apunkt/plano
1
0
Fork 0
mirror of https://github.com/katanemo/plano.git synced 2026-05-06 14:22:51 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions 3

add history of messages

Browse source
This commit is contained in:
cotran 2024-10-10 15:37:08 -07:00
parent 1dba11bfd9
commit 60cadf3e8a
3 changed files with 4 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

2
arch/src/consts.rs
View file

@ -1,7 +1,7 @@
pub const DEFAULT_EMBEDDING_MODEL: &str = "katanemo/bge-large-en-v1.5";
pub const DEFAULT_INTENT_MODEL: &str = "katanemo/bart-large-mnli";
pub const DEFAULT_PROMPT_TARGET_THRESHOLD: f64 = 0.8;
pub const DEFAULT_HALLUCINATED_THRESHOLD: f64 = 0.1;
pub const DEFAULT_HALLUCINATED_THRESHOLD: f64 = 0.25;
pub const RATELIMIT_SELECTOR_HEADER_KEY: &str = "x-arch-ratelimit-selector";
pub const SYSTEM_ROLE: &str = "system";
pub const USER_ROLE: &str = "user";

3
arch/src/stream_context.rs
View file

@ -756,7 +756,8 @@ impl StreamContext {
}
}
}
} else {
}
else {
user_messages = callout_context.user_message.as_ref().unwrap().clone();
}
info!("user messages: {}", user_messages);

2
model_server/app/main.py
View file

@ -179,7 +179,7 @@ async def hallucination(req: HallucinationRequest, res: Response):
"""
Take input as text and return the prediction of hallucination for each parameter
"""
logger.info(f"hallucination request: {req}")
if req.model != zero_shot_model["model_name"]:
raise HTTPException(status_code=400, detail="unknown model: " + req.model)