Comments (5)
Hey Eyal, thanks for bringing this up
We do have plans for this and is in our roadmap but it might take a while. All our testing is with OpenAI so we can't ensure accuracy with opensource models. That being said I was wondering what were the reasons you wanted open-source models and how big of a model can you run at your end. Also, would you be okay running a ~30B model for evaluations, do you have a hosted version available right now?
This will help us get more sense of how to prioritise this
from ragas.
Hello,
I am in a similar situation, as I am building a RAG pipeline but I cannot use the OpenAI key (due to cost reasons); I am using a quantized version of Llama2 that fits in the free version of Colab.
The idea I had was to use the same model for generation and evaluation, aware that the results won't be that accurate as in case I was using GPT3.5/GPT4, but still in some ways informative for optimization.
Looking here (https://github.com/explodinggradients/ragas/blob/main/docs/guides/llms.ipynb) I thought it was possible to do that by changing the llm called by the metrics, but looking at your comment I don't know if it is convenient (?)
Do you have any suggestions?
from ragas.
Hey @Robs1999, full support for opensource LLMs will be coming shortly, I'll update this once its in
from ragas.
Thanks a lot for the update @jjmachan! I was wondering if you could give us a rough idea of when we might expect full support for open-source language models? Even a rough estimate would be really helpful!
Thanks again, and keep up the great work! 😊🚀
from ragas.
appologies for the delay but we have shipped this feature you can check the docs here https://docs.ragas.io/en/latest/howtos/customisations/llms.html#evaluating-with-open-source-llms
do check it out and feel free to open it this doesn't address the issue completely.
from ragas.
Related Issues (20)
- Azure has not provided the response due to a content filter being triggered HOT 2
- Ragas scores look inconsistent HOT 7
- evaluate not working HOT 2
- OPENAI key error even on open-source llm
- Rate Limit Exceeded Exceptions for OpenAI when `is_async=True` HOT 5
- evaluate forcefully stopped
- AttributeError("'list' object has no attribute 'get'") when calculating metrics score HOT 6
- time out
- Azure Open AI Error when creating synthetic test data - Chat Completion Models Not Supported HOT 2
- Python library must be updated to the latest version HOT 3
- When I use my own LLM model to evaluate answer_relevancy, the following error occurs.
- Context Precision Error using own LLMs
- I'm not able to reproduce documentation example HOT 6
- How can we use Azure OpenAI key with Ragas? HOT 1
- ImportError: cannot import name 'LangchainLLM' from 'ragas.llms' HOT 3
- TypeError: create() got an unexpected keyword argument 'deployment_name' HOT 1
- IndexError: Invalid key: 0 is out of bounds for size 0 HOT 2
- ValueError: Project root not found!
- ValueError: diag requires an array of at least two dimensions
- answer_correctness : too many values to unpack (expected 3) HOT 8
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from ragas.