GitPlanet

Cheap and reliable Node.js hosting starts at $3/month, and $1/month static HTML hosting

Created with love in Canada, visit hostnodejs.com today

Feel like to post an Ad? Learn Details
All Projects โ†’ jahongir7174 โ†’ EffcientNetV2

jahongir7174 / EffcientNetV2

Licence: Apache-2.0 license
EfficientNetV2 implementation using PyTorch

Programming Languages

python
139335 projects - #7 most used programming language
shell
77523 projects

Labels

trainingpytorchimagenetefficientnetv2

Projects that are alternatives of or similar to EffcientNetV2

Dawn Bench Entries
DAWNBench: An End-to-End Deep Learning Benchmark and Competition
Stars: โœญ 254 (+170.21%)
Mutual labels:  training, imagenet
ModelZoo.pytorch
Hands on Imagenet training. Unofficial ModelZoo project on Pytorch. MobileNetV3 Top1 75.64๐ŸŒŸ GhostNet1.3x 75.78๐ŸŒŸ
Stars: โœญ 42 (-55.32%)
Mutual labels:  training, imagenet
efficientnetv2.pytorch
PyTorch implementation of EfficientNetV2 family
Stars: โœญ 366 (+289.36%)
Mutual labels:  imagenet, efficientnetv2
Php Certification Training
๐ŸŽ“ Zend Certified PHP Engineer Certification (free training)
Stars: โœญ 181 (+92.55%)
Mutual labels:  training
React Workshop
The course material for our React Hooks workshop
Stars: โœญ 184 (+95.74%)
Mutual labels:  training
Dynamic Training Bench
Simplify the training and tuning of Tensorflow models
Stars: โœญ 210 (+123.4%)
Mutual labels:  training
Lunatech Scala 2 To Scala3 Course
Lunatech course - "Moving forward from Scala 2 to Scala 3"
Stars: โœญ 174 (+85.11%)
Mutual labels:  training
Symfony Pack
A series of questions to prepare for the Symfony certification
Stars: โœญ 241 (+156.38%)
Mutual labels:  training
Improved Body Parts
Simple Pose: Rethinking and Improving a Bottom-up Approach for Multi-Person Pose Estimation
Stars: โœญ 202 (+114.89%)
Mutual labels:  training
Hyperlpr Training
่ฝฆ็‰Œ่ฏ†ๅˆซ่ฎญ็ปƒไปฃ็  Training Tools Related to HyperLPR
Stars: โœญ 190 (+102.13%)
Mutual labels:  training
Training
๐Ÿ A fast, easy and collaborative open source image annotation tool for teams and individuals.
Stars: โœญ 2,615 (+2681.91%)
Mutual labels:  training
Germanwordembeddings
Toolkit to obtain and preprocess german corpora, train models using word2vec (gensim) and evaluate them with generated testsets
Stars: โœญ 189 (+101.06%)
Mutual labels:  training
Easymodels
Easily find and view pre-trained AI models and deep learning projects through the command line ๐Ÿ’ป
Stars: โœญ 186 (+97.87%)
Mutual labels:  training
Purrr Tutorial
A introduction to purrr
Stars: โœญ 222 (+136.17%)
Mutual labels:  training
Training Material
A collection of Galaxy-related training material
Stars: โœญ 184 (+95.74%)
Mutual labels:  training
Malware Samples
Malware samples, analysis exercises and other interesting resources.
Stars: โœญ 241 (+156.38%)
Mutual labels:  training
Frida Boot
Frida Boot ๐Ÿ‘ข- A binary instrumentation workshop, with Frida, for beginners!
Stars: โœญ 179 (+90.43%)
Mutual labels:  training
Apprenticeships.me
Dev & design apprenticeships to the rescue! ๐Ÿ”ง
Stars: โœญ 188 (+100%)
Mutual labels:  training
Split Folders
๐Ÿ—‚ Split folders with files (i.e. images) into training, validation and test (dataset) folders
Stars: โœญ 203 (+115.96%)
Mutual labels:  training
R-Bridge-Tutorial-Notebooks
Jupyter notebooks demonstrating setup and use of the R-ArcGIS bridge. The repo includes datasets required to run the Jupyter notebooks.
Stars: โœญ 56 (-40.43%)
Mutual labels:  training
View All Similar Projects โž”

EfficientNetV2 implementation using PyTorch

Steps

  • imagenet path by changing data_dir in main.py
  • bash ./main.sh $ --train for training model, $ is number of GPUs
  • EfficientNet class in nets/nn.py for different versions

Note

  • the default training configuration is for EfficientNetV2-S

Parameters and FLOPS

  • python main.py --benchmark
Number of parameters: 21458488
Time per operator type:
        1504.95 ms.    80.5982%. Conv
        225.509 ms.    12.0772%. Sigmoid
        115.112 ms.     6.1649%. Mul
        12.7341 ms.   0.681982%. Add
        7.50523 ms.   0.401946%. AveragePool
        1.40185 ms.  0.0750768%. FC
      0.0112697 ms. 0.000603555%. Flatten
        1867.22 ms in Total
FLOP per operator type:
        16.7287 GFLOP.     99.708%. Conv
      0.0412707 GFLOP.   0.245986%. Mul
     0.00516096 GFLOP.  0.0307609%. Add
       0.002561 GFLOP.  0.0152643%. FC
        16.7777 GFLOP in Total
Feature Memory Read per operator type:
        291.409 MB.    51.8224%. Mul
        224.497 MB.    39.9231%. Conv
        41.2877 MB.    7.34234%. Add
        5.12912 MB.   0.912131%. FC
        562.323 MB in Total
Feature Memory Written per operator type:
        165.083 MB.    50.2087%. Mul
        143.062 MB.    43.5114%. Conv
        20.6438 MB.    6.27867%. Add
          0.004 MB. 0.00121657%. FC
        328.793 MB in Total
Parameter Memory per operator type:
        79.9537 MB.    93.9773%. Conv
          5.124 MB.    6.02273%. FC
              0 MB.          0%. Add
              0 MB.          0%. Mul
        85.0777 MB in Total

Results

  • python main.py --test for trained model testing
name resolution acc@1 acc@5 #params FLOPS resample training loss
EfficientNetV2-S 384x384 83.9 96.7 21.46 16.7777 BILINEAR CrossEntropy
EfficientNetV2-S 384x384 - - 21.46 16.7777 BILINEAR PolyLoss
EfficientNetV2-M - - - - - - -
EfficientNetV2-L - - - - - - -
Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].