Comments (6)
jackytu256
commented on August 25, 2024
1
Thanks a lot !! Fried
Get Outlook for Android<https://aka.ms/ghei36>
…________________________________
From: FriedRonaldo <[email protected]>
Sent: Tuesday, October 6, 2020 11:25:42 PM
To: clovaai/tunit <[email protected]>
Cc: jackytu256 <[email protected]>; Author <[email protected]>
Subject: Re: [clovaai/tunit] OOM (#15)
Actually, it is totally from the memory capacity. I already tried running the code with a single 2080ti that is the same model you have.
More specifically, that behavior is intended. Untill 65 epoch, the code trains the guiding network only for the stability.
After the point, GANs training starts and the guiding network and GANs are trained with the overall objectives.
So, only small memory is required before the entire training so that you can run the code while 65 epochs but cannot after 65 epoch.
Please follow issue #9<#9>.
Thanks.
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub<#15 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ACA5TFB7DHNBEHFU6YCYMNTSJMZHNANCNFSM4SF4HBSA>.
from tunit.
FriedRonaldo
commented on August 25, 2024
Well... In case of OOM, you can try the experiment with GPU having larger memory or reduce the size of the networks.
Please refer to issue #9
from tunit.
jackytu256
commented on August 25, 2024
The gpu I have is 2080ti which contains 11g memory so I reckon memory capacity is not a main issue for this??
Get Outlook for Android<https://aka.ms/ghei36>
…________________________________
From: FriedRonaldo <[email protected]>
Sent: Tuesday, October 6, 2020 7:21:49 PM
To: clovaai/tunit <[email protected]>
Cc: jackytu256 <[email protected]>; Author <[email protected]>
Subject: Re: [clovaai/tunit] OOM (#15)
Well... In case of OOM, you can try the experiment with GPU having larger memory or reduce the size of the networks.
Please refer to issue #9<#9>
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub<#15 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ACA5TFBLBDADMFWJV5TCMRTSJL4U3ANCNFSM4SF4HBSA>.
from tunit.
FriedRonaldo
commented on August 25, 2024
Unfortunately, memory capacity is the main issue. I also have 2080ti but I can not run the code on a single 2080ti.
This project is mainly done with V100 and P40 (2xP40).
With batch_size=32 and img_size=128, you might need more than 24GB.
from tunit.
jackytu256
commented on August 25, 2024
Thanks for your suggestions @FriedRonaldo
The model is initialized with roughly usage of 6g memory and can be trained till 65 epochs to occur OOM. Do I need to provide more info for you to figure out what the issue is ? or I just follow the steps with issue #9.
Thanks for the quick reply!
Get Outlook for Android<https://aka.ms/ghei36>
…________________________________
From: FriedRonaldo <[email protected]>
Sent: Tuesday, October 6, 2020 8:08:16 PM
To: clovaai/tunit <[email protected]>
Cc: jackytu256 <[email protected]>; Author <[email protected]>
Subject: Re: [clovaai/tunit] OOM (#15)
Unfortunately, memory capacity is the main issue. I also have 2080ti but I can not run the code on a single 2080ti.
This project is mainly done with V100 and P40 (2xP40).
With batch_size=32 and img_size=128, you might need more than 24GB.
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub<#15 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ACA5TFDJSXJDCPRWD6WIUCDSJMCDBANCNFSM4SF4HBSA>.
from tunit.
FriedRonaldo
commented on August 25, 2024
Actually, it is totally from the memory capacity. I already tried running the code with a single 2080ti that is the same model you have.
More specifically, that behavior is intended. Untill 65 epoch, the code trains the guiding network only for the stability.
After the point, GANs training starts and the guiding network and GANs are trained with the overall objectives.
So, only small memory is required before the entire training so that you can run the code while 65 epochs but cannot after 65 epoch.
Please follow issue #9.
Thanks.
from tunit.
Related Issues (20)
- Why the argumentation transform of the orginal image contains colorjitter HOT 6
- Seek the AnimalFaces Dataset HOT 2
- question about the animal faces dataset HOT 1
- FFHQ: IndexError list index out of range HOT 4
- Training on non-square images HOT 1
- It looks like the requirements.txt file needs to be modified. HOT 1
- About custom dataset HOT 1
- afhq dataset error HOT 11
- Validation Error？ HOT 8
- About Food-10 dataset settings HOT 5
- Average style vectors? HOT 2
- Why updating the gradients of the guiding network C while training G? HOT 2
- Why the style encoder maps all the images to the same style code? HOT 7
- Can TUNIT do few-shot generation? HOT 4
- Training Discriminator vs Generator in TUNIT HOT 2
- pretarined food10 weights HOT 5
- I trained to 65 epochs and found an error：“fid_ema, fid_ema_mean is not defined” HOT 9
- FFHQ results HOT 1
- Own Dataset HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from tunit.