How train a googlenet model adversarially on an own image classification dataset?
For example: Using cleverhans library, the data that has batches to run the attacks on are MNIST and CIFAR.
I trained an image classifier with my own data (Googlenet) using Tensorflow, now I want to train the model with the adversarial examples. Any ideas that I can do with the cleverhans library. Thanks.
The easiest is probably to start from your own code to train GoogleNet and modify its loss. You can find an example modification of the loss that adds a penalty to train on adversarial examples in the CleverHans tutorial. It uses the loss implementation found here to define a weighted average between the cross-entropy on clean images and the cross-entropy on adversarial images.
Related
How do you guys visualize the training history of your pytorch model like in keras here.
I have a pytorch trained model and I want to see the graph of its training.
Can I do this using only matplotlib? If yes, can someone give me resources to follow.
You have to save the loss while training. A trained model won't have history of its loss. You need to train again.
Save the loss while training then plot it against the epochs using matplotlib. In your training function, where loss is being calculated save that to a file and visualize it later.
Also, you can use tensorboardX if you want to visualize in realtime.
This is a tutorial for tensorboardX: http://www.erogol.com/use-tensorboard-pytorch/
I have searched this for a while, but it seems Keras only has quantization feature after the model is trained. I wish to add Tensorflow fake quantization to my Keras sequential model. According to Tensorflow's doc, I need these two functions to do fake quantization: tf.contrib.quantize.create_training_graph() and tf.contrib.quantize.create_eval_graph().
My question is has anyone managed to add these two functions in a Keras model? If yes, where should these two function be added? For example, before model.compile or after model.fit or somewhere else? Thanks in advance.
I worked around by post-training quantization. Since my final goal is to train a mdoel for mobile device, instead of fake quantization during training, I exported keras .h5 file and converted to Tenforflow lite .tflite file directly (with post_training_quantize flag set to true). I tested this on a simple cifar-10 model. The original keras model and the quantized tflite model have very close accuracy (the quantized one a bit lower).
Post-training quantization: https://www.tensorflow.org/performance/post_training_quantization
Convert Keras model to tensorflow lite: https://github.com/tensorflow/tensorflow/blob/master/tensorflow/contrib/lite/toco/g3doc/python_api.md
Used the tf-nightly tensorflow here: https://pypi.org/project/tf-nightly/
If you still want to do fake quantization (because for some model, post-training quantization may give poor accuracy according to Google), the original webpage is down last week. But you can find it from github: https://github.com/tensorflow/tensorflow/tree/master/tensorflow/contrib/quantize
Update: Turns out post-quantization does not really quantize the model. During inference, it still uses float32 kernels to do calculations. Thus, I've switched to quantization-aware training. The accuracy is pretty good for my cifar10 model.
I am working on the recently released "SSD-Mobilenet" model by google for object detection.
Model downloaded from following location: https://github.com/tensorflow/models/blob/master/object_detection/g3doc/detection_model_zoo.md
The frozen graph file downloaded from the site is working as expected, however after quantization the accuracy drops significantly (mostly random predictions).
I built tensorflow r1.2 from source, and used following method to quantize:
bazel-bin/tensorflow/tools/graph_transforms/transform_graph --in_graph=frozen_inference_graph.pb --out_graph=optimized_graph.pb --inputs='image_tensor' --outputs='detection_boxes','detection_scores','detection_classes','num_detections' --transforms='add_default_attributes strip_unused_nodes(type=float, shape="1,224,224,3") fold_constants(ignore_errors=true) fold_batch_norms fold_old_batch_norms quantize_weights strip_unused_nodes sort_by_execution_order'
I tried various combinations in the "transforms" part, and the transforms mentioned above gave sometimes correct predictions, however no where close to the original model.
Is there any other way to improve performance of the quantized model?
In this case SSD uses mobilenet as it's feature extractor . In-order to increase the speed. If you read the mobilenet paper , it's a lightweight convolutional neural nets specially using separable convolution inroder to reduce parameters .
As I understood separable convolution can loose information because of the channel wise convolution.
So when quantifying a graph according to TF implementation it makes 16 bits ops and weights to 8bits . If you read the tutorial in TF for quantization they clearly have mentioned how this operation is more like adding some noise in to already trained net hoping our model has well generalized .
So this will work really well and almost lossless interms of accuracy for a heavy model like inception , resnet etc. But with the lightness and simplicity of ssd with mobilenet it really can make a accuracy loss .
MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
How to Quantize Neural Networks with TensorFlow
I want to implements a faster-rcnn model using distributed tensorflow, But I have difficult to load a pretrained vgg model,How to do it? thanks
The TensorFlow tutorial on retraining inception is a good start to read. Then try to reproduce what it does starting from an already trained vgg model.
I am working on a project which is to localize object in a image. The method I am going to adopt is based on the localization algorithm in CS231n-8.
The network structure has two optimization heads, classification head and regression head. How can I minimize both of them when training the network?
I have one idea that summarizing both of them into one loss. But the problem is classification loss is softmax loss and regression loss is l2 loss, which means they have different range. I don't think this is the best way.
It depends on your network status.
If your network is just able to extract features [you're using weights kept from some other net], you can set this weights to be constants and then train separately the two classification heads, since the gradient will not flow trough the constants.
If you're not using weights from a pre-trained model, you
Have to train the network to extract features: thus train the network using the classification head and let the gradient flow from the classification head to the first convolutional filter. In this way your network now can classify objects combining the extracted features.
Convert to constant tensors the learned weights of the convolutional filters and the classification head and train the regression head.
The regression head will learn to combine the features extracted from the convolutional layer adapting its parameters in order to minimize the L2 loss.
Tl;dr:
Train the network for classification first.
Convert every learned parameter to a constant tensor, using graph_util.convert_variables_to_constants as showed in the 'freeze_graph` script.
Train the regression head.