maxmustermann/DeepLearningExamples
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity

Merge pull request #718 from NVIDIA/gh/release

Browse source 
[ConvNets/Pyt] Pretrained weights usage guidelines
This commit is contained in:
nv-kkudrynski 2020-10-21 17:24:37 +02:00 committed by GitHub
parent 0a120b2a1e 9061083098
commit d1b185459b
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
5 changed files with 141 additions and 28 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

9
PyTorch/Classification/ConvNets/classify.py
View file

@ -63,17 +63,16 @@ def main(args):
if args.weights is not None:
weights = torch.load(args.weights)
#Temporary fix to allow NGC checkpoint loading
weights = {k.replace("module.", ""): v for k, v in weights.items()}
weights = {
k.replace("module.", ""): v for k, v in weights.items()
}
model.load_state_dict(weights)
model = model.cuda()
if args.precision in ["AMP", "FP16"]:
model = model.half()
model = network_to_half()
model.eval()

8
PyTorch/Classification/ConvNets/main.py
View file

@ -363,10 +363,10 @@ def main(args):
)
)
pretrained_weights = torch.load(args.pretrained_weights)
#Temporary fix to allow NGC checkpoint loading
pretrained_weights = {k.replace("module.", ""): v for k, v in pretrained_weights.items()}
# Temporary fix to allow NGC checkpoint loading
pretrained_weights = {
k.replace("module.", ""): v for k, v in pretrained_weights.items()
}
else:
print("=> no pretrained weights found at '{}'".format(args.resume))

50
PyTorch/Classification/ConvNets/resnet50v1.5/README.md
View file

@ -281,17 +281,21 @@ Example:
### 6. Start inference
To run inference on ImageNet on a checkpointed model, run:
You can download pretrained weights from NGC:
`python ./main.py --arch resnet50 --evaluate --epochs 1 --resume <path to checkpoint> -b <batch size> <path to imagenet>`
```bash
wget --content-disposition https://api.ngc.nvidia.com/v2/models/nvidia/resnet50_pyt_amp/versions/20.06.0/zip -O resnet50_pyt_amp_20.06.0.zip
To run inference on JPEG image, you have to first extract the model weights from checkpoint:
unzip resnet50_pyt_amp_20.06.0.zip
```
`python checkpoint2model.py --checkpoint-path <path to checkpoint> --weight-path <path where weights will be stored>`
To run inference on ImageNet, run:
Then run classification script:
`python ./main.py --arch resnet50 --evaluate --epochs 1 --pretrained-weights nvidia_resnet50_200821.pth.tar -b <batch size> <path to imagenet>`
`python classify.py --arch resnet50 -c fanin --weights <path to weights from previous step> --precision AMP|FP32 --image <path to JPEG image>`
To run inference on JPEG image using pretrained weights:
`python classify.py --arch resnet50 -c fanin --weights nvidia_resnet50_200821.pth.tar --precision AMP|FP32 --image <path to JPEG image>`
## Advanced
@ -445,6 +449,19 @@ Metrics gathered through training:
- `train.data_time` - time spent on waiting on data
- `train.compute_time` - time spent in forward/backward pass
To restart training from checkpoint use `--resume` option.
To start training from pretrained weights (e.g. downloaded from NGC) use `--pretrained-weights` option.
The difference between those two is that the pretrained weights contain only model weights,
and checkpoints, apart from model weights, contain optimizer state, LR scheduler state, RNG state.
Checkpoints are suitable for dividing the training into parts, for example in order
to divide the training job into shorter stages, or restart training after infrastructure fail.
Pretrained weights can be used as a base for finetuning the model to a different dataset,
or as a backbone to detection models.
### Inference process
Validation is done every epoch, and can be also run separately on a checkpointed model.
@ -470,6 +487,27 @@ Then run classification script:
`python classify.py --arch resnet50 -c fanin --weights <path to weights from previous step> --precision AMP|FP32 --image <path to JPEG image>`
You can also run ImageNet validation on pretrained weights:
`python ./main.py --arch resnet50 --evaluate --epochs 1 --pretrained-weights <path to pretrained weights> -b <batch size> <path to imagenet>`
#### NGC Pretrained weights:
Pretrained weights can be downloaded from NGC:
```bash
wget --content-disposition https://api.ngc.nvidia.com/v2/models/nvidia/resnet50_pyt_amp/versions/20.06.0/zip -O resnet50_pyt_amp_20.06.0.zip
unzip resnet50_pyt_amp_20.06.0.zip
```
To run inference on ImageNet, run:
`python ./main.py --arch resnet50 --evaluate --epochs 1 --pretrained-weights nvidia_resnet50_200821.pth.tar -b <batch size> <path to imagenet>`
To run inference on JPEG image using pretrained weights:
`python classify.py --arch resnet50 -c fanin --weights nvidia_resnet50_200821.pth.tar --precision AMP|FP32 --image <path to JPEG image>`
## Performance

52
PyTorch/Classification/ConvNets/resnext101-32x4d/README.md
View file

@ -266,17 +266,21 @@ Example:
### 6. Start inference
To run inference on ImageNet on a checkpointed model, run:
You can download pretrained weights from NGC:
`python ./main.py --arch resnext101-32x4d --evaluate --epochs 1 --resume <path to checkpoint> -b <batch size> <path to imagenet>`
```bash
wget --content-disposition https://api.ngc.nvidia.com/v2/models/nvidia/resnext101_32x4d_pyt_amp/versions/20.06.0/zip -O resnext101_32x4d_pyt_amp_20.06.0.zip
To run inference on JPEG image, you have to first extract the model weights from checkpoint:
unzip resnext101_32x4d_pyt_amp_20.06.0.zip
```
`python checkpoint2model.py --checkpoint-path <path to checkpoint> --weight-path <path where weights will be stored>`
To run inference on ImageNet, run:
Then run classification script:
`python ./main.py --arch resnext101-32x4d --evaluate --epochs 1 --pretrained-weights nvidia_resnext101-32x4d_200821.pth.tar -b <batch size> <path to imagenet>`
`python classify.py --arch resnext101-32x4d -c fanin --weights <path to weights from previous step> --precision AMP|FP32 --image <path to JPEG image>`
To run inference on JPEG image using pretrained weights:
`python classify.py --arch resnext101-32x4d -c fanin --weights nvidia_resnext101-32x4d_200821.pth.tar --precision AMP|FP32 --image <path to JPEG image>`
## Advanced
@ -431,6 +435,19 @@ Metrics gathered through training:
- `train.data_time` - time spent on waiting on data
- `train.compute_time` - time spent in forward/backward pass
To restart training from checkpoint use `--resume` option.
To start training from pretrained weights (e.g. downloaded from NGC) use `--pretrained-weights` option.
The difference between those two is that the pretrained weights contain only model weights,
and checkpoints, apart from model weights, contain optimizer state, LR scheduler state, RNG state.
Checkpoints are suitable for dividing the training into parts, for example in order
to divide the training job into shorter stages, or restart training after infrastructure fail.
Pretrained weights can be used as a base for finetuning the model to a different dataset,
or as a backbone to detection models.
### Inference process
Validation is done every epoch, and can be also run separately on a checkpointed model.
@ -454,8 +471,29 @@ To run inference on JPEG image, you have to first extract the model weights from
Then run classification script:
`python classify.py --arch resnext101-32x4d -c fanin --weights <path to weights from previous step> --precision AMP|
`python classify.py --arch resnext101-32x4d -c fanin --weights <path to weights from previous step> --precision AMP|FP32 --image <path to JPEG image>`
You can also run ImageNet validation on pretrained weights:
`python ./main.py --arch resnext101-32x4d --evaluate --epochs 1 --pretrained-weights <path to pretrained weights> -b <batch size> <path to imagenet>`
#### NGC Pretrained weights:
Pretrained weights can be downloaded from NGC:
```bash
wget --content-disposition https://api.ngc.nvidia.com/v2/models/nvidia/resnext101-32x4d_pyt_amp/versions/20.06.0/zip -O resnext101-32x4d_pyt_amp_20.06.0.zip
unzip resnext101-32x4d_pyt_amp_20.06.0.zip
```
To run inference on ImageNet, run:
`python ./main.py --arch resnext101-32x4d --evaluate --epochs 1 --pretrained-weights nvidia_resnext101-32x4d_200821.pth.tar -b <batch size> <path to imagenet>`
To run inference on JPEG image using pretrained weights:
`python classify.py --arch resnext101-32x4d -c fanin --weights nvidia_resnext101-32x4d_200821.pth.tar --precision AMP|FP32 --image <path to JPEG image>`
## Performance

50
PyTorch/Classification/ConvNets/se-resnext101-32x4d/README.md
View file

@ -267,17 +267,21 @@ Example:
### 6. Start inference
To run inference on ImageNet on a checkpointed model, run:
You can download pretrained weights from NGC:
`python ./main.py --arch se-resnext101-32x4d --evaluate --epochs 1 --resume <path to checkpoint> -b <batch size> <path to imagenet>`
```bash
wget --content-disposition https://api.ngc.nvidia.com/v2/models/nvidia/seresnext101_32x4d_pyt_amp/versions/20.06.0/zip -O seresnext101_32x4d_pyt_amp_20.06.0.zip
To run inference on JPEG image, you have to first extract the model weights from checkpoint:
unzip seresnext101_32x4d_pyt_amp_20.06.0.zip
```
`python checkpoint2model.py --checkpoint-path <path to checkpoint> --weight-path <path where weights will be stored>`
To run inference on ImageNet, run:
Then run classification script:
`python ./main.py --arch se-resnext101-32x4d --evaluate --epochs 1 --pretrained-weights nvidia_se-resnext101-32x4d_200821.pth.tar -b <batch size> <path to imagenet>`
`python classify.py --arch se-resnext101-32x4d -c fanin --weights <path to weights from previous step> --precision AMP|FP32 --image <path to JPEG image>`
To run inference on JPEG image using pretrained weights:
`python classify.py --arch se-resnext101-32x4d -c fanin --weights nvidia_se-resnext101-32x4d_200821.pth.tar --precision AMP|FP32 --image <path to JPEG image>`
## Advanced
@ -432,6 +436,19 @@ Metrics gathered through training:
- `train.data_time` - time spent on waiting on data
- `train.compute_time` - time spent in forward/backward pass
To restart training from checkpoint use `--resume` option.
To start training from pretrained weights (e.g. downloaded from NGC) use `--pretrained-weights` option.
The difference between those two is that the pretrained weights contain only model weights,
and checkpoints, apart from model weights, contain optimizer state, LR scheduler state, RNG state.
Checkpoints are suitable for dividing the training into parts, for example in order
to divide the training job into shorter stages, or restart training after infrastructure fail.
Pretrained weights can be used as a base for finetuning the model to a different dataset,
or as a backbone to detection models.
### Inference process
Validation is done every epoch, and can be also run separately on a checkpointed model.
@ -457,6 +474,27 @@ Then run classification script:
`python classify.py --arch se-resnext101-32x4d -c fanin --weights <path to weights from previous step> --precision AMP|FP32 --image <path to JPEG image>`
You can also run ImageNet validation on pretrained weights:
`python ./main.py --arch se-resnext101-32x4d --evaluate --epochs 1 --pretrained-weights <path to pretrained weights> -b <batch size> <path to imagenet>`
#### NGC Pretrained weights:
Pretrained weights can be downloaded from NGC:
```bash
wget --content-disposition https://api.ngc.nvidia.com/v2/models/nvidia/seresnext101_32x4d_pyt_amp/versions/20.06.0/zip -O seresnext101_32x4d_pyt_amp_20.06.0.zip
unzip seresnext101_32x4d_pyt_amp_20.06.0.zip
```
To run inference on ImageNet, run:
`python ./main.py --arch se-resnext101-32x4d --evaluate --epochs 1 --pretrained-weights nvidia_se-resnext101-32x4d_200821.pth.tar -b <batch size> <path to imagenet>`
To run inference on JPEG image using pretrained weights:
`python classify.py --arch se-resnext101-32x4d -c fanin --weights nvidia_se-resnext101-32x4d_200821.pth.tar --precision AMP|FP32 --image <path to JPEG image>`
## Performance