*Zvakakwana nekuda kwekudzidza Muchina Kudzidza, hongu. Pasi pekutarisa kusingagutsikane zvishoma kwemukadzi wake waaida.
Panogona kunge pasina application iri nyore kune iyo spinal reflexes seTinder. Kuti uishandise, iwe unongoda chigunwe chimwe chete chekutsvaira uye neuroni shoma kuti usarudze vasikana kana varume vaunoda zvakanyanya. Kuitwa kwakanaka kwe brute force mukusarudza vaviri.
Ndakasarudza kuti iyi ingave nzira yakanaka yekuwana kunzwa kushoma kwekudzidza kwemuchina pane kadhi nyowani yemifananidzo. Chasara kutsanangurira mukadzi wangu kuti ini handidi mukadzi akafuta, uye ndiri kungodzidzisa neural network.
Dambudziko nderei pakupfimba network?
Paiva nechinhu chakadaro - Ashley Madison. Yakananga, ine sirogani inoti βUpenyu hupfupi. Itai zvenyu." Vateereri vakuru varume vakaroora vari kutsvaga kusvirana paside. Kuita mari kunonakidza zvakare - mukuwedzera kune yakajairwa "shandisa mapoinzi kuda nekunyora," ivo vakakumbira $19 yekudzima account yemushandisi pasina kuteedzera.
Muna 2015, saiti yacho yakadonhedza uye 60 GB yedata remunhu rakaburitswa munzvimbo yeruzhinji. Pamusoro pemhuri zhinji dzakaparadzwa, kuvuza uku kwakapa ruzivo rwakawanda runonakidza kune vanoongorora. Ini ndaigara ndichifungidzira kuti kune varume vakawanda vakawanda panzvimbo dzekupfimbana, asi munyaya iyi zvakazonyanya kunakidza. Mutori wenhau Annalee Newitz yakawana kuti pavashandisi vane mamiriyoni mashanu, 5 12 chete vakanga vakafanana neakaundi evasikana chaivo uye vaishandiswa nguva dzose. Vamwe vese vaingove bots vaikurukura nevaenzi vechirume.
Kuwedzeredza kwakadai kumaakaundi echirume kwakajairwa kwete chete kune iyi sosi, asiwo kune dzakawanda dzimwe nzvimbo dzekufambidzana. Ndine chokwadi chokuti vazhinji vakasangana nemamiriro ezvinhu aya pasina mubvunzo, kana iwe uchifanira kunyatsoronga waunoziva, asi musikana anongoda kunyoresa. Ngatisiyei kunaka kweiyi mhomho yevateveri parutivi, asi chokwadi hachirambiki kuti chiyero chekugovera uye chinodiwa chinoshandurwa zvakajeka mukufarira vasikana.
Tinder Feature
Iyo yakakodzera brute force muhukama hwevakadzi
Chinhu chikuru chepuratifomu iyi mutengo wakaderera pane wekuziva. Kusangana kwemaswipe maviri kwakakwana uye unenge uchitotaurirana nemunhu angangove anonakidza. Dambudziko nderekuti kusaenzana kwevarume nevakadzi kunokonzera kuti vasikana vazhinji vanenge vaine mitambo yakawanda pazuva. Izvi zvinoreva kuti vangangove vasina nguva yekukuteerera iwe pakati pevamwe vavhoti.
Zvakanyatsojeka kuti chikuva chinoreva mukana mudiki wekuongorora iyo yakadzika yemukati nyika yemunhu kubva kune imwe nehafu kwesekondi kutarisa pamufananidzo mune yekushambira kana kutyaira yefashoni tinted mota. Naizvozvo, kana iwe usingataridzi sehumwari mumifananidzo yako, hauna sarudzo kunze kwekuwedzera mikana yako nekutora. mune mamwe marudzi. Zvichitaurwa zviri nyore, isu tichamanikidza hutsinye uye kutora huwandu kuti tiwedzere mikana yedu yekubudirira kubereka. Sezvo iwe dzimwe nguva uchifanirwa kutsauswa nechikafu uye kurara, uye swipes ishoma, iwe ungangoda kuti otomatiki sarudza vasikana kana varume vanonyatsoenderana nezvaunofarira. Short redheads kana akareba brunettes - zviri kwauri.
Kuunganidza data
Chekutanga pane zvese, iwe unoda data rakawanda rekurongeka chaiko. Chero ani akasangana nemichina yekudzidza anoziva kuoma kwazvingaite kugadzira dataset rakaunganidzwa nemazvo. Nechepfungwa, chero sosi yakafanana ichave yakakodzera sesosi yedata, ingave Instagram kana mamwe masocial network. Asi zviri nani kudzidzisa pane iwo masampuli ayo network ichashanda mune ramangwana.
Ngatitorei repository sehwaro . Mapikicha eTinder anogara achiwanikwa pachena, asi basa re "se" rakatoganhurwa. Naizvozvo, zvinodikanwa kuburitsa zvese zvipenyu mukati meradius uye nekunyatso kumaka. Kutanga iwe unofanirwa kushandisa script iri nyore:
from skimage.io import imread, imsave, imshow, show
import matplotlib.pyplot as plt
import pynder
from helpers import get_access_token, get_login_credentials
from io_helper import save_image
email, password, FBID = get_login_credentials()
FBTOKEN = get_access_token(email, password)
session = pynder.Session(facebook_token=FBTOKEN)
while True:
users = session.nearby_users()
for user in users:
photos = user.get_photos()
print("Fetched user photos..")
for photo in photos:
print(photo)
image = imread(photo)
imshow(image)
show()
input_string = "Write 1 to like. Write 2 to dislike."
ans = str(input(input_string)).lower()
if ans == "1":
save_image(image, photo, True)
else:
save_image(image, photo, False)
Ichakubvumidza iwe kumaka iyo dataset nekukurumidza sezvinobvira nemabhatani maviri chete. Gomba rakakosha riri pakuti raibhurari yewerkzeug yakatyora kuenderana nekumashure uye ichafanira kumanikidzwa kuidzikisa. Zvikasadaro inokanda kukanganisa uku.
Traceback (most recent call last):
File "img_scrape.py", line 4, in <module>
from helpers import get_access_token, get_login_credentials
File "/home/someone/tmp/TinderAutomation/helpers.py", line 1, in <module>
import robobrowser
File "/home/someone/tmp/TinderAutomation/venv/lib/python3.6/site-packages/robobrowser/__init__.py", line 3, in <module>
from .browser import RoboBrowser
File "/home/someone/tmp/TinderAutomation/venv/lib/python3.6/site-packages/robobrowser/browser.py", line 8, in <module>
from werkzeug import cached_property
ImportError: cannot import name 'cached_property'
Naizvozvo, mune zvinodiwa.txt unofanira kunyora Werkzeug==0.16.1. Ipapo ichasimuka.
Dambudziko rechipiri nderekuwana chiratidzo ichi. Iyo yakajairwa nzira kubva kune repository haina kushanda kwandiri, asi ndakakwanisa kuitora kubva kune yekuvandudza console. Kuti uite izvi, enda ku uye bvisa mhinduro kune POST chikumbiro mukati . Mukati tinotarisa 'access_token'. Nokuda kwechimwe chikonzero hazvina kushanda kekutanga, asi ipapo ndakaiwana uye hardcode iyo mune script.
Dataset zvinodiwa
Pane akati wandei akakosha anodikanwa emuchina kudzidza datasets:
- Kukwana
- Uniformity
- Zvakasiyana-siyana
Kukwana mune iyi kesi kunoda anokwana zviuru gumi mafoto kuti uvake muenzaniso wakakwana. Hongu, zvakawanda. Ichi ndicho chikonzero nei masevhisi achida , kwaunogona kugovera chiratidzo che dataset kune vamwe vanhu nemubhadharo. Kune rimwe divi, unonyatsoda kuti bot yako ifarire inoshamisa mwedzi-akatarisana nevasikana vekuAsia kana vasikana vakanaka zvakaenzana vane midzi yeIndia? Zvakadaro, modhi yacho inofanirwa kuratidza kuravira kwako.
Iko hakuna mamwe matambudziko nekusiyana-siyana; mafoto ese anoratidzwa kubva kwakasiyana makona uye mwenje. Mumagirazi, madhirezi, swimsuits uye ski suits. Dambudziko rinogona kumuka nekufanana kwedataset. Sezvineiwo, kana tikanyora sampuli yedu, inofanira kunge ine zvikamu zvakaenzana. Kana iwe ukapedzisira uine "yakashongedzwa" dhatabheti, iwe uchafanirwa kuidzikisa nemifananidzo kubva kune mamwe masosi. Iwe unozofanirwa kuwedzera mamwe anoyevedza, kana zvinopesana, iwe unovasarudza zvichibva pane iyo markup mhedzisiro. Ndine chimwe chinhu chakakomberedza 60% chakanaka. Pamwe ini handina kunyanya kusarudzwa, kana kuti ndine rombo rakanaka uye kune vasikana vakanaka vakawanda.
Ini zvakare handidzikisi iyo hypothesis yekuti kune akawanda bots pakati pavo. Isu tinodzidzisa bot inoda mamwe mabhoti. Pane kusekesa mune izvi.
Kugadziriswa kwedata
Tine boka remapikicha akamaka, asi akasanganiswa zvakanyanya. Masikati, husiku, kubva kumashure uye nevamwe. Nekuzvidemba, ndinonzwisisa kuti kudzidzisa kubva kumifananidzo kubva kune reverse angle hakuzonyanya kushanda, sezvo sampuli ichave isina kuenzana. Naizvozvo, sarudzo yakanakisa ndeye kushandisa zviso sechiratidzo che "kunaka". Zvakadaro, kwatiri, sekune mamwe maprimates, iyi ndiyo yakakosha parameter.
Naizvozvo, ngatishandisei . Iyi yakanakisa algorithm iyo inokutendera iwe kuti uwane zviso mumifananidzo ine yakaderera muzana yekukanganisa kwakanaka kwenhema.
Izvi zvinotsanangurwa zvakadzama mubhuku rekushandisa
Padanho rinotevera, mushure mekunge zviso chete zviri mumuenzaniso, zvine musoro kubvisa ruvara. Sezvineiwo, haumbozofanire kusarudza pakati pePandora yakanaka yebhuruu denizen kana yegirinhi-ganda runako.
Muvanhu veHue, ruvara rweganda harwupi mupiro wakakura pakuyerwa kwekukwezva.
Naizvozvo, zvakakodzera kurerutsa basa reneural network uye kusiya chete greyscale.
Model kuvaka
Ini ndinoda kutaura ipapo ipapo kuti pasina yakanaka vhidhiyo kadhi uye CUDA, iwe kazhinji haungazowana yakadzidziswa modhi munguva yakakwana. Naizvozvo, nekukurumidza tarisa kuverenga mumakore akasarudzika kana kushandisa python-CUDA.
Ini ndakatora yakakosha-yechitatu-layer muenzaniso kubva kumunyori weiyo repository uye, zvinoshamisa, yakaratidza huchokwadi hwekutenderedza 72%, inova mhedzisiro yakanaka.
model = Sequential()
model.add(Convolution2D(32, 3, 3, activation='relu', input_shape=(img_size, img_size, 3)))
model.add(MaxPooling2D(pool_size=(2,2)))
model.add(Convolution2D(32, 3, 3, activation='relu'))
model.add(MaxPooling2D(pool_size=(2,2)))
model.add(Convolution2D(64, 3, 3, activation='relu'))
model.add(MaxPooling2D(pool_size=(2,2)))
model.add(Flatten())
model.add(Dense(128, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(2, activation='softmax'))
adam = optimizers.SGD(lr=1e-4, decay=1e-6, momentum=0.9, nesterov=True)
model.compile(loss='categorical_crossentropy',
optimizer= adam,
metrics=['accuracy'])
Kana paine sampuli yakanaka, saka inogona kunge yakakwana kuwana modhi inoshanda.
Ngatitangei bot
Kutenda kumunyori weiyo repository kune yakagadzirira-yakagadzirwa sarudzo yekukurumidza kuyedza iyo pfungwa. Muchokwadi, inoshanda nemazvo mune yekutanga vhezheni uye inogona, mumusimboti, kutangwa pane yedu . Izvo hazvichagone kudzidzisa parizvino; parizvino isu hatipe chaiwo michina ine CUDA rutsigiro rwekuverenga, asi unogona kumhanya chimwe chinhu 24/7 pasina matambudziko. Iyo bot yakareruka, saka zvingave zvine pundutso kutora mutero unobhadhara zviwanikwa zvinoshandiswa.
Mhinduro
Ndinofungidzira kuti ndiri tsvarakadenga. Uye ndine nyika yakapfuma yemukati. Ndakawana chimwe chinhu semachisi gumi nematatu mukati meawa. Uyezve, kakawanda vasikana vakatanga kunyora.
Nekuda kweizvozvo, takapedzisira tave nenhaurirano dzakanaka kwazvo, kwandakataura kuti ndauya kuzotamba nekudzidza kwemichina uye kunyora data. Mumwe wevasikana vacho aifarira zvikuru, sezvo iye pachake ari mugadziri. Pane kunzwa kwakasimba kuti anozopedzisira averenga chinyorwa ichi paHabrΓ©. Ndinovimba chaizvo kuti Oksana achachengetedza kusazivikanwa kwangu. π
*achisimudza ruoko ndokuti hi
Zvishoma nezverutivi rwetsika rwenyaya
Kutaura chokwadi, ini handifarire zano rese rekuita hukama pakati pevarume nevasikana. Pane chimwe chinhu chakanaka pamusoro pekukanda bhachi rako pamapfudzi emutorwa anotonhora akamira ega. Kana kuswedera kumusikana akanaka mune cafe yezhizha uye kunwa kofi pamwe chete. Ibva kuseri kwemamonitor kare.
Zhizha rakapoteredza. Yasvika nguva yekuzivana.
Source: www.habr.com