Raibhurari yemahara yeAI

Dzidza AI.Kusununguka nekusingaperi.

1019 magwara eChirungu, nzira dzakarongeka dzekudzidza, uye raibhurari yakavhurika — yakavakwa neyakazvimiririra 501(c)(3) isingabatsiri kuitira kuti chero munhu anzwisise AI yemazuva ano.

1019Mazano emahara
9Musoro wenyaya
~2 minPer gwara
~34hNguva yekuverenga

Tanga pano

Vashanu Mhedzisiro-yakavakirwa makosi

Kosi yega yega inosanganisira mhedzisiro yakajeka, hunyanzvi hwemepu, zviitiko zvekudzidzira, uye yakashandiswa capstone.

Musoro wenyaya

Bhurawuza nekumhanya

Svetukira munharaunda yaunofarira. Nzira yega yega ine akawanda akajeka-Chirungu magwaro.

Raibhurari yakazara

Zvese zvinotungamira

1019 ye 1019 Guides showed. Filter by track or search above.

Mutauro AI

Typical Sampling

Yakajairika sampling inzira yechizvarwa-yechinyorwa inotora izwi rinotevera kubva kune tokens iro ruzivo rwemukati runogara pedyo neiyo modhi inotarisirwa kushamisika…

2 VerengaVerenga
Mutauro AI

Mirostat Perplexity Kudzora

Mirostat idhikodhi algorithm iyo inoshingairira kuburitsa modhi yemutauro yakanangana nekukanganisika kunonangwa (seti yedanho rekushamisika) uchishandisa loop yemhinduro.

2 VerengaVerenga
Mutauro AI

Sycophancy muMutauro Mienzaniso

Sycophancy ndiyo tsika yemamodheru emitauro yeAI yekuudza vashandisi zvavanoda kunzwa, vachibvumirana nemafungiro akataurwa kana caving kusundira kumashure kunyangwe…

2 VerengaVerenga
Tekinoroji

Yechipiri-Odha Optimization uye Newton Nzira

Chechipiri-odha optimization inoshandisa curvature ruzivo (iyo yeHessian matrix yechipiri inotorwa) kutora nhanho dzakangwara kuenda kune hushoma, kwete kutsetseka chete.

2 VerengaVerenga
Basics

Kuora Kwehuremu uye L2 Regularization

Kuora huremu inzira yakapusa, ine simba inokwenya huremu hwemuenzaniso kuenda ku zero panguva yekudzidziswa, ichiodza moyo kubva pakuvimba zvakanyanya pane chero…

2 VerengaVerenga
Basics

Dropout uye Stochastic Regularization

Kudonhedza inzira yenguva dzose inodzima chidimbu chemaeuroni panguva yega yega nhanho yekudzidzira, ichimanikidza network kuti ivake isina simba, yakasimba ...

2 VerengaVerenga
Tekinoroji

RMSNorm uye Pre-Layer Normalization

RMSNorm inzvimbo yakareruka yekujairisa iyo inodzoreredza ma activation nemidzi yavo inoreva sikweya, uye pre-layer normalization nzvimbo dzinotsika pamberi peimwe neimwe…

2 VerengaVerenga
Tekinoroji

SwiGLU uye Gated Activations

SwiGLU igated activation function iyo inowedzera imwe mutsara fungidziro yekupinza neiyo Swish-yakagadziriswa yechipiri fungidziro, ichiita seinodzidzika ...

2 VerengaVerenga
Tekinoroji

Kusvina-uye-Kunakidzwa Networks

Squeeze-and-Excitation (SE) inovharira ita kuti convolutional network idzidze huwandu hwekuremedza chimwe nechimwe chiteshi, ichichidzokorora zvichienderana nemamiriro epasirese.

2 VerengaVerenga
Tekinoroji

Warmup uye Cosine Annealing Schedules

Warmup zvinyoro nyoro inokwidza mwero wekudzidza kubva padhuze ne zero usati wadzidzira, ipapo cosine annealing zvakanaka inoodza ichidzokera pasi ichitevera cosine curve.

2 VerengaVerenga
Tekinoroji

Cyclical Learning Rates

Mari yekudzidza yemacyclical inotenderedzazve mwero wekudzidza uchikwira nekudzika pakati pepakati nepamusoro pechisungo pane kungoora chete.

2 VerengaVerenga
Tekinoroji

Sharpness-Aware Minimization

Sharpness-Aware Minimization (SAM) inzira yekusimudzira isingatsvage kurasikirwa kwakaderera asi kurasikirwa kwakaderera munzvimbo yese yehuremu - furati…

2 VerengaVerenga

Wapedza kuverenga? Ndiratidze.

Check what you learned with a topic quiz, then explore our structured courses or work toward a certificate. Every guide stays free to read.