Fetching metadata from the HF Docker repository...
Jan 26 08:54 Update app.py added: import uuid 08:51 Update app.py Added
# We add 'collection_name' with a random ID (uuid)
# This forces Chroma to create a brand new, empty bucket for every PDF
vectordb = Chroma.from_documents(
documents=chunks,
embedding=embedding_model,
collection_name=f"pdf_chat_{uuid.uuid4().hex}"
) 07:35 Update app.py Add:
from huggingface_hub import hf_hub_download
from langchain_community.llms import LlamaCpp 07:32 Update app.py Add the following line : LLM_INSTANCE = None 07:29 Update app.py correct the word mglobal to global 07:14 Update requirements.txt Add llama-cpp-python 07:11 Update app.py CPU-compatible version of Mistral 7B 06:52 Update app.py Changed the server address from 127.0.0.1 to 0.0.0.0 06:42 Update app.py Correct pipeline mistake in the transformer import 06:25 Update app.py corrected 'pipeline' to 'pipe' in get_llm() 06:17 Update app.py Correct error message from "Pipeline" to "pipeline" in get_llm() 06:12 Upload RAG chatbot files Firest version of granite 3.0 8b 05:38 initial commit 
