Cishe uke waphawula ukuthi i-inthanethi igcwele izithombe ezidalwe ubuhlakani bokwenziwa. Abaningi bayazibuza ukuthi kungenzeka yini ukufeza leyo miphumela ekhaya. ngaphandle kokukhokha izikhokhelo zanyanga zonke kanye nokugcina ubumfihlo obuphelele bezicelo zakho. Impendulo inguyebo ozwakalayo, futhi ithuluzi elinamandla kakhulu lokufeza lokhu okwamanje yi-ComfyUI, isikhombimsebenzisi, nakuba siyesabisa ekuqaleni, esikunikeza ukulawula okuphelele ekudalweni okubonakalayo.
Ngokungafani nezinketho ezilula, i-ComfyUI isebenza ngokusebenzisa uhlelo lwegrafu lapho uxhuma khona amabhokisi noma ama-node ahlukene. Lokhu kusho ukuthi awunqunyelwe yifomu eqinile, kodwa ungenza... klama ipayipi lakho lokukhiqizaKungakhathaliseki ukuthi uneseva enamandla ye-Linux noma i-laptop ye-Windows, ukusetha le ecosystem kuzokuvumela ukuthi uzame amamodeli wakamuva, njenge-FLUX noma i-Stable Diffusion, ngaphandle kokuthi othile ahlole lokho okudalayo.
Izidingo Zehadiwe: Umhlaba we-VRAM
Ngaphambi kokuthi uqale ukufaka, sidinga ukukhuluma ngalokho okubaluleke ngempela: imemori yevidiyo (i-VRAM). Akunandaba ukuthi iprosesa yakho inamandla kangakanani noma ukuthi une-RAM ejwayelekile kangakanani; uma ikhadi lehluzo lingasekeli ikhadi lehluzo, uhlelo luzosebenza ngesivinini esikhulu noma lumane luphahlazeke. Ukuze uhlole isimo sakho samanje se-VRAM ku-Linux, ungasebenzisa umyalo. i-nvidia-smi ye-NVIDIA noma i-rocm-smi ye-AMD.
Uma une-4GB ye-VRAM, lungela ukusebenza okuhamba kancane futhi usebenzise izinqumo eziphansi. Nge-8GB ungakwazi ukusingatha imisebenzi eminingi kahle, yize emidlalweni enzima njenge-FLUX.2 kuzodingeka usebenzise Ukulinganiswa kwe-GGUF (njenge-Q4_K_M)Lokhu kunciphisa usayizi wemodeli ngaphandle kokulahlekelwa ikhwalithi. Uma unenhlanhla yokuba ne-RAM engu-16 GB noma ngaphezulu, njengakwi-RTX 3090 noma i-4090, ungalayisha amamodeli e-FP8 noma ngisho ne-FP16 ngaphandle kokujuluka.
Ukufakwa Kwesinyathelo Ngesinyathelo Ezinhlelweni Ezihlukene
Kuye ngesistimu yakho yokusebenza, inqubo iyahlukahluka. Ku-Windows, indlela elula ukulanda i- iphakheji ephathekayo esemthethweniYivule bese usebenzisa ifayela le-.bat elihambisana ne-GPU yakho. Lena yindlela esheshayo ngoba isivele ifaka i-Python kanye nokuncika okulungiselelwe, okugwema ukungqubuzana neziguquguquko zemvelo.
Uma ukhetha i-Linux noma i-macOS, inqubo ilula kakhulu kodwa ihlanzekile. Okokuqala, udinga ukulingisa indawo yokugcina izinto evela ku-GitHub bese udala i- Imvelo ebonakalayo yePython (venv)Lokhu kubalulekile ukuze ugweme ukuphula imitapo yolwazi yesistimu yokusebenza. Uma indawo isisebenza, faka izinto ezixhomeke kuzo ezibalwe kufayela le-requirements.txt kanye nenjini yezibalo ye-PyTorch. Kubasebenzisi be-NVIDIA, inguqulo [inombolo yenguqulo engekho embhalweni wokuqala] iyanconywa. I-CUDA 12.1kuyilapho abasebenzisi be-AMD kumele bakhethe inguqulo ye ROCm.
Ukucushwa Okuthuthukisiwe kanye namaNode Abalulekile
Uma usuqalise i-ComfyUI okokuqala ku-http://127.0.0.1:8188, uzobona ideshibhodi egcwele amakhebula. Ukuze ugweme ukudideka, into yokuqala okufanele uyifake yi- Umphathi we-ComfyUILe plugin iyigugu elibalulekile, njengoba ikuvumela ukuthi useshe futhi ufake amanye ama-node enziwe ngokwezifiso, futhi okubaluleke kakhulu, ufake ngokuzenzakalelayo noma yimaphi ama-node angekho uma ungenisa umsebenzi womunye umuntu.
Ukuze usebenze namamodeli esimanje afana ne-FLUX, uzodinga i-node I-ComfyUI-GGUFNgokungafani ne-Stable Diffusion, lapho yonke into ivame ukungena khona kufolda yezindawo zokuhlola, i-FLUX idinga ukuhlukanisa izingxenye zayo: imodeli yokusabalalisa (i-UNet) iyangena amamodeli/unet/I-text encoder (CLIP) itholakala ku-models/clip/ kanti i-VAE itholakala ku-models/vae/. Uma uyibeka endaweni engafanele, i-loader ngeke iwathole futhi uzothola isikrini esingenalutho.
Ukuqonda kahle ukuhamba komsebenzi kanye ne-API
Ukugeleza okuyisisekelo kuhlanganisa ukulayisha imodeli, ukubhala i-prompt, ukudlula ku-sampler (lapho kwenzeka khona umlingo wokusabalalisa), kanye nokubhala ikhodi yesithombe. Icebo elibalulekile le-FLUX ukugcina I-CFG ku-1.0 bese usebenzisa izinyathelo ezimbalwa (phakathi kuka-4 no-8) nge-Euler sampler; uma ukhulisa i-CFG, izithombe zizophuma zishile futhi zinemibala engenangqondo.
Kulabo abanehadiwe kuseva futhi abafuna ukusebenza besebenzisa enye ikhompyutha, i-ComfyUI ivumela acceso remoto mediante el parámetro –listen 0.0.0.0Lokhu kuvula i-HTTP API ekuvumela ukuthi uthumele imisebenzi ngefomethi ye-JSON futhi uthole isithombe esicutshunguliwe. Kuyindlela ekahle yokuguqula i-PC yakho ene-GPU ibe isevisi yokukhiqiza izithombe yangasese, ekuvumela ukuthi udale izithombe eziningi ngokuzenzakalela zokumaketha noma ze-e-commerce usebenzisa izikripthi ze-Python.
Ukuxazulula izinkinga Nokuthuthukisa
Kuvamile ukuhlangana nephutha le- I-CUDA Ayisekho EmemoriUma lokhu kwenzeka kuwe, ungaqala kabusha i-ComfyUI ukuze ususe ukuhlukana kwememori noma uqale uhlelo ngomyalo –lowvramokuhambisa idatha phakathi kwe-GPU ne-RAM yesistimu. Uma kwenzeka i-AMD, uma ikhadi lingatholakali, ngokuvamile lixazululwa ngokwengeza umsebenzisi emaqenjini ukudweba kanye nevidiyo noma usebenzisa i-HSA_OVERRIDE_GFX_VERSION environment variable.
Uma ubona ukusebenza okungekuhle, hlola ukuthi awusebenzisi i-CPU ngephutha. Futhi, cabanga ngokuthuthukisa i-RAM yesistimu yakho ibe ngu-32GB noma u-64GB, njengoba i-ComfyUI iyisebenzisa njengebhuloho lapho i-VRAM inganele. Ukuze uthole indawo yokugcina, i- idrayivu ye-NVMe esheshayo Lokhu kubalulekile, njengoba ukulayisha amamodeli angu-10 GB njalo uma ushintsha imisebenzi kungaba yinto engapheli kudiski yomshini.
Ukusetha indawo yakho yokusebenza yezithombe yendawo kukukhulula emikhawulweni yamafu kanye nezindleko zesikweletu. Kusukela ekuphathweni kwe-VRAM kanye nokusethwa kwemvelo ebonakalayo kuya e-API ezenzakalelayo kanye nokusetshenziswa kwamamodeli e-GGUF quantized, manje usunawo wonke amathuluzi okuguqula i-RTX ibe yifektri yobuciko bedijithali, uthuthukise isinyathelo ngasinye ngokusebenzisa ama-node futhi uqinisekise ukuthi ihadiwe isebenza kahle kakhulu.