Embedding Models

​

Prerequisites

1. Create an API key for authentication.
2. Save the API key to an environment variable:
   export NEBIUS_API_KEY=<API_key>
   
3. Install LlamaIndex packages:
   pip3 install llama-index-embeddings-nebius llama-index
   

   Show If you get error: externally_managed_ environment

   Create a virtual Python environment. You can install packages there that are isolated from the basic environment.To prepare a virtual environment:

   1. Create it:
      python3 -m venv <environment_name>
      
   2. Activate the environment:
      source <environment_name>/bin/activate
      
      A directory with the environment name is created.

   Now, you can install required Python packages. When you no longer need the created virtual environment, run the deactivate command and delete the environment directory.

​

Prepare a script

1. Copy the following part of the script:
   from llama_index.embeddings.nebius import NebiusEmbedding
   import os
   
   # Take the API key from the environment variable
   NEBIUS_API_KEY = os.getenv("NEBIUS_API_KEY")
   
   # Load an AI model
   embed_model = NebiusEmbedding(api_key=NEBIUS_API_KEY)
   
2. Add one of the following methods, depending on your use case:

   Use case	Description	How to implement
   Regular request	Get a vector for a text.	Call the embed_model.get_text_embedding(<text>) method and print its output as print(embeddings[:5], sep="\n").
   Asynchronous request	Get a vector asynchronously, so methods that follow your request do not wait for the response.	Call the await embed_model.aget_text_embedding(<text>) method.
   Batch of texts	Get vectors for several texts.	Call the embed_model.get_text_embedding_batch(texts) method where texts is an array of strings. Next, print vectors as print(*[x[:3] for x in embeddings], sep="\n").
   Asynchronous request with a batch of texts	Get vectors for several texts asynchronously.	Call the await embed_model.aget_text_embedding_batch(texts) method where texts is an array of strings.

​

Examples

​

Asynchronous request

from llama_index.embeddings.nebius import NebiusEmbedding
import os
import asyncio

# Take the API key from the environment variable
NEBIUS_API_KEY = os.getenv("NEBIUS_API_KEY")

# Load an AI model
embed_model = NebiusEmbedding(api_key=NEBIUS_API_KEY)

# Request a vector for the specified text
text = "Everyone loves justice at another person's expense"
async def complete():
    embeddings = await embed_model.aget_text_embedding(text)
    assert len(embeddings) == 4096
    print(len(embeddings), embeddings[:5], sep="\n")
asyncio.run(complete())

4096
[-0.0024051666259765625, 0.0083770751953125, -0.005413055419921875, 0.007396697998046875, -0.022247314453125]

​

Request with a batch of texts

from llama_index.embeddings.nebius import NebiusEmbedding

import os

# Take the API key from the environment variable
NEBIUS_API_KEY = os.getenv("NEBIUS_API_KEY")

# Load an AI model
embed_model = NebiusEmbedding(api_key=NEBIUS_API_KEY)

# Create a batch of texts
texts = [
    "As the hours pass",
    "I will let you know",
    "That I need to ask",
    "Before I'm alone",
]

# Request a vector for the batch
embeddings = embed_model.get_text_embedding_batch(texts)
assert len(embeddings) == 4
assert len(embeddings[0]) == 4096
print(*[x[:3] for x in embeddings], sep="\n")

[-0.0003848075866699219, 0.0004799365997314453, 0.011199951171875]
[-0.0037078857421875, 0.0114288330078125, 0.00878143310546875]
[0.005924224853515625, 0.005153656005859375, 0.001438140869140625]
[-0.009490966796875, -0.004852294921875, 0.004779815673828125]