ONNX na Model Interoperability
ONNX (Open Neural Network Exchange) bụ usoro ọkọlọtọ mepere emepe maka ịnọchite anya ụdị mmụta igwe ka ha nwee ike ịgagharị n'enweghị ihe ọ bụla n'etiti etiti na oge ọsọ.
Nchịkọta
It lets you train a model in one tool, like PyTorch, and deploy it in another environment without rewriting it.
Ime miri emi
Usoro dị iche iche (PyTorch, TensorFlow, scikit-learn) na-echekwa ụdị n'ụdị na-ekwekọghị, nke na-eme ka ntinye ya na-egbu mgbu. ONNX, nke ewepụtara na 2017 site na Microsoft na Facebook ma ugbu a na-achịkwa n'okpuru Linux Foundation, na-edozi nke a site n'ịkọwa usoro faịlụ nkịtị na nhazi nke ndị na-arụ ọrụ (dị ka Conv, MatMul, Relu) nke na-akọwa ihe nlereanya dị ka eserese mgbakọ. Ị na-ebupụ ụdị zụrụ azụ na faịlụ .onnx, na oge ọ bụla dakọtara nwere ike ibu ya. Oge ojiri oge ONNX na-eme eserese ahụ nke ọma n'ofe ngwaike dị iche iche, na-etinye njikarịcha dị ka njikọta ndị ọrụ na ọnụọgụ, yana ịgbakọ gbakọọ azụ azụ dị ka CPUs, NVIDIA GPUs (site na TensorRT), ma ọ bụ ngwa ngwa pụrụ iche. Nke a na-ewepụ ọzụzụ nlereanya site na ntinye.
Nghọta nka nka
Ihe Nlereanya ONNX bụ eserese ngụkọ serialized: ọnụ bụ ndị na-arụ ọrụ sitere na setịpụ onye ọrụ mbipute (opset), na akụkụ na-ebu tenors nwere ụdị na ụdị akọwapụtara. Ndị mbupụ na-achọpụta ma ọ bụ detuo ihe nlereanya gị iji weghara eserese a. N'ịtụle, ONNX Runtime partitions the graf n'ofe 'ndị na-enye ọrụ' (CPU, CUDA, TensorRT, wdg), onye ọ bụla na-ejizi ndị na-arụ ọrụ ọ na-akwado nke ọma, ma na-etinye nkwalite ọkwa ọkwa eserese dị ka mpịaji mgbe niile na ngwakọta ọnụ iji mee ngwa ngwa.
Mmetụta atụmatụ
Ọnụ ego na mmefu ego
Mkpebi ihe owuwu ụlọ na-akwalite arụmọrụ yana ọnụ ahịa ọrụ ruo ọtụtụ afọ.
Mkpebi doro anya
Nkà mmụta nka na-enyere ndị otu egwuregwu aka ịhọrọ nchịkọta ziri ezi, ọ bụghị naanị nke kachasị ọhụrụ.
Quality akara
Nhọrọ injinia ka mma na-ebelata ihe omume ntụkwasị obi na mmepụta.
Ọdịnihu nke ONNX na Model Interoperability
ONNX na-eme onwe ya ka ọ bụrụ asụsụ asụsụ maka mbugharị ihe atụ, ọkachasị maka ozi ihu na cross-platform. Na-atụ anya mkpuchi mkpuchi onye ọrụ maka nnukwu ụdị asụsụ na mgbanwe mgbanwe, nkwado siri ike maka ntinye ọnụ ọgụgụ na obere ntakịrị, yana njikọta miri emi na oge ndị na-ere ngwaike. Dị ka usoro gburugburu ebe obibi nke ibe AI pụrụ iche na-eto eto, usoro na-adịghị na-ere ahịa dị ka ONNX na-aghọwanye ihe bara uru, na-ahapụ ndị otu na-agbanwe ngwaike na-enweghị ụdị nhazigharị, na ONNX Runtime na-aga n'ihu na-agbasawanye n'ime mobile na web (site na WebAssembly).
Mmejuputa n'ezie n'ụwa
Na-ebupụ ihe nhazi ihe oyiyi PyTorch na ONNX ma jiri ONNX Runtime mee ya na ihe nkesa mmepụta C++ na-enweghị ndabere Python.
Na-ebuga ihe nlereanya na mkpanaka ma ọ bụ ihe nchọgharị site na ONNX Runtime Web (WebAssembly) maka ntinye aka na ngwaọrụ.
Na-eme ngwa ngwa ngwa ngbanwe mbupụ na NVIDIA TensorRT dị ka onye na-eweta ONNX Oge ojiri gaa maka igbu oge dị ala.
Ịkọwapụta ụdị ONNX ka ọ bụrụ int8 iji belata nha ya ma mee ka ntinye aka na CPUs ọnụ.
Ihe ize ndụ & okporo ụzọ nche
Ịkwalite otu akara ngosi nwere ike zoo adịghị ike sistemụ sara mbara.
A na-eledakarị ihe akụrụngwa na ụgwọ ọrụ anya.
Ọdịiche nchekwa na nleba anya nwere ike itolite ka sistemu na-adịwanye mgbagwoju anya.
Map mmejuputa
Kọwaa latency, ịdịmma na ebumnuche ọnụ ahịa tupu mmejuputa ya.
Benchmark n'okpuru ibu dị adị na ọnọdụ data.
Nleba anya akụrụngwa maka mperi, ịkpafu na mmetụta onye ọrụ.
Kwadebe ụzọ nzaghachi azụghachi azụ na ihe omume tupu ịchachaa.
Nọgide na-eme nchọpụta
Free newsletter
Get the daily AI briefing
Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Take the ONNX and Model Interoperability quiz
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
Ntuziaka na-esote
Ngwakọta Model
Ajụjụ a na-ajụkarị
What is ONNX and Model Interoperability?
ONNX (Open Neural Network Exchange) bụ usoro ọkọlọtọ mepere emepe maka ịnọchite anya ụdị mmụta igwe ka ha nwee ike ịgagharị n'enweghị ihe ọ bụla n'etiti etiti na oge ọsọ. Ọ na-enye gị ohere ịzụ ihe nlereanya n'otu ngwá ọrụ, dị ka PyTorch, ma tinye ya na mpaghara ọzọ na-edeghachighị ya.
Kedu isi nsogbu ONNX na-edozi isi?
ONNX na-akọwa usoro nkekọrịta ka enwere ike ibunye ụdị a zụrụ n'otu usoro na gburugburu ebe ọzọ na-edoghị ya ọzọ.
Kedu otu faịlụ ONNX si anọchi anya ihe nlereanya?
Ihe Nlereanya ONNX bụ eserese mgbako nke ọnụ ya bụ ndị na-arụ ọrụ ahaziri ahazi (dị ka Conv, MatMul, Relu) ndị tenors jikọtara.
Kedu ụlọ ọrụ bidoro ONNX na 2017?
Microsoft na Facebook webatara ONNX na 2017 ma na-akwado ya ugbu a n'okpuru Linux Foundation.
Na ONNX Oge ojiri gaa, kedu ihe bụ 'onye na-eweta ọrụ'?
Ndị na-eweta igbu egbu bụ nkwụnye azụ azụ (CPU, CUDA, TensorRT, na ndị ọzọ) nke ONNX Runtime na-eji na-agba ọsọ ndị ọrụ nke ọ bụla na-akwado kacha mma.
Kedu ihe ụdị 'opset' (operator set) na ụdị ONNX na-akọwa?
Ụdị oset na-akọwapụta nke ahaziri ahazi na ụdị ndị na-arụ ọrụ ihe nlereanya a dabere na ya, na-eme ka oge ọsọ ọsọ dakọtara na-akọwa ya nke ọma.