github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 16:11:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #3629] Wrong handling of empty stop-parameter ( on empty array ) via Ollama API #27996

New Issue
Closed
opened 2026-04-22 05:42:44 -05:00 by GiteaMirror · 1 comment
GiteaMirror commented 2026-04-22 05:42:44 -05:00
Owner
Copy Link

Originally created by @TheMasterFX on GitHub (Apr 13, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/3629

What is the issue?

Some implementations of the Ollama API (e.g. Langchain.llm ollama) are always sending an empty "stop": [] instead of "stop": null or just not even sending the stop parameter at all. When "stop": [] is send in the Json request, the default stop Token (set by the Modelfile) is overwritten, leading sometimes to a very long text, end ending with the stop string, e.g. <|im_end|> in the case of dolphin-mistral.

What did you expect to see?

Stop parameter stays untouched.

Steps to reproduce

from langchain.llms import Ollama

lang_client = Ollama(model="dolphin-mistral:latest")
response = lang_client.invoke("Hello")
print(response)

Output: Hello! How can I assist you today?<|im_end|>

Or:
curl -X POST http://localhost:11434/api/generate -d '{ "model": "dolphin-mistral:latest","prompt":"Hello" ,"options": {"stop": []}, "stream": false}'

Are there any recent changes that introduced the issue?

I never saw it working.

Or is it the fault of the 3rd party implementations like langchain?

OS

Windows

Architecture

amd64

Platform

No response

Ollama version

0.1.31

GPU

Nvidia

GPU info

+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 546.01                 Driver Version: 546.01       CUDA Version: 12.3     |
|-----------------------------------------+----------------------+----------------------+
| GPU  Name                     TCC/WDDM  | Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |         Memory-Usage | GPU-Util  Compute M. |
|                                         |                      |               MIG M. |
|=========================================+======================+======================|
|   0  NVIDIA GeForce RTX 3080      WDDM  | 00000000:02:00.0  On |                  N/A |
|  0%   47C    P8              27W / 370W |   6294MiB / 10240MiB |      0%      Default |
|                                         |                      |                  N/A |
+-----------------------------------------+----------------------+----------------------+

CPU

Intel

Other software

No response

Originally created by @TheMasterFX on GitHub (Apr 13, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/3629 ### What is the issue? Some implementations of the Ollama API (e.g. Langchain.llm ollama) are always sending an empty `"stop": []` instead of `"stop": null` or just not even sending the stop parameter at all. When `"stop": []` is send in the Json request, the default stop Token (set by the Modelfile) is overwritten, leading sometimes to a very long text, end ending with the stop string, e.g. <|im_end|> in the case of dolphin-mistral. ### What did you expect to see? Stop parameter stays untouched. ### Steps to reproduce ``` from langchain.llms import Ollama lang_client = Ollama(model="dolphin-mistral:latest") response = lang_client.invoke("Hello") print(response) ``` Output: `Hello! How can I assist you today?<|im_end|>` Or: `curl -X POST http://localhost:11434/api/generate -d '{ "model": "dolphin-mistral:latest","prompt":"Hello" ,"options": {"stop": []}, "stream": false}'` ### Are there any recent changes that introduced the issue? I never saw it working. Or is it the fault of the 3rd party implementations like langchain? ### OS Windows ### Architecture amd64 ### Platform _No response_ ### Ollama version 0.1.31 ### GPU Nvidia ### GPU info ``` +---------------------------------------------------------------------------------------+ | NVIDIA-SMI 546.01 Driver Version: 546.01 CUDA Version: 12.3 | |-----------------------------------------+----------------------+----------------------+ | GPU Name TCC/WDDM | Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. | | | | MIG M. | |=========================================+======================+======================| | 0 NVIDIA GeForce RTX 3080 WDDM | 00000000:02:00.0 On | N/A | | 0% 47C P8 27W / 370W | 6294MiB / 10240MiB | 0% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+ ``` ### CPU Intel ### Other software _No response_
GiteaMirror added the bug label 2026-04-22 05:42:44 -05:00
GiteaMirror commented 2026-04-22 05:42:46 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Jul 18, 2024):

@TheMasterFX Sorry for the slow response. I think this ended up getting fixed a while ago as I'm not able to duplicate it:

% curl -X POST http://localhost:11434/api/generate -d '{ "model": "dolphin-mistral:latest","prompt":"Hello" ,"options": {"stop": []}, "stream": false}'
{"model":"dolphin-mistral:latest","created_at":"2024-07-18T18:56:00.790381Z","response":"Hi! How can I assist you today? What do you need help with?","done":true,"done_reason":"stop","context":[32001,1587,13,1976,460,15052,721,262,28725,264,10865,16107,13892,28723,13,32000,28705,13,32001,2188,13,16230,32000,28705,13,32001,13892,13,23809,28808,1602,541,315,6031,368,3154,28804,1824,511,368,927,1316,395,28804],"total_duration":4645989208,"load_duration":4286676208,"prompt_eval_count":29,"prompt_eval_duration":93417000,"eval_count":17,"eval_duration":264221000}%

I'm going to go ahead and close the issue. LMK if you're still seeing the problem.

<!-- gh-comment-id:2237330406 --> @pdevine commented on GitHub (Jul 18, 2024): @TheMasterFX Sorry for the slow response. I think this ended up getting fixed a while ago as I'm not able to duplicate it: ``` % curl -X POST http://localhost:11434/api/generate -d '{ "model": "dolphin-mistral:latest","prompt":"Hello" ,"options": {"stop": []}, "stream": false}' {"model":"dolphin-mistral:latest","created_at":"2024-07-18T18:56:00.790381Z","response":"Hi! How can I assist you today? What do you need help with?","done":true,"done_reason":"stop","context":[32001,1587,13,1976,460,15052,721,262,28725,264,10865,16107,13892,28723,13,32000,28705,13,32001,2188,13,16230,32000,28705,13,32001,13892,13,23809,28808,1602,541,315,6031,368,3154,28804,1824,511,368,927,1316,395,28804],"total_duration":4645989208,"load_duration":4286676208,"prompt_eval_count":29,"prompt_eval_duration":93417000,"eval_count":17,"eval_duration":264221000}% ``` I'm going to go ahead and close the issue. LMK if you're still seeing the problem.
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 bug
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#27996
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?