How to Train BERT Over Several Epochs to Improve Accuracy
BERT PART 4
In this episode, we continue our movie review sentiment analysis investigation into BERT as we expand our first run, single-epoch model into a model trained over five epochs to improve our accuracy, then test it out with our own ad-hoc tests to see how it fares. This time around, our resulting graphs actually have meaning, since there is data to plot. Can we train a model with TinyBERT on our computer to accurately classify movie reviews? Let's find out.
Use promo code SMDE50SIX for 50% off a six month subscription for a virtual machine at:
https://beaverdamtechnologies.ca
Promo Expires Aug 31, 2026
Check out my easy time entry system for your business:
https://system.billzone.com/billzone
Join me on Patreon!
/ mackenziedataengineering
Want the code from this video?
http://mackenziemackenzie.com/downloads
Avoid hidden costs and complexity. Get Virtual machines, S3 storage, and more, on your own cloud network.
https://beaverdamtechnologies.ca
Want my team to do a project for you? Let's get to it!
https://buy.billzone.com/smdeprojects...
Contact form you can find at http://www.mackenziemackenzie.com/
Follow me on social media:
/ mackenziedataanalytics
/ seamacke
/ seamacke
/ seamacke
/ psmackenzie
Thank you to my viewers!
Mention Promo Code SMDE25TWELVE for 25% off a twelve-month virtual machine subscription at:
https://beaverdamtechnologies.ca
Got a YouTube Channel? I use TubeBuddy, it is awesome. Give it a try:
https://www.tubebuddy.com/seanmackenz...
#dataengineering #datascience #nlp