updated Embeddings.ipynb to use latest python sdk

[ROHANDWIVEDI2005]

Can anyone tell me if the syntax for taskType of the models.embedContent method has changed in the new python sdk.
I cheked the embeddings docs for it.

does this mean taskType can only be used for embedding-001 because in the embeddings notebook is using it with embedding-004.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[markmcd]

Line #2.    result = client.models.embed_content(model="text-embedding-004", content=text)

Please run the colab and make sure everything is correct.

Reply via ReviewNB

[markmcd]

Line #5.    print(str(result.embeddings)[:50], '... TRIMMED]')

This isn't quite the same output as before... Can you use result.embeddings[0].values instead of result.rembeddings here?

Reply via ReviewNB

[ROHANDWIVEDI2005]

done

[markmcd]

Line #1.    print(len(result.embeddings)) # The embeddings have 768 dimensions

This is also incorrect.

Since this is batching/unbatching, it might be clearer to use lists as input and unpack the results when they're generated. e.g.

texts = ["Hello world"]

result = client.models.embed_content(model="text-embedding-004", contents=texts)

[embedding] = result.embeddings
print(len(embedding.values))

Reply via ReviewNB

[ROHANDWIVEDI2005]

yes, i updated this as well

[markmcd]

This is broken too. Please run this & ensure it executes and that the outputs are correct.

Reply via ReviewNB

[ROHANDWIVEDI2005]

Ok I have ran the entire notebook on colab all the outputs are correct.

[markmcd]

See reviewnb comments

[markmcd]

Thanks for the contibution!

does this mean taskType can only be used for embedding-001 because in the embeddings notebook is using it with embedding-004.

No that doc comment looks wrong. It should work with the text004 model too (as seen in this notebook)

edit: I'll get this updated. It'll take a little time to approve and propagate but thanks for pointing it out.