'content': message}], stream=True)
       
        complete_message = ''
        is_first_content = True

        async for part in response:
            if stop_event.is_set():
                return 'Error'

            content = part['message']['content']

            if is_first_content:
                if not content.startswith(('E'):
                    complete_message = 'Error'
                    print(complete_message)
                    stop_event.set()
                    return complete_message

                is_first_content = False

            print(content, end='', flush=True)
            complete_message += content
       
        return complete_message

    except Exception as e:
        return 'Error'
   
    finally:
        if client:
            await client._client.aclose()

I’m currently using client._client.aclose() to close the connection, but I’m not sure if it’s the correct way to do it.

<!-- gh-comment-id:2731387274 --> @20246688 commented on GitHub (Mar 18, 2025): For example： async def async_chat(message, stop_event):     client = None     try:         client = AsyncClient(host=HOST)         response = await client.chat(model=MODEL, messages=[{'role': 'user', 'content': message}], stream=True)                 complete_message = ''         is_first_content = True         async for part in response:             if stop_event.is_set():                 return 'Error'             content = part['message']['content']             if is_first_content:                 if not content.startswith(('E'):                     complete_message = 'Error'                     print(complete_message)                     stop_event.set()                     return complete_message                 is_first_content = False             print(content, end='', flush=True)             complete_message += content                 return complete_message     except Exception as e:         return 'Error'         finally:         if client:             await client._client.aclose() I’m currently using client._client.aclose() to close the connection, but I’m not sure if it’s the correct way to do it.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label feature request
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#32182
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?