fix(llm): improve reasoning support and provider stability#420
Merged
fix(llm): improve reasoning support and provider stability#420
Conversation
This commit addresses several issues across different LLM providers related to reasoning handling and API compatibility. For `llamacpp`, reasoning is now explicitly controlled via the `reasoning_format` and `chat_template_kwargs` fields, ensuring the server handles template formatting instead of manually wrapping content in tags. It also adds support for standard parameters like temperature and max tokens, and ensures reasoning is flushed before tool calls. For `ollama`, the `think` parameter is now explicitly set to false when reasoning is disabled, preventing models from defaulting to active reasoning. For `openrouter`, the reasoning configuration is now always sent to the API. When disabled, it uses a combination of minimal effort and exclusion to avoid errors with models that reject disabled reasoning. Additionally, the `ToolCall` type was converted to a struct to improve deserialization compatibility across different underlying providers. Signed-off-by: Jean Mertz <git@jeanmertz.com>
Signed-off-by: Jean Mertz <git@jeanmertz.com>
Signed-off-by: Jean Mertz <git@jeanmertz.com>
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This commit addresses several issues across different LLM providers related to reasoning handling and API compatibility.
For
llamacpp, reasoning is now explicitly controlled via thereasoning_formatandchat_template_kwargsfields, ensuring the server handles template formatting instead of manually wrapping content in tags. It also adds support for standard parameters like temperature and max tokens, and ensures reasoning is flushed before tool calls.For
ollama, thethinkparameter is now explicitly set to false when reasoning is disabled, preventing models from defaulting to active reasoning.For
openrouter, the reasoning configuration is now always sent to the API. When disabled, it uses a combination of minimal effort and exclusion to avoid errors with models that reject disabled reasoning. Additionally, theToolCalltype was converted to a struct to improve deserialization compatibility across different underlying providers.