Md Zuhair

icon company Upcoming SWE @ Google icon location Hello! I'm Md Zuhair from CS, IIT(BHU). I interned at Google this summer. I love developing in Python and also love developing things integrating LLMs!

Results 34 comments of Md Zuhair

Update prompt.jinja2 with an example for more consistent format reply (https://github.com/stitionai/devika/issues/429)

What exactly did you change?

Update prompt.jinja2 with an example for more consistent format reply (https://github.com/stitionai/devika/issues/429)

> The code prompt also has problems with LLAMA and some other models. Any way to improve that? What issue exactly did you face?

Bedrock Support

I'm guessing there might be some API key in bedrock too for claude? That doesn't work right?

the front is here i can choose the model the search engine select a project but the agent doesnt respond to my prompt

It'd help if you show the logs rather than the API keys. I'd also suggest to remove the API keys as you have shown the OpenAI API key and this...

the front is here i can choose the model the search engine select a project but the agent doesnt respond to my prompt

Why don't I see your message in the logs? The "Hi.... " one

Not working with LLM

What I feel is it's llama2's inability to follow the prompt and give the correct format response. Try with another LLM, like Claude3 (for which you get free 5$ initially)...

No response from devika gave apis like gemini and gpt but still no response

Does this happen even after re-running devika.py?

What's most optimal context size for the average workflow to work well ?

Umm what i feel is it at first didn't work because of the wrong response from the LLM on the prompts which are there. Because it shows "Invalid response from...

Blank UI: Unhandled Promise Rejection: ReferenceError: Cannot access uninitialized variable.

Fix white ui issues: https://youtu.be/3ULryo20mdc?si=5UoES_wVMxQTdrSq

Addition of web search feature for better error resolution and new feature implementation

Here is a video of the same (here although error was hardcoded and browser opening was hardcoded just to show) [demo.webm](https://github.com/stitionai/devika/assets/62602902/3617cfb6-0991-4dc9-91b9-9069c10b3810)