search

c100rczyk / Team_project_PWR

This is team project implemented as part of the IT specialization in the field of Automation and Robotics, in cooperate with company. Main technologies are: Computer Vision. Language: Python.
MIT License
2 stars 0 forks source link

Connecting with Kaggle #28

Closed c100rczyk closed 4 months ago

c100rczyk commented 4 months ago

How to succesfully use Kaggle Tesla GPU, to train models. Transfer code to Kaggle.

c100rczyk commented 4 months ago

How i connected with Kaggle:

c100rczyk commented 4 months ago

  1. Open Kaggle Notebook

  2. Name opened notebook

  3. File -> Link to GitHub

  4. Type in new code cell: !git clone https://repozytorium

  5. Kaggle settings / create new API

  6. Put loaded file.json to : Notebook / Input / Upload / (our file.json)

  7. Run this code cell in Notebook:

c100rczyk commented 4 months ago

(it creating new directory, in Kaggle)

import os import shutil os.makedirs('/root/.kaggle', exist_ok = True)

shutil.copy('/kaggle/input/kagglejson/kaggle.json', '/root/.kaggle/kaggle.json')

os.chmod('/root/.kaggle/kaggle.json', 0o600)

c100rczyk commented 4 months ago

add this befor importing everything:

import sys sys.path.append('/kaggle/working/Team_project_PWR')

c100rczyk commented 4 months ago

Next download prepared dataset. Create dataset.zip and put this to Inputs/Upload/Dataset on Kaggle in our notebook

c100rczyk commented 4 months ago

Then change paths to every file add kaggle/input/prepareddata/...

To install what you need : install like in your terminal yust put in kaggle notebook code cell for example: !pip install livelossplot