Nicholas Liu

I ran this is a notebook and get the following: ``` import numpy as np import pandas as pd from langchain.agents import create_pandas_dataframe_agent from langchain.llms import OpenAI from langchain.callbacks import...

Do you have more details on reproducing? Are you using `google_drive.ipynb`? Looking through the code, if `folder_id` is provided, the loader only fetches the files after `load()` is called. This...

> It's working well if I go to the trash in Google Drive, and click "Empty Trash". +1, I just tested this and saw the same behavior. This seems like...

I found a way to check if a file is trashed, fill send out a PR tonight.

I believe this `max_tokens` parameter is passed to the openAI API (https://platform.openai.com/docs/api-reference/completions/create#completions/create-max_tokens). Changing this parameter won't trigger the model to provide longer responses. Instead, I'd suggest you to try using...

Made the stricter typing fix. Added an example to the existing notebooks. lmk if there's anything else to change.

Is this just waiting for a maintainer to merge? Also if this change makes it to twitter, would appreciate a s/o `nicholasliu77` :smile:

I am also curious what the effects of the limited vocabulary has on model performance. Eagle 3 paper mentions the 2 main contributions are training time test and feature fusion....

It seems the authors took it down when preparing to release EAGLE 3. The generating script exists on older commits such as: https://github.com/SafeAILab/EAGLE/tree/7e885b2ba9cf6d116cd041790f06143720821a68 Alternatively, I'm maintaining an unofficial implementation here...

I've implemented my understanding of training-time-test here: https://github.com/NickL77/BaldEagle?tab=readme-ov-file#eagle-3-status It performs 11.7% faster and has 8.4% higher acceptance rate than my Eagle 2 baseline. You can see benchmark results in the...