Welcome to forums.fast.ai


Is there a way to create a new post? I can’t see anywhere on how to do this. Or, is the only way to ask a new question via responding to existing posts?


(zoooo) #84

Can anyone help me with this?

When I try to import “fastai.text” library, it returns this error message instead:

SyntaxError: invalid syntax



(siddharth) #86

What is the format to create a data bunch for regression problems using images like bounding box problem?


(Thomas Boogaerts) #87

Thanks :slight_smile:


(Sashank Gondala) #88

You can also use google colab - it provides GPU completely free of cost.


(Shridhan) #89

google colab is so inconsistent - the GPU goes off suddenly…wondering if anyone tried gpuoncloud.com GPU’s for practice? Seems to be reasonably priced and stable for instant start…


(Rose) #90

I’m new to this forum so hopefully this is the right way to post this:

I’m trying to build an ML model and would like to hire someone to help me finish my project. If anyone is interested please DM me!


(Michael) #91


new here.
Just started the lessons. HAve been waiting for a while before posting for the first time.
Hope i’ll find some answers

Thanks !


(Xandre Dmitri Clementsmith) #92

Hey guys, I’m looking to post a question. Attempting to reply in order to receive privileges. Thanks!!


(JamesT) #93

One thing you could try is Google Cloud Compute and AWS, they are quite reliable and reasonably priced as well too.


(Arjun Singh) #94

Hello I am 2 year b.tech student and know only basic python,i recently watched 1 video on deep learning but could not understand the whole stuff. please help me out ,what is the right approach to get started in fast ai. Thanks


(Kiran Mundy) #95

As Jeremy explains at the beginning of the first lesson, don’t worry about not understanding - you’re not really expected to understand the details. He’s teaching top down. So your first objective, after listening to the lesson 1, should be to -

  1. Choose & setup your environment.
  2. Run the exercise
  3. Try and execute the exercise algorithm on your own image dataset.
1 Like

(John Chadwell) #96

Hey All!

New here. I’ve got all of about a week of coding under my belt. Im still stumbling around getting set up efficiently (Literally having like 97 copies of everything all over my computer). I’ve watched lessons 1 and 2 already and feel pretty confident executing the code (Its taking me a bit to figure everything out). Im currently working through putting together the pet category CNN. It’s slow, and I’m probably on the duller side of things.

One of the things thats absolute wrecking my brain is getting grips on handling using python from the terminal(mac), using jupyter notebook, and using anaconda, and using all the different environments that all depend on who is teaching.

I will get this dang CNN to label these cats and dogs. If its the last thing I do.


(sathvik udupa) #97

Have you tried google colabotary? I could run Jeremy’s pet classifier code on it in 8 mins.


(Sergii) #98

Hello everyone! Love this quiz with replying here to be able to create a new post :slight_smile:


(Rohan) #99

Simple. When you come across something you don’t know just pause the video and learn everything that you can find about it. Then continue and repeat. Soon you will find out you know more technical terms and can easily correlate stuff.

1 Like

(Rohan) #100

Usually we build deep learning models on platforms like Google Colab, Kaggle. Since you are a beginner, setting up on Colab will also be a little difficult (getting files from Google drive, importing code). I use the Clouderizer platform which has a neat and abstracted interface. You can use it even with your Mac in the backend(or any other Linux machine).

1 Like

(JamesT) #101

If it helps, I have one suggestion to help you to sort of ‘ignore’ the environments issue you are facing. You could set up virtual instances on Google Cloud Compute, which you can use for free since they give you $300 worth of credits. And as for why environments isn’t going to be an issue, well you can just delete the instance while you are done with it, not to mention that things run a lot faster on the cloud!

Instructions to set it up can be found here https://course-v3.fast.ai/start_gcp.html



Hello! I want to be able to post a question. So I’m replying to achieve that privilege. Thank you!


(Ben Hoover) #103

Starting to watch these videos and I am really enjoying the FastAI library. Looking forward to being part of this community!