Title:
NEURAL NETWORK MODEL TRAINING METHOD AND APPARATUS, ELECTRONIC DEVICE AND MEDIUM
Document Type and Number:
WIPO Patent Application WO/2023/232031
Kind Code:
A1
Abstract:
The present application discloses a neural network model training method and apparatus, an electronic device and a storage medium. The method comprises: in response to a neural network model to be trained not satisfying a predetermined convergence condition, dividing a plurality of pre-collected pictures into N types; taking one of the N types as a current reference type, and extracting a picture from among pictures of the current reference type as a current reference picture; constructing a current positive sample pair and a current negative sample pair on the basis of the current reference picture, the pictures in the current reference type and the pictures in types other than the current reference type; and on the basis of the current positive sample pair and the current negative sample pair, training the neural network model, and repeatedly performing the above operations until the neural network model meets the convergence condition.
Inventors:
JIN BAIXIN (CN)
SHANG BINGXU (CN)
CHEN ZHIXIN (CN)
LIU YANG (CN)
WANG HONGFENG (CN)
ZHANG YONG (CN)
ZHANG ZHONGJU (CN)
HE LIU (CN)
SHANG BINGXU (CN)
CHEN ZHIXIN (CN)
LIU YANG (CN)
WANG HONGFENG (CN)
ZHANG YONG (CN)
ZHANG ZHONGJU (CN)
HE LIU (CN)
Application Number:
PCT/CN2023/097127
Publication Date:
December 07, 2023
Filing Date:
May 30, 2023
Export Citation:
Assignee:
CHINA FAW CO LTD (CN)
International Classes:
G06V40/18; G06N3/04; G06N3/08; G06V10/764; G06V10/774; G06V10/82
Foreign References:
| CN114898454A | 2022-08-12 | |||
| CN112418189A | 2021-02-26 | |||
| CN112085041A | 2020-12-15 | |||
| CN113191215A | 2021-07-30 |
Attorney, Agent or Firm:
BEIJING YZH IP FIRM (CN)
Download PDF:
Previous Patent: HIGH-TEMPERATURE SINGLE CRYSTAL ALLOY BRAZING MATERIAL AND METHOD
Next Patent: CHANNEL ACCESS METHOD AND APPARATUS
Next Patent: CHANNEL ACCESS METHOD AND APPARATUS