Added kalshi bet predictor example #85
Conversation
| import os | ||
| from exa_py import Exa | ||
|
|
||
| def getKalshiQuestion(market_ticker)->Tuple[str,str]: |
There was a problem hiding this comment.
why make 2 requests to this endpoint? Make 1 and just store the results to be used
|
|
||
|
|
||
| class BetPredictor: | ||
| def __init__(self, model_name: str = "Qwen/Qwen3-4B-Instruct-2507"): |
There was a problem hiding this comment.
What made you go with this model out of curioisity?
There was a problem hiding this comment.
The 3B-6B model size was enough to produce reasonable responses and its really only Qwen and Llama variants in that range. The instruct models are fine tuned for instruction following, and the Qwen3 instruct 2507 model preforms better on instruction following then Llama 3.2 instruct model.
| numpy==2.3.4 | ||
| openai==2.5.0 | ||
| packaging==25.0 | ||
| playwright==1.55.0 |
There was a problem hiding this comment.
Do you really need all these? I dont see playright, jinja etc in your code
There was a problem hiding this comment.
Fixed. Jinja is a torch dependency but playwright was leftover from web scraping attempt
2 participants
No description provided.