Training custom models on Vertex AI

Показать описание

A managed ML training service can help you automate experimentation at scale or retain models for a production application. In this episode of Prototype to Production, Developer Advocate, Nikita Namjoshi, walks through the steps required to train custom models on Vertex AI. Watch along and learn about the benefits of a managed training service that helps keep your results fresh.

Chapters:
0:00 - Intro
0:22 - Why do I need a machine learning training service?
1:26 - What are containers?
2:19 - Update custom training code
3:23 - Cloud storage for machine learning
4:50 - Containerizing code for machine learning
5:39 - Dockerfile syntax
6:42 - How to store container images in Google Cloud
7:21 - How to launch a training job on Vertex AI
8:12 - Wrap up

#Prototype2Production

Рекомендации по теме

Комментарии

Wow this was so much better than the videos in the google ml path courses thanks for this

SomethingRandomChannel

nice and helpful video. I started learning how to train custom model code with vertex AI and found this video helpful to me.

MMReza-xf

This is what I am looking for. Thanks for the video 👍🏼

getjawa

Thanks for this video, It's helpful!

prajwalsyallur

can i not directly select run time gpu on colab enterprise

AbdulMusawarSoomro

Is vertex custom training done on our local computer? In other words, does vertex custom training allow us to train our model using our local computing, and ram on our personal computer instead of Google cloud?

LoneLeagle

3:48 its slash or forward slash. Not backslash!

SukantaPaul

i got this error by model deploying: FileNotFoundError: [Errno 2] No such file or directory:

zisang

Como es posible que esta info no esté en español, acabo de ver que es funcional hace mucho y apenas me enteró todo lo que hubiera podido hacer si lo hubiera conocido antes... Algunas veces si es demasiado tarde... Igual me alegra haber encontrado esto

HectorFonsecaCano

I might just have developed a tiny bit of a crush on Nikita Namjoshi 👀. She's so cool

deepansharya

I can not use the /gcs/MY_BUCKET/ path. My service account has the "Storage Object Admin" Role and I could mount it manually to a given folder. But I cant use it by default as shown here. The bucket is in the same location as the notebook and in the same project. Did you skip a necessary step here?

Edit: never mind. This only works in the training pipeline but when you want to test initially in the notebook you need to mount a folder manually

bytblaster

Training custom models on Vertex AI

Training custom models on Vertex AI

Build a custom ML model with Vertex AI

Fuel your custom models on Vertex AI

How To Build and Train Custom Models

Building and training ML models with Vertex AI

Build, tune, and deploy foundation models with Vertex AI

Vertex AI Training and Serving a Custom Model

Google Vertex AI : Building Model and Training Hands-on Lab

Building a Machine Learning Model with Google Vertex AI: A Step-by-Step Guide

Training Custom Chatbot Models with Labelbox and GCP Vertex AI

Building Efficient PyTorch Models on Vertex AI - Joinal Ahmed

Training and Deploying a TensorFlow Model in Vertex AI #qwiklabs || #coursera [With Explanation🗣️]...

Deploy a custom model to Vertex AI

Practical AI 012a: Training an Image Classification Model with Vertex AI

Custom Model Training & Deployment on Google Cloud: A Vertex AI Guide in Go! | AI Heroes 2023 Ta...

PyTorch on Vertex AI

Run custom training job with custom container in Vertex AI

Training a model with Vertex AI Forecast

Build, tune, and deploy foundation models with Generative AI Support in Vertex AI

How to train ML models with Google Vertex AI Training

Inference in Vertex AI: manually deploying a custom model

Build Custom Models with Vertex AI's AutoML Suite

Custom Training Pipeline | Vertex AI Demo | GCP

How to tune LLMs in Generative AI Studio