github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 00:22:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #3851] Why Ollama is so terribly slow when I set format="json" #48897

New Issue
Open
opened 2026-04-28 09:59:30 -05:00 by GiteaMirror · 8 comments
GiteaMirror commented 2026-04-28 09:59:30 -05:00
Owner
Copy Link

Originally created by @marksalpeter on GitHub (Apr 23, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/3851

What is the issue?

This is a duplicate of #3154, which was closed, I'm assuming, by mistake.
The performance of the format="json" param is 10x slower than regular inference when additional context is included

A prompt like this takes ~24s to return on an NVIDIA T4 with CUDA enabled and format="json". The same exact prompt without format json takes ~2s to return. This has got to be a bug right?

Use the following pieces of context to answer the question at the end. If you don't know the answer, just say that you don't know, don't try to make up an answer.

${context}

Please respond in the following JSON schema
{
   "${schema.fieldName}": {
      "type": ${schema.type},
      "description": ${schema.description}
     }
}

Question: ${schema.description}
Helpful Answer:

OS

Linux

GPU

Nvidia

CPU

Intel

Ollama version

0.1.32

Originally created by @marksalpeter on GitHub (Apr 23, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/3851 ### What is the issue? This is a duplicate of #3154, which was closed, I'm assuming, by mistake. The performance of the `format="json"` param is 10x slower than regular inference when additional context is included A prompt like this takes ~24s to return on an NVIDIA T4 with CUDA enabled and ` format="json"`. The same exact prompt without format json takes ~2s to return. This has got to be a bug right? ``` Use the following pieces of context to answer the question at the end. If you don't know the answer, just say that you don't know, don't try to make up an answer. ${context} Please respond in the following JSON schema { "${schema.fieldName}": { "type": ${schema.type}, "description": ${schema.description} } } Question: ${schema.description} Helpful Answer: ``` ### OS Linux ### GPU Nvidia ### CPU Intel ### Ollama version 0.1.32
GiteaMirror added the performancebugapi labels 2026-04-28 09:59:31 -05:00
GiteaMirror commented 2026-04-28 09:59:34 -05:00
Author
Owner
Copy Link

@sebdg commented on GitHub (Apr 23, 2024):

I think this is related to a loop detection piece of code in the server.go code, the detection code allows the model to cycle for a number of tokens over whitespace, if the last token is repeated a number of times or only whitespace is detected for like 30 iterations it aborts the prediction, it would be useful if you could run your request while setting ollama in debug mode more about this here : troubleshooting

if this relates to the loop detection logic you will a line like 'prediction aborted, token repeat limit reached' in the log.
on the other side some other bugs on stop detection relate to llama.cpp and could also be a cause.

I would be useful to try your request with streaming enabled this will show what the model returns

curl http://127.0.0.1:11434/api/generate -d '{ 
   "model": "llama3:8b", 
   "prompt": "You are a helpful writer, respond with an address in the US in JSON format.", 
   "stream": true, "format": "json" }'

Now as a workaround I would recommend on not using the format=json for now and just mention it in the prompt itself, depending on what your integration is you might be better by capturing the json part of the response by using a regex orso, I've had some flaws and inconsistent behaviors using format=json across different model, the regex might be a more robust solution to this

<!-- gh-comment-id:2073507410 --> @sebdg commented on GitHub (Apr 23, 2024): I think this is related to a loop detection piece of code in the server.go code, the detection code allows the model to cycle for a number of tokens over whitespace, if the last token is repeated a number of times or only whitespace is detected for like 30 iterations it aborts the prediction, it would be useful if you could run your request while setting ollama in debug mode more about this here : [troubleshooting](https://github.com/ollama/ollama/blob/main/docs/troubleshooting.md#how-to-troubleshoot-issues) if this relates to the loop detection logic you will a line like 'prediction aborted, token repeat limit reached' in the log. on the other side some other bugs on stop detection relate to llama.cpp and could also be a cause. I would be useful to try your request with streaming enabled this will show what the model returns ```sh curl http://127.0.0.1:11434/api/generate -d '{ "model": "llama3:8b", "prompt": "You are a helpful writer, respond with an address in the US in JSON format.", "stream": true, "format": "json" }' ``` Now as a workaround I would recommend on not using the format=json for now and just mention it in the prompt itself, depending on what your integration is you might be better by capturing the json part of the response by using a regex orso, I've had some flaws and inconsistent behaviors using format=json across different model, the regex might be a more robust solution to this
GiteaMirror commented 2026-04-28 09:59:35 -05:00
Author
Owner
Copy Link

@coder543 commented on GitHub (Apr 24, 2024):

I will say... I've observed that some models are slower with json mode than others. I'm not sure if it is a bug in the implementation, or if the models themselves are just trained in interesting ways.

Observing the streaming response, it seems to respond quickly, but then it waits around for awhile before deciding the message is complete. A well-defined grammar would realize that the JSON message is over, and immediately terminate, rather than waiting for some kind of end-of-stream token, and this could be the issue here.

<!-- gh-comment-id:2073883100 --> @coder543 commented on GitHub (Apr 24, 2024): I will say... I've observed that some models [are slower](https://github.com/open-webui/open-webui/discussions/1692#discussioncomment-9196102) with json mode than others. I'm not sure if it is a bug in the implementation, or if the models themselves are just trained in interesting ways. Observing the streaming response, it seems to respond quickly, but then it waits around for awhile before deciding the message is complete. A well-defined grammar would realize that the JSON message is over, and immediately terminate, rather than waiting for some kind of end-of-stream token, and this could be the issue here.
GiteaMirror commented 2026-04-28 09:59:37 -05:00
Author
Owner
Copy Link

@not-nullptr commented on GitHub (Apr 24, 2024):

getting this exact issue on llama3:8b but not with mistral:latest, weirdly enough? speeds for regular text between both models are the exact same on my 3080. i think @coder543 is correct, except this is a bug in the implementation. why are we outputting whitespace in the json in the first place?

<!-- gh-comment-id:2075771062 --> @not-nullptr commented on GitHub (Apr 24, 2024): getting this exact issue on `llama3:8b` but not with `mistral:latest`, weirdly enough? speeds for regular text between both models are the exact same on my 3080. i think @coder543 is correct, except this is a bug in the implementation. why are we outputting whitespace in the json in the first place?
GiteaMirror commented 2026-04-28 09:59:38 -05:00
Author
Owner
Copy Link

@not-nullptr commented on GitHub (Apr 25, 2024):

https://github.com/ollama/ollama/assets/62841684/a91cb579-4160-445d-ad47-caf888f17a39

https://github.com/ollama/ollama/assets/62841684/fbc5a9b4-0113-4d2a-8467-5b24083433f7

the first video demonstrates my function calling without "format": "json", and the second demonstrates it with "format": "json". you can see the speed difference is insane; same prompt and everything.

<!-- gh-comment-id:2076577295 --> @not-nullptr commented on GitHub (Apr 25, 2024): https://github.com/ollama/ollama/assets/62841684/a91cb579-4160-445d-ad47-caf888f17a39 https://github.com/ollama/ollama/assets/62841684/fbc5a9b4-0113-4d2a-8467-5b24083433f7 the first video demonstrates my function calling without `"format": "json"`, and the second demonstrates it with `"format": "json"`. you can see the speed difference is insane; same prompt and everything.
GiteaMirror commented 2026-04-28 09:59:39 -05:00
Author
Owner
Copy Link

@coder543 commented on GitHub (Apr 25, 2024):

Unfortunately your video isn’t visually showing it generating JSON in both modes. If the model can’t respond with the correct JSON without JSON mode at least some of the time, it makes it harder to know for sure where the issue is in a piece of software like this.

It would also be helpful if the response were streaming (with a clear visual indication of when the streaming response has finished) so we could see if it pauses after generating the JSON, or if it is just generating the JSON really slowly character by character.

I would definitely like this slow JSON situation to be fixed.

<!-- gh-comment-id:2076958040 --> @coder543 commented on GitHub (Apr 25, 2024): Unfortunately your video isn’t visually showing it generating JSON in both modes. If the model can’t respond with the correct JSON without JSON mode at least some of the time, it makes it harder to know for sure where the issue is in a piece of software like this. It would also be helpful if the response were streaming (with a clear visual indication of when the streaming response has finished) so we could see if it pauses after generating the JSON, or if it is just generating the JSON really slowly character by character. I would definitely like this slow JSON situation to be fixed.
GiteaMirror commented 2026-04-28 09:59:39 -05:00
Author
Owner
Copy Link

@mitar commented on GitHub (May 29, 2024):

There are some known upstream issues with grammar restrictions (which JSON format uses): https://github.com/ggerganov/llama.cpp/issues/4218

<!-- gh-comment-id:2138275507 --> @mitar commented on GitHub (May 29, 2024): There are some known upstream issues with grammar restrictions (which JSON format uses): https://github.com/ggerganov/llama.cpp/issues/4218
GiteaMirror commented 2026-04-28 09:59:40 -05:00
Author
Owner
Copy Link

@ZeyBal commented on GitHub (Nov 25, 2025):

Hello! Is it fixed?

<!-- gh-comment-id:3576239629 --> @ZeyBal commented on GitHub (Nov 25, 2025): Hello! Is it fixed?
GiteaMirror commented 2026-04-28 09:59:41 -05:00
Author
Owner
Copy Link

@willson556 commented on GitHub (Feb 10, 2026):

It seems like llama.cpp has moved forward in supporting faster grammar restrictions (https://github.com/ggml-org/llama.cpp/pull/10224 for example has merged) but it doesn't seem like that's accessible through Ollama yet.

<!-- gh-comment-id:3880083831 --> @willson556 commented on GitHub (Feb 10, 2026): It seems like llama.cpp has moved forward in supporting faster grammar restrictions (https://github.com/ggml-org/llama.cpp/pull/10224 for example has merged) but it doesn't seem like that's accessible through Ollama yet.
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 api bug performance
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#48897
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?