How to: Get the fastest GPU on google colab for free ?

Поделиться
HTML-код
  • Опубликовано: 13 янв 2025

Комментарии • 60

  • @yahyaefhaima1238
    @yahyaefhaima1238 3 года назад +8

    what do i do if it says "'You cannot currently connect to a GPU due to usage limits in Colab''

  • @samay1163
    @samay1163 3 года назад +1

    hello...runtime is for per day or month?i mean if we are taking google colab pro than we are getting 24 hours runtime for a day or its for overall month?

  • @RADEONR
    @RADEONR Год назад

    i haven't factory reset what do you prefer command to change T4 GPU?

  • @babayaga.mp4
    @babayaga.mp4 4 года назад +2

    More power to you.

  • @tflame7925
    @tflame7925 2 года назад

    On certain google accounts, my programs will run different iteration speeds. For example, when I’m logged in here, it’ll run at 8.1 iterations/ second, but my other google acc will run at 5.7 iterations/ second, what could be the reason for this and how can I make this faster but why does it do this?

  • @AnkitGupta-sc1ys
    @AnkitGupta-sc1ys 4 года назад +2

    You are future of Nepal surely you know so much knowledge about ML. Can I know what u are recently studying. Please bro. I m from nepal too at butwal

    • @Cominlooks
      @Cominlooks  4 года назад +2

      Thanks. Appreciate it. I am studying electronics at pulchowk engineering campus

    • @AnkitGupta-sc1ys
      @AnkitGupta-sc1ys 4 года назад

      @@Cominlooks Oh.... Thank you for telling me.

    • @titanide
      @titanide 3 года назад

      @@Cominlooks from IPython.display import HTML
      from subprocess import getoutput
      s = getoutput('nvidia-smi')
      if 'K80' in s:
      gpu = 'K80'
      elif 'T4' in s:
      gpu = 'T4'
      elif 'P100' in s:
      gpu = 'P100'
      display(HTML(f"{gpu}"))
      print(s)
      # or simply
      !nvidia-smi -L
      FOR CONSTANTLY SEARCHING P100

  • @EthioGoldenAge-uu8ev
    @EthioGoldenAge-uu8ev Год назад

    My memory crashed while using the normal GPU , Is there anyone to tell me how can I use memory without limitation of processing, No crashing of available memory

  • @AnirbanBiswas123
    @AnirbanBiswas123 3 года назад +4

    P100 is not the fastest, V100 is the fastest
    then comes T4

    • @Cominlooks
      @Cominlooks  3 года назад +2

      Sure v100 is faster, but only available for colab pro.

  • @elliykollek
    @elliykollek 2 года назад

    Thanks shukran, but when I try too much times it stops working. Jazakum Ullahu khayra

  • @oguzhanbolukbas9621
    @oguzhanbolukbas9621 3 года назад +1

    NameError: name 'GPU_model' is not defined

  • @murtazahussain8308
    @murtazahussain8308 3 года назад +7

    Please do not apply background music in your videos. I just started watching your video and it's killing my ears. i cannot even listen your words.

    • @Cominlooks
      @Cominlooks  3 года назад +2

      Sure

    • @a.9589
      @a.9589 3 года назад +2

      i can hear perfectly

  • @gabogc
    @gabogc 3 года назад +2

    U need to edit your sound better! i cant not even hear your voice under the music. Great tip!

  • @kamleshsahane9439
    @kamleshsahane9439 3 года назад

    [3]
    !nvidia-smi
    NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running. getting this

    • @Cominlooks
      @Cominlooks  3 года назад

      Have you set your runtime with GPU?

    • @somesh7010
      @somesh7010 3 года назад

      @@Cominlooks is it working now i get only Tesla T4 pls reply

  • @Whatyelooking
    @Whatyelooking 3 года назад +3

    import os
    def restart_runtime():
    os.kill(os.getpid(), 9)
    if ('Tesla P100-PCIE...' in GPU_model):
    print("You are good to go")
    else:
    print("Changing GPU")
    restart_runtime()

    • @robrob2835
      @robrob2835 3 года назад

      doesn't work for me: "File "", line 3
      os.kill(os.getpid(), 9)
      ^
      IndentationError: expected an indented block"

    • @forwork5516
      @forwork5516 3 года назад +1

      @@robrob2835 you should put tab space to the os.kill line

    • @detonasmart9848
      @detonasmart9848 3 года назад

      @@forwork5516 help-me ?

    • @detonasmart9848
      @detonasmart9848 3 года назад

      File "", line 6 print("You are good to go") ^ IndentationError: expected an indented block

    • @leonspremo2739
      @leonspremo2739 3 года назад

      @@detonasmart9848 press tab

  • @mipe7755
    @mipe7755 3 года назад +1

    If you choose TPU instead of GPU, is it going to run faster?

    • @Cominlooks
      @Cominlooks  3 года назад +2

      If the application support tpu then better choose that.

  • @nevillechristopherraja3470
    @nevillechristopherraja3470 3 года назад +3

    I'm unable to hear your voice, please stop the background music

    • @Cominlooks
      @Cominlooks  3 года назад +1

      Sure will do from next time

  • @osamaelawad1428
    @osamaelawad1428 4 года назад +3

    Believe me please , you are the best , thank you too much ,please would you make easy tutorial also about deepfake like same face deep fake but in dance or fighting movie? , PLEASE

  • @MegaJaddi
    @MegaJaddi 4 года назад

    Tanks very imported 👌👌👌

  • @salmansourov8932
    @salmansourov8932 3 года назад

    not working now

  • @raminafrah3418
    @raminafrah3418 2 года назад

    what the hell music is playing in background. mute it pls

  • @EthioGoldenAge-uu8ev
    @EthioGoldenAge-uu8ev Год назад

    I am processing big data and I really challenged .

    • @Futureyouth-be1bo
      @Futureyouth-be1bo 6 месяцев назад

      did you solve the problem i am facing now pro tell me how you did if you did ?

  • @a.9589
    @a.9589 3 года назад

    so wich is the best?

    • @Cominlooks
      @Cominlooks  3 года назад

      It depends but most of the time its p100

  • @annwang5530
    @annwang5530 4 года назад

    So which is the fastest?

    • @Cominlooks
      @Cominlooks  4 года назад +1

      In almost every case, tela p100 is the fastest

    • @annwang5530
      @annwang5530 4 года назад

      @@Cominlooks thanks cute

    • @adamfils
      @adamfils 3 года назад

      With ColabPro+ has A100 which is the fastest

  • @Pinkyrosebithy1
    @Pinkyrosebithy1 4 года назад

    voice clone not working

    • @Cominlooks
      @Cominlooks  4 года назад

      Please do check that video (AI clones voice)(ruclips.net/video/2duHg9XFgFw/видео.html). I have responded to your comment. Try the new approach. Hope it solves your issue

  • @soumyaranjansethi1790
    @soumyaranjansethi1790 4 года назад +1

    Nice 👌👌

  • @hanzjeoelbernabe6819
    @hanzjeoelbernabe6819 4 года назад

    Please give the script... I'm stuck

    • @Cominlooks
      @Cominlooks  4 года назад +1

      All you need is "!nvidia-smi" you dont need any script..

  • @jaimelannister8160
    @jaimelannister8160 4 года назад

    yeah man u the best

  • @relaxingsounds7256
    @relaxingsounds7256 4 года назад

    Thanks a lot .. how to update the gpu drivers in the colab vps

    • @Cominlooks
      @Cominlooks  4 года назад

      The same way

    • @relaxingsounds7256
      @relaxingsounds7256 4 года назад

      @@Cominlooks hey .. hope you have a good time .. What do you mean by the same way ?

    • @relaxingsounds7256
      @relaxingsounds7256 4 года назад

      when I try to open a game in steam or even use the hardware encoder (NVCC ) to stram on OBS studio I get a message that i must update the GPU drivers etc please help me out
      thanks again

  • @sinaasadiyan
    @sinaasadiyan 4 года назад +2

    👌👌👌

  • @mr.schaeffer5399
    @mr.schaeffer5399 4 года назад

    Bullshit you are just saying to try until it finds it the best gpu but that is obvious!! Why make a video about it...🙄