xpltdco/chrysopedia
2
0
Fork 0
Code Issues 2 Pull requests Projects Releases Packages Wiki Activity Actions

fix: Add max_tokens=16384 to LLM requests (OpenWebUI defaults to 1000, truncating pipeline JSON)

Browse source
This commit is contained in:
jlightner 2026-03-30 04:08:29 +00:00
parent 8e96fae64f
commit cf759f3739
3 changed files with 8 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

3
.env.example
View file

@ -29,6 +29,9 @@ LLM_FALLBACK_MODEL=fyn-qwen35-chat
#LLM_STAGE5_MODEL=fyn-qwen35-thinking #LLM_STAGE5_MODEL=fyn-qwen35-thinking
#LLM_STAGE5_MODALITY=thinking #LLM_STAGE5_MODALITY=thinking
# Max tokens for LLM responses (OpenWebUI defaults to 1000 — pipeline needs much more)
LLM_MAX_TOKENS=16384
# Embedding endpoint (Ollama container in the compose stack) # Embedding endpoint (Ollama container in the compose stack)
EMBEDDING_API_URL=http://chrysopedia-ollama:11434/v1 EMBEDDING_API_URL=http://chrysopedia-ollama:11434/v1
EMBEDDING_MODEL=nomic-embed-text EMBEDDING_MODEL=nomic-embed-text

3
backend/config.py
View file

@ -43,6 +43,9 @@ class Settings(BaseSettings):
llm_stage5_model: str | None = None # synthesis — thinking model recommended llm_stage5_model: str | None = None # synthesis — thinking model recommended
llm_stage5_modality: str = "chat" llm_stage5_modality: str = "chat"
# Max tokens for LLM responses (OpenWebUI defaults to 1000 which truncates pipeline JSON)
llm_max_tokens: int = 16384
# Embedding endpoint # Embedding endpoint
embedding_api_url: str = "http://localhost:11434/v1" embedding_api_url: str = "http://localhost:11434/v1"
embedding_model: str = "nomic-embed-text" embedding_model: str = "nomic-embed-text"

2
backend/pipeline/llm_client.py
View file

@ -136,6 +136,7 @@ class LLMClient:
response = self._primary.chat.completions.create( response = self._primary.chat.completions.create(
model=primary_model, model=primary_model,
messages=messages, messages=messages,
max_tokens=self.settings.llm_max_tokens,
**kwargs, **kwargs,
) )
raw = response.choices[0].message.content or "" raw = response.choices[0].message.content or ""
@ -156,6 +157,7 @@ class LLMClient:
response = self._fallback.chat.completions.create( response = self._fallback.chat.completions.create(
model=fallback_model, model=fallback_model,
messages=messages, messages=messages,
max_tokens=self.settings.llm_max_tokens,
**kwargs, **kwargs,
) )
raw = response.choices[0].message.content or "" raw = response.choices[0].message.content or ""