You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Add Support for Groq in Giskard or show it in the documentation if already supported
🔈 Motivation
The integration of Groq as an alternative to OpenAI in the Giskard library is highly beneficial. Groq offers speed that is almost equivalent or even faster than GPT-4 and significantly faster than any Ollama model. This will provide users with a high-performance alternative for their machine learning tasks.
🛰 Alternatives
Currently, the only alternative is using OpenAI's models, which may not meet all users' needs for speed and performance. Can use Ollama as well but using it is a lot of hassle especially in colab or kaggle notebooks.Incorporating Groq provides a competitive edge and additional flexibility.
📎 Additional context
Here is the proposed code to integrate Groq with Giskard:
from groq import `Groq`
from giskard.llm.client.openai import OpenAIClient
import giskard as gsk
_client = Groq(api_key="YOUR_GROQ_API_KEY")
oc = OpenAIClient(model="any_model_in_groq", client=_client)
gsk.llm.set_default_client(oc)
This simple addition allows users to set up Groq as their default client in Giskard, enabling them to take advantage of Groq's superior speed and performance. Detailed documentation and examples will also be necessary to guide users through the setup and usage process.
The text was updated successfully, but these errors were encountered:
Hey @Vikas9758, thanks for your suggestion!
Indeed it would be very useful to use Groq API itself instead of implementing a custom client.
Would you be interested in implementing this?
🚀 Feature Request
Add Support for Groq in Giskard or show it in the documentation if already supported
🔈 Motivation
The integration of Groq as an alternative to OpenAI in the Giskard library is highly beneficial. Groq offers speed that is almost equivalent or even faster than GPT-4 and significantly faster than any Ollama model. This will provide users with a high-performance alternative for their machine learning tasks.
🛰 Alternatives
Currently, the only alternative is using OpenAI's models, which may not meet all users' needs for speed and performance. Can use Ollama as well but using it is a lot of hassle especially in colab or kaggle notebooks.Incorporating Groq provides a competitive edge and additional flexibility.
📎 Additional context
Here is the proposed code to integrate Groq with Giskard:
This simple addition allows users to set up Groq as their default client in Giskard, enabling them to take advantage of Groq's superior speed and performance. Detailed documentation and examples will also be necessary to guide users through the setup and usage process.
The text was updated successfully, but these errors were encountered: