90 lines
3.3 KiB
Markdown
90 lines
3.3 KiB
Markdown
# Resnet-family Convolutional Neural Networks for Image Classification in Tensorflow
|
|
|
|
In this repository you will find implementation of Resnet and its variations for image
|
|
classification
|
|
|
|
## Table Of Contents
|
|
|
|
* [Models](#models)
|
|
* [Validation accuracy results](#validation-accuracy-results)
|
|
* [Training performance results](#training-performance-results)
|
|
* [Training performance: NVIDIA DGX A100 (8x A100 40G)](#training-performance-nvidia-dgx-a100-8x-a100-40g)
|
|
* [Training performance: NVIDIA DGX-1 (8x V100 16G)](#training-performance-nvidia-dgx-1-8x-v100-16g)
|
|
* [Release notes](#release-notes)
|
|
* [Changelog](#changelog)
|
|
|
|
|
|
## Models
|
|
|
|
The following table provides links to where you can find additional information on each model:
|
|
|
|
| **Model** | **Link**|
|
|
|-----------|---------|
|
|
| resnet50 | [README](./resnet50v1.5/README.md) |
|
|
| resnext101-32x4d | [README](./resnext101-32x4d/README.md) |
|
|
| se-resnext101-32x4d | [README](./se-resnext101-32x4d/README.md) |
|
|
|
|
## Validation accuracy results
|
|
|
|
Our results were obtained by running the applicable training scripts in the tensorflow-20.06-tf1-py3 NGC container
|
|
on NVIDIA DGX-1 with (8x V100 16G) GPUs. The specific training script that was run is documented in the corresponding model's README.
|
|
|
|
The following table shows the validation accuracy results of the
|
|
three classification models side-by-side.
|
|
|
|
|
|
| **arch** | **AMP Top1** | **AMP Top5** | **FP32 Top1** | **FP32 Top5** |
|
|
|:-:|:-:|:-:|:-:|:-:|
|
|
| resnet50 | 78.35 | 94.21 | 78.34 | 94.21 |
|
|
| resnext101-32x4d | 80.21 | 95.00 | 80.21 | 94.99 |
|
|
| se-resnext101-32x4d | 80.87 | 95.35 | 80.84 | 95.37 |
|
|
|
|
## Training performance results
|
|
|
|
### Training performance: NVIDIA DGX A100 (8x A100 40G)
|
|
|
|
Our results were obtained by running the applicable
|
|
training scripts in the tensorflow-20.06-tf1-py3 NGC container
|
|
on NVIDIA DGX A100 with (8x A100 40G) GPUs.
|
|
Performance numbers (in images per second)
|
|
were averaged over an entire training epoch.
|
|
The specific training script that was run is documented
|
|
in the corresponding model's README.
|
|
|
|
The following table shows the training accuracy results of the
|
|
three classification models side-by-side.
|
|
|
|
|
|
| **arch** | **Mixed Precision XLA** | **TF32 XLA** | **Mixed Precision speedup** |
|
|
|:-:|:-:|:-:|:-:|
|
|
| resnet50 | 16400 img/s | 6300 img/s | 2.60x |
|
|
| resnext101-32x4d | 8000 img/s | 2630 img/s | 3.05x |
|
|
| se-resnext101-32x4d | 6930 img/s | 2400 img/s | 2.88x |
|
|
|
|
### Training performance: NVIDIA DGX-1 (8x V100 16G)
|
|
|
|
Our results were obtained by running the applicable
|
|
training scripts in the tensorflow-20.06-tf1-py3 NGC container
|
|
on NVIDIA DGX-1 with (8x V100 16G) GPUs.
|
|
Performance numbers (in images per second)
|
|
were averaged over an entire training epoch.
|
|
The specific training script that was run is documented
|
|
in the corresponding model's README.
|
|
|
|
The following table shows the training accuracy results of the
|
|
three classification models side-by-side.
|
|
|
|
|
|
| **arch** | **Mixed Precision XLA** | **FP32 XLA** | **Mixed Precision speedup** |
|
|
|:-:|:-:|:-:|:-:|
|
|
| resnet50 | 9510 img/s | 3170 img/s | 3.00x |
|
|
| resnext101-32x4d | 4160 img/s | 1210 img/s | 3.44x |
|
|
| se-resnext101-32x4d | 3360 img/s | 1120 img/s | 3.00x |
|
|
|
|
## Release notes
|
|
|
|
### Changelog
|
|
June 2020
|
|
- ConvNets repo restructurization
|
|
- Initial release of ResNext and SE-Resnext
|