Is my embedding code up to date?

API
1

Hello,

Is this correct embedding code? Please look at the combined code as well.


import pandas as pd

input_datapath = 'E:/ProgramData/openai/qa/csv/qa.csv'
df = pd.read_csv(input_datapath, header=0)
df = df[['Question', 'Answer']]
df = df.dropna()
df['combined'] = "Question: " + df.Question.str.strip() + "; Content: " + df.Answer.str.strip()
df.head(2)

from transformers import GPT2TokenizerFast
tokenizer = GPT2TokenizerFast.from_pretrained("gpt2")

from openai.embeddings_utils import get_embedding

df['ada_similarity'] = df.combined.apply(lambda x: get_embedding(x, engine='text-similarity-ada-001'))
df['ada_search'] = df.combined.apply(lambda x: get_embedding(x, engine='text-search-ada-doc-001'))
df.to_csv('E:/ProgramData/openai/qa/embeddings/qa_embeddings.csv')

Regards,
William Johnston

2

Hello,

I found the latest embedding code on the OpenAI website.

I also had GPT-3 generate Python code that creates lists for the maximum number of tokens for Ada embeddings.

Thanks.

William Johnston