In this example, we will upload food review embeddings to Atlas to visualize the embeddings.
What is Atlas?
Atlas is a machine learning tool used to visualize massive datasets of embeddings in your web browser. Upload millions of embeddings to Atlas and interact with them in your web browser or jupyter notebook.
1. Login to Atlas.
!pip install nomic
import pandas as pd
import numpy as np
from ast import literal_eval
# Load the embeddings
datafile_path = "data/fine_food_reviews_with_embeddings_1k.csv"
df = pd.read_csv(datafile_path)
# Convert to a list of lists of floats
embeddings = np.array(df.embedding.apply(literal_eval).to_list())
df = df.drop('embedding', axis=1)
df = df.rename(columns={'Unnamed: 0': 'id'})
import nomic
from nomic import atlas
nomic.login('7xDPkYXSYDc1_ErdTPIcoAR9RNd8YDlkS3nVNXcVoIMZ6') #demo account
data = df.to_dict('records')
project = atlas.map_embeddings(embeddings=embeddings, data=data,
id_field='id',
colorable_fields=['Score'])
map = project.maps[0]
2. Interact with your embeddings in Jupyter
map
Project: meek-laborer
Projection ID: 463f4614-7689-47e4-b55b-1da0cc679559
Hide embedded project