Fix "<|channel|>analysis" appearing when chatting via llama-server API #15124
+1
−1
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
tarruda
force-pushed
the
fix-gpt-oss-reasoning-parse
branch
from
606ea13 to
896f2e6
Compare
This also requires patching the Jinja2 template:
```diff
--- gpt-oss.j2.old 2025-08-06 10:59:18
+++ gpt-oss.j2.new 2025-08-06 11:00:43
@@ -303,5 +303,5 @@
{#- Generation prompt #}
{%- if add_generation_prompt -%}
-<|start|>assistant
+<|start|>assistant<|channel|>analysis
{%- endif -%}
```
tarruda
force-pushed
the
fix-gpt-oss-reasoning-parse
branch
from
896f2e6 to
34786fe
Compare
tarruda
changed the title
tarruda
changed the title
|
I had the same issue (using https://huggingface.co/unsloth/gpt-oss-120b-GGUF with --jinja) and can confirm this fixed it for me. |
Collaborator
|
I would prefer a fix without changing jinja template. I think we should simply extend the builder class to accept a sequence of strings instead of a single string, something like this: One of the goals of having a jinja template is to provide a ground truth template cross runtimes. Adding |
Author
|
@ngxson makes sense, closing this one then! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
TBH I'm not sure if this is the correct fix, but it seems to work fine locally if I also patch the jinja2 template (https://huggingface.co/ggml-org/gpt-oss-120b-GGUF) with this:
This might be a partial fix for #15110