Getting invalid_token error with NVIDIAEmbeddings Langchain APIs

[ashish-kumar-hpe]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

from langchain_nvidia_ai_endpoints import NVIDIAEmbeddings
import os

os.environ['REQUESTS_CA_BUNDLE']='./security/chain.pem'

base_url= "https://embedding-deployment-1-predictor-ashish-bc181805/v1"

token = 'xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx'

headers = {'Authorization': 'Bearer {token}'}
# Initialize and connect to a NeMo Retriever Text Embedding NIM (nvidia/nv-embedqa-e5-v5) running at localhost:8000
embedding_model = NVIDIAEmbeddings(model="nvidia/nv-embedqa-e5-v5", base_url=base_url, headers=headers, verify="./security/chain.pem")

# Create vector embeddings of the query
embedding_model.embed_query("How much memory does the NVIDIA H200 have?")[:10]

Description

Traceback (most recent call last):
File "/home/ashish/conversational-ai-4.0/test_embedding_model.py", line 15, in
embedding_model.embed_query("How much memory does the NVIDIA H200 have?")[:10]
File "/home/ashish/venv310/lib/python3.10/site-packages/langchain_nvidia_ai_endpoints/embeddings.py", line 157, in embed_query
return self._embed([text], model_type=self.model_type or "query")[0]
File "/home/ashish/venv310/lib/python3.10/site-packages/langchain_nvidia_ai_endpoints/embeddings.py", line 143, in _embed
response = self._client.client.get_req(
File "/home/ashish/venv310/lib/python3.10/site-packages/langchain_nvidia_ai_endpoints/_common.py", line 354, in get_req
response, session = self._post(invoke_url, payload)
File "/home/ashish/venv310/lib/python3.10/site-packages/langchain_nvidia_ai_endpoints/_common.py", line 214, in _post
self._try_raise(response)
File "/home/ashish/venv310/lib/python3.10/site-packages/langchain_nvidia_ai_endpoints/_common.py", line 311, in _try_raise
raise Exception(f"{header}\n{body}") from None
Exception: [401] Unauthorized
Bearer realm="https://10.224.200.139:8012/v1/embeddings"
error="invalid_token"
Please check or regenerate your API key.

System Info

pip freeze|grep -i langchain
langchain==0.1.11
langchain-community==0.0.25
langchain-core==0.1.52
langchain-nvidia-ai-endpoints==0.1.4
langchain-text-splitters==0.0.2

[feijoes]

@ashish-kumar-hpe
Try

os.environ["NVIDIA_API_KEY"] = nvapi_key

docs

[ashish-kumar-hpe]

Model is hosted locally with local endpoint. and api endpoint is protected through api token. now i am wondering how to pass this api token to langchain.

[feijoes]

@ashish-kumar-hpe you could create a custom NVIDIAEmbeddings and implement the logic yourself to autenticate

[ashish-kumar-hpe]

can u share an example to implement a light weight custom NVIDIAEmbeddings