Introduction to LlamaIndex v0.10
ฝัง
- เผยแพร่เมื่อ 21 ก.ค. 2024
- This video gives a comprehensive overview of LlamaIndex v0.10, the largest release to our Python package to date.
We cover the following topics:
1. Creating a `llama-index-core` package and splitting integrations/templates into separate packages
2. LlamaHub Revamp
3. ServiceContext Deprecation
Throughout these sections, we'll cover a migration guide, usage examples, and a contributing guide.
If you have questions come hop in our Discord: / discord
Timeline:
00:00-02:30 Intro
02:30-09:30 Package Refactor Overview
09:30-13:17 Package Refactor Usage Example
13:17-16:00 Package Refactor Migration Guide
16:00-18:17 LlamaHub
18:17-23:28 Deprecate ServiceContext
23:28 Contributing to LlamaIndex v0.10
Additional Notes:
- If you're installing in a virtual environment and running into installation/import issues, try deleting your entire virtual environment and trying again. Another example that works is this:
```
pip uninstall llama-index
pip install llama-index --upgrade --no-cache-dir --force-reinstall
```
amazing changes :) and brilliant explanation 🎉
Kudos! Living without that annoying "ServiceContext," ;) especially mixed with global settings, shrinks the codebase by up to 10 times. Thx for the video! It was a headache figuring out these new things without this intro.
Hello, are there plans for deeper DSPy integration for optimized prompt capabilities?
Thanks, very good. 😀
What's the process for contributing to README's? Is it the same as packages and readers?
This seems like a mostly structural change, so all the demo videos should still work, just with the updated paths right?
yep exactly!
Hello, can you create a video showing the main differences between Llma Index and Langchain. When to use which and so on. Thanks, amazing work!
Choosing Between Llama Index and Lang-Chain:
When choosing between the Llama Index and Lang-Chain, consider the following factors:
Use Case: Determine the specific tasks and requirements you have for text analysis. If you need a wide range of NLP capabilities and have access to significant computational resources, the Llama Index may be more suitable. If you prioritize decentralization, data privacy, and community-driven approaches, Lang-Chain could be a better fit.
Resource Availability: Assess your access to computational resources, including hardware and expertise. The Llama Index may require substantial resources for setup and maintenance, whereas Lang-Chain may offer a more accessible infrastructure for decentralized processing.
Security and Privacy: Consider the sensitivity of your data and whether you require a decentralized approach for enhanced security and privacy. Lang-Chain's blockchain-based architecture may offer advantages in terms of data control and privacy.
Community and Ecosystem: Evaluate the support and ecosystem around each tool. The availability of documentation, community forums, and developer resources can impact ease of use and troubleshooting.
Ultimately, the choice between the Llama Index and Lang-Chain depends on your specific needs, priorities, and available resources. It may be beneficial to experiment with both solutions and assess their suitability for your use case before making a decision.
from llama_index.embeddings.huggingface import HuggingFaceEmbedding
from llama_index.core import Settings
Settings.embed_model = HuggingFaceEmbedding(
model_name="BAAI/bge-small-en-v1.5"
)
I tried to change to the new version. Can anyone help me out how to fix this? Its not working.
It throws an attribute error: module 'torch._subclasses' has no attribute 'functional_tensor'