Uyifaka njani i-DeepFaceLab kwiLinux ukwenza iividiyo ezinzulu

Luis Lopez

Imizuzu ye2

Iminyaka eliqela kwi-intanethi kukho oko kuthiwa nzulu iividiyo, kodwa kwiinyanga nje ezidlulileyo ukuthandwa kwayo kuye kwanda kakhulu ngenxa yezixhobo ezahlukeneyo kunye nokusetyenziswa okwenza kube lula ukwenza.

I-Deepfake yindlela esetyenziselwa ukuguqula imifanekiso okanye iividiyo kusetyenziswa ubukrelekrele bokufakelwa, kwiimeko ezininzi isetyenziselwa ukuphakamisa ubuso ngobuso, ngaloo ndlela kusenziwa iividiyo ezingezizo ezineziphumo ezintle.

Ngaphambili, ukwenza iividiyo ezinzulu yayisisakhono esasithatha iminyaka ukuba sifunde, kodwa ngoku, ngenkqubela phambili yetekhnoloji, kulula kakhulu ukuyenza.

I-DeepFaceLab sisixhobo esifumanekayo kwiLinux kunye nezinye iinkqubo ezikuvumela ukuba wenze iividiyo ezinzulu ngendlela elula kakhulu usebenzisa umgca wokuyalela.

Kungabonakala ngathi ukusebenzisa i-DeepFaceLab kunzima kakhulu ukuyisebenzisa, kodwa akukho nto ikude ngokuchaseneyo, ngokusebenzisa nje izikripthi ezimbalwa onokuguqula ividiyo kwaye wongeze ubuso obutsha kwi-protagonist.

Uyifaka njani i-DeepFaceLab kwiLinux

Ukufaka i-DeepFaceLab kuya kufuneka kuqala ufake i-Anaconda3 usebenzisa i iphepha elisemthethweni kwaye kuya kufuneka uyiqale ngale miyalelo ilandelayo:

ukuthumela ngaphandle PATH = ~ / anaconda3 / bin: $ PATH conda init bash

Ngoku ungaqhubekeka kufakelo lwe-DeepFacebLab ngale migca ilandelayo yekhowudi:

i-conda yenza -y -n-deepfacelab python = 3.6.6 cudatoolkit = 9.0 cudnn = 7.3.1 conda activate deepfacelab git clone https://github.com/lbfs/DeepFaceLab_Linux.git cd DeepFaceLab_Linux python -m pip install -r requirements-cuda .txt

Ukuba une-Ubuntu 16.04 okanye i-18.04 ungajonga iphepha lokwakha elisemthethweni ngeendlela zokufaka i-DeepFaceLab kuzo zombini iinkqubo.

Sebenzisa i-DeepFaceLab kunokuba yinkohliso kodwa wakube ufundile oku kusoloko kukho umhla wokufundisa uya kuba nakho ukwenza iividiyo zakho ezinzulu njenge pro.

Ukuba unengxaki yokufaka okanye ukusebenzisa i-DeepFaceLab, siya kukuvuyela ukukunceda kwizimvo.


Shiya uluvo lwakho

Idilesi yakho ye email aziyi kupapashwa. ezidingekayo ziphawulwe *

*

*

  1. Inoxanduva lwedatha: I-AB Internet Networks 2008 SL
  2. Injongo yedatha: Ulawulo lwe-SPAM, ulawulo lwezimvo.
  3. Umthetho: Imvume yakho
  4. Unxibelelwano lwedatha: Idatha ayizukuhanjiswa kubantu besithathu ngaphandle koxanduva lomthetho.
  5. Ukugcinwa kweenkcukacha
  6. Amalungelo: Ngalo naliphi na ixesha unganciphisa, uphinde uphinde ucime ulwazi lwakho.

  1.   I-Rugher sitsho
    yenzayo 5 iminyaka

    Ndingayiqala njani inkqubo xa sele ifakiwe okanye ndingaqinisekisa njani ukuba iyasebenza?

    Phendula uRugher
  2.   UHUGO FERNANDO CARRERA TOASA sitsho
    yenzayo 5 iminyaka

    enkoso ngokwabelana
    njengoko ndisombulula oku kulandelayo.
    Impazamo: Ayifumananga ingxelo eyanelisa imfuno tensorflow-gpu == 1.12.0 (from -r requirements-cuda.txt (line 5)) (from versions: 1.13.0rc1, 1.13.0rc2, 1.13.1, 1.13.2 .1.14.0, 0rc1.14.0, 1rc1.14.0, 1.15.0, 0rc2.0.0, 0a2.0.0, 0b2.0.0, 1b2.0.0, 0rc2.0.0, 1rcXNUMX).
    Impazamo: Akukho luhambiso luthelekiso lufunyenweyo lwe-tensorflow-gpu == 1.12.0 (from -r requirements-cuda.txt (line 5))
    Nceda undincede

    Phendula HUGO FERNANDO CARRERA TOASA
  3.   Ricardo sitsho
    yenzayo 4 iminyaka

    Sawubona,

    Inginika impazamo efanayo noHUGO FERNANDO CARRERA TOASA izimvo

    Impazamo: Ayifumananga ingxelo eyanelisa imfuno tensorflow-gpu == 1.12.0 (from -r requirements-cuda.txt (line 5)) (from versions: 1.13.0rc1, 1.13.0rc2, 1.13.1, 1.13.2 .1.14.0, 0rc1.14.0, 1rc1.14.0, 1.15.0, 0rc1.15.0, 1rc1.15.0, 2rc1.15.0, 3rc1.15.0, 2.0.0, 0a2.0.0, 0b2.0.0, 1b2.0.0, 0rc2.0.0. , 1rc2.0.0, 2rc2.0.0, XNUMX)
    Impazamo: Akukho luhambiso luthelekiso lufunyenweyo lwe-tensorflow-gpu == 1.12.0 (from -r requirements-cuda.txt (line 5))

    Phendula uRicardo
  4.   UFrancisco Diaz Carsí sitsho
    yenzayo 3 iminyaka

    Sawubona,
    Uthi "Ukusebenzisa i-DeepFaceLab kunokuba yinto ekhohlisayo kodwa nje ukuba ufunde le nto ihlala ihlaziyiwe unokwenza iividiyo zakho ezinzulu njenge pro." kodwa iphi indawo yokusebenzisa?

    Enkosi ukubulisa.

    Phendula uFrancisco Díaz Carsí
  5.   Guillermo sitsho
    yenzayo 3 iminyaka

    Molo, molweni ekuseni, ndingathanda ukwazi ukuba umntu angandinceda na nge-deepface yewindows 10. Inqaku kukuba ndiyalikhuphela kwaye ndilandele amanyathelo esifundo sevidiyo, fakela abaqhubi bekhadi lemizobo le-nvidia, i cuda_9.0.176 _win10.exe, i-cuda_9.0.176.1_windows.exe patch, i-cudnn-9.0-windows10-x64-v7.6.5.32.zip ithala leencwadi kunye nenkqubo ye-deepface. Ndenze onke amanyathelo kodwa xa ndivula inkqubo ndifumana loo festile incinci ingenanto ethi: Changelog:
    -Iqela le-squirrel
    -Idatha yedatha eyongeziweyo
    Impazamo engekho ebusweni
    - Yongeza kwakhona ukhetho lokudibanisa
    -Uphuculelwe i-TF 1.5, CUDA 9
    Khumbula ukukhuphela kunye nokufaka ithala leencwadi eliphambili ngokunjalo. Nje ukuba ikhutshelwe kwaye ifakwe ngokomyalelo kwiforum, layisha kwakhona usetyenziso kwaye lo myalezo uza kuthi
    nyamalala.
    Oko kukuthi, le nkqubo ayindivulelanga. Ndingaluvuyela kakhulu naluphi na uncedo onokuthi undinike lona. Enkosi kakhulu. UWilliam

    Phendula u Guillermo