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 #1135] json response stalls? #47081

New Issue
Closed
opened 2026-04-28 02:59:58 -05:00 by GiteaMirror · 11 comments
GiteaMirror commented 2026-04-28 02:59:58 -05:00
Owner
Copy Link

Originally created by @hemanth on GitHub (Nov 15, 2023).
Original GitHub issue: https://github.com/ollama/ollama/issues/1135

/tmp 
❯ ollama --version
ollama version 0.1.9

https://github.com/jmorganca/ollama/assets/18315/d0d8ecb1-142f-464c-bb49-8d147eb3d322

Sometimes we see an empty response:

{"model":"llama2","created_at":"2023-11-15T05:46:21.685664Z","response":"{} ","done":true,"context":[29961,25580,29962,3532,14816,29903,29958,5299,829,14816,29903,6778,13,13,29911,514,592,263,270,328,2212,446,518,29914,25580,29962,6571,29871],"total_duration":216306917,"load_duration":982333,"prompt_eval_count":1,"eval_count":3,"eval_duration":192199000}
Originally created by @hemanth on GitHub (Nov 15, 2023). Original GitHub issue: https://github.com/ollama/ollama/issues/1135 ``` /tmp ❯ ollama --version ollama version 0.1.9 ``` https://github.com/jmorganca/ollama/assets/18315/d0d8ecb1-142f-464c-bb49-8d147eb3d322 Sometimes we see an empty response: ```json {"model":"llama2","created_at":"2023-11-15T05:46:21.685664Z","response":"{} ","done":true,"context":[29961,25580,29962,3532,14816,29903,29958,5299,829,14816,29903,6778,13,13,29911,514,592,263,270,328,2212,446,518,29914,25580,29962,6571,29871],"total_duration":216306917,"load_duration":982333,"prompt_eval_count":1,"eval_count":3,"eval_duration":192199000} ```
GiteaMirror added the bug label 2026-04-28 02:59:59 -05:00
GiteaMirror commented 2026-04-28 03:00:02 -05:00
Author
Owner
Copy Link

@jmorganca commented on GitHub (Nov 15, 2023):

Hi @hemanth ! Thanks for the issue. would it be possible to share an example prompt that causes this? Also I think your video may be of another project 😊

<!-- gh-comment-id:1812389341 --> @jmorganca commented on GitHub (Nov 15, 2023): Hi @hemanth ! Thanks for the issue. would it be possible to share an example prompt that causes this? Also I think your video may be of another project 😊
GiteaMirror commented 2026-04-28 03:00:04 -05:00
Author
Owner
Copy Link

@technovangelist commented on GitHub (Dec 4, 2023):

@hemanth can you provide an example of this?

<!-- gh-comment-id:1839707566 --> @technovangelist commented on GitHub (Dec 4, 2023): @hemanth can you provide an example of this?
GiteaMirror commented 2026-04-28 03:00:07 -05:00
Author
Owner
Copy Link

@hemanth commented on GitHub (Dec 5, 2023):

Just the video. I didn't change any prompt.

@technovangelist @jmorganca

image
<!-- gh-comment-id:1839768365 --> @hemanth commented on GitHub (Dec 5, 2023): Just the video. I didn't change any prompt. @technovangelist @jmorganca <img width="764" alt="image" src="https://github.com/jmorganca/ollama/assets/18315/7ca82fb2-63fb-4024-9c03-8af08026c86a">
GiteaMirror commented 2026-04-28 03:00:09 -05:00
Author
Owner
Copy Link

@technovangelist commented on GitHub (Dec 8, 2023):

But the video and this screenshot don't show anything having to do with ollama. Is there anything you can show that demonstrates the problem. We would love to solve any issues you have. But we need a little info about what is the problem and how did you get there.

<!-- gh-comment-id:1847973496 --> @technovangelist commented on GitHub (Dec 8, 2023): But the video and this screenshot don't show anything having to do with ollama. Is there anything you can show that demonstrates the problem. We would love to solve any issues you have. But we need a little info about what is the problem and how did you get there.
GiteaMirror commented 2026-04-28 03:00:15 -05:00
Author
Owner
Copy Link

@hemanth commented on GitHub (Dec 9, 2023):

Ah, sorry, my bad! [/me had mapped this issue mentally as the create-llama ticket :D]

Here is the correct recording...

https://github.com/jmorganca/ollama/assets/18315/b900d6ef-423d-4674-88ed-3b94ccfa2a7c

<!-- gh-comment-id:1848612267 --> @hemanth commented on GitHub (Dec 9, 2023): Ah, sorry, my bad! [/me had mapped this issue mentally as the create-llama ticket :D] Here is the correct recording... https://github.com/jmorganca/ollama/assets/18315/b900d6ef-423d-4674-88ed-3b94ccfa2a7c
GiteaMirror commented 2026-04-28 03:00:18 -05:00
Author
Owner
Copy Link

@atorr0 commented on GitHub (Jan 1, 2024):

Hi @hemanth , did you try it with newer versions (like https://github.com/jmorganca/ollama/releases/tag/v0.1.17)?

<!-- gh-comment-id:1873421004 --> @atorr0 commented on GitHub (Jan 1, 2024): Hi @hemanth , did you try it with newer versions (like https://github.com/jmorganca/ollama/releases/tag/v0.1.17)?
GiteaMirror commented 2026-04-28 03:00:19 -05:00
Author
Owner
Copy Link

@rmallick6806 commented on GitHub (Jan 1, 2024):

Running into the same issue and I'm using v0.0.17. JSON format mode always hangs. Without the flag, I am able to get responses.

<!-- gh-comment-id:1873447428 --> @rmallick6806 commented on GitHub (Jan 1, 2024): Running into the same issue and I'm using v0.0.17. JSON format mode always hangs. Without the flag, I am able to get responses.
GiteaMirror commented 2026-04-28 03:00:24 -05:00
Author
Owner
Copy Link

@technovangelist commented on GitHub (Jan 3, 2024):

@hemanth Have you tried with the latest version? What response are you getting?

@rmallick6806 can you tell me more about what you are trying and the result. what platform are you on?

<!-- gh-comment-id:1875686856 --> @technovangelist commented on GitHub (Jan 3, 2024): @hemanth Have you tried with the latest version? What response are you getting? @rmallick6806 can you tell me more about what you are trying and the result. what platform are you on?
GiteaMirror commented 2026-04-28 03:00:28 -05:00
Author
Owner
Copy Link

@horiacristescu commented on GitHub (Jan 6, 2024):

Same problem, using mistral:7b-instruct-v0.2-q5_K_M but it happens for any model in my experience.

I tried with v0.1.18 and for every --format json request it doesn't stop. After printing the JSON it continues to print empty lines forever. Because it never stops printing empty lines, it is as if it hangs forever.

Does the JSON format model know when to stop? What stop words should I use? I tried stop="\n\n\n" without success.

<!-- gh-comment-id:1879635186 --> @horiacristescu commented on GitHub (Jan 6, 2024): Same problem, using `mistral:7b-instruct-v0.2-q5_K_M` but it happens for any model in my experience. I tried with v0.1.18 and for every --format json request it doesn't stop. After printing the JSON it continues to print empty lines forever. Because it never stops printing empty lines, it is as if it hangs forever. Does the JSON format model know when to stop? What stop words should I use? I tried stop="\n\n\n" without success.
GiteaMirror commented 2026-04-28 03:00:30 -05:00
Author
Owner
Copy Link

@seanmavley commented on GitHub (Mar 9, 2024):

Related to #1910

<!-- gh-comment-id:1986843050 --> @seanmavley commented on GitHub (Mar 9, 2024): Related to #1910
GiteaMirror commented 2026-04-28 03:00:31 -05:00
Author
Owner
Copy Link

@BruceMacD commented on GitHub (Mar 11, 2024):

Hi all, this is an ongoing issue with how the response is mapped to JSON. If you see this issue please try adding a system prompt that specifies a JSON response is expected.

For example in the API:

curl http://127.0.0.1:11434/api/chat -d '{
    "model": "llama2",
    "format": "json",
    "stream": false,
    "messages": [
        {
            "role": "system",
            "content": "Respond only in JSON."
        },
        {
            "role": "user",
            "content": "Give me a list of emojis to descriptions."
        }
    ]
}'

Resolving this issue to consolidate it with #1182 , which I will continue work in.

<!-- gh-comment-id:1989189585 --> @BruceMacD commented on GitHub (Mar 11, 2024): Hi all, this is an ongoing issue with how the response is mapped to JSON. If you see this issue please try adding a system prompt that specifies a JSON response is expected. For example in the API: ```bash curl http://127.0.0.1:11434/api/chat -d '{ "model": "llama2", "format": "json", "stream": false, "messages": [ { "role": "system", "content": "Respond only in JSON." }, { "role": "user", "content": "Give me a list of emojis to descriptions." } ] }' ``` Resolving this issue to consolidate it with #1182 , which I will continue work in.
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#47081
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?