Comments (10)
Thank you very much!
-
Minimum requirements for a PC
I am sorry. I have not checked it exactly.
We will confirm it at the beginning of the week. -
About "requirements.txt"
We will also check it at the beginning of the week.
As soon as we get information, we will report you.
from dcase2020_task2_baseline.
Thank you very much.
I want to know your environment.
- OS version
- Python version
For answering the second question, I changed versions to
- Keras-Preprocessing == 1.1.0
- gast == 0.3.3
and ran the baseline.
The same results were obtained.
from dcase2020_task2_baseline.
@nakamura419
Thanks for your reply.
I run the code in
Windows10 Home (version 1903),
Pycharm 2019.2.4(Community),
Python 3.6.10.
from dcase2020_task2_baseline.
Answer questions 1.
As a result of training only with the largest amount of data, Toy Car, the maximum memory usage was 7578MB.
Even if batch_size was changed, this usage did not change much, so I think it is better not to use all dev_data with 8GB memory.
from dcase2020_task2_baseline.
We have found that the memory usage can be reduced to 4136MB if the parameter of "n_mels" is changed from 128 to 64.
Please try it.
from dcase2020_task2_baseline.
@Yuki-Nikaido
Thank you. I changed the 'n_mels' from 128 to 64, and 'batch_size' from 512 to 1, it still does not work. However, I reduced half of the training data, then it works. For example, there are 3291 files in valve, and the training process works when only 1496 files are used.
from dcase2020_task2_baseline.
Because train_data = list_to_vector_array() read all data in to train_data, it is very memory-consuming. It could save memory consumption by reading data in batch_size.
from dcase2020_task2_baseline.
If only one model (ToyCar, valve ... etc) is trained, it can be executed with 8GB memory, so if you use all data, how about training only one model in one execution?
When using n_mels = 64 and using all dev_data, it was confirmed that a maximum of 8527 MB of memory was used, so in an 8 GB memory environment, training using all dev_data would be difficult.
(Maximum memory usage depends on garbage collection)
from dcase2020_task2_baseline.
@20050710212 Thank you very much!
We understand that loading data little by little will save memory consumption.
However, processing speed will be slower. So, we will not revise this repository in such a direction.
At this time, it is recommended to train one machine-type model at a time as @Yuki-Nikaido said.
We hope you understand.
from dcase2020_task2_baseline.
@y-kawagu @Yuki-Nikaido @nakamura419
Thank you for the reply. I have already add a memory chip in my pc.
from dcase2020_task2_baseline.
Related Issues (3)
- eroor HOT 20
- shape error 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 dcase2020_task2_baseline.