Find answers from the community

Updated 6 months ago

from llama_index.core.response import

At a glance

The community member is having trouble importing StreamingResponse from llama_index.core.response. Other community members suggest trying a fresh virtual environment, and note that the import has been moved to llama_index.core.base.response.schema. The community members also discuss using LangChainLLM to wrap the Text Generation Interface (TGI) and the OpenAILike class for the OpenAI API version of TGI.

Useful resources

ddrewskidang

from llama_index.core.response import StreamingResponse
i cant impore this for somereason

16 comments

LLogan M

Can you try with a fresh venv?

In a new terminal

Plain Text

pip uninstall llama-index  # remove any potential global install
python -m venv venv
source venv/bin/activate
pip install llama-index

ddrewskidang

No bueno 😦

ddrewskidang

Attachment

LLogan M

Typo

LLogan M

StreamingResponse

LLogan M

actually nvm its not there

ddrewskidang

lmaoo

LLogan M

Looks like it got moved, and the backwards compatibility import got missed during the v0.10.x explosion

LLogan M

from llama_index.core.base.response.schema import StreamingResponse

LLogan M

that will work

ddrewskidang

it worked!

ddrewskidang

im currently using langchain llms does llama index have a huggingface text generation wrapper
I see this but i have my own dedicated serve with a url
https://docs.llamaindex.ai/en/stable/examples/llm/huggingface.html#hugging-face-llms

LLogan M

Are you using the openai api version of TGI?

LLogan M

If so, you can use the OpenAILike class

LLogan M

Otherwise you can wrap TGI with the LangChainLLM class

ddrewskidang

gotcha gotcha

Add a reply