I am trying to speed up the segmentation model(unet-mobilenet-512x512). I converted my tensorflow model to tensorRT with FP16 precision mode. And the speed is lower than I expected.
Before the optimization i had 7FPS on inference with .pb frozen graph. After tensorRT oprimization I have 14FPS.
Here is benchmark results of Jetson NX from their site
You can see, that unet 256x256 segmentation speed is 146 FPS. I thought, the speed of my unet512x512 should be 4 times slower in the worst case.
Here is my code for optimizing tensorflow saved model using TensorRt:
import numpy as np
from tensorflow.python.compiler.tensorrt import trt_convert as trt
import tensorflow as tf
params = trt.DEFAULT_TRT_CONVERSION_PARAMS
params = params._replace(
max_workspace_size_bytes=(1<<32))
params = params._replace(precision_mode="FP16")
converter = tf.experimental.tensorrt.Converter(input_saved_model_dir='./model1', conversion_params=params)
converter.convert()
def my_input_fn():
inp1 = np.random.normal(size=(1, 512, 512, 3)).astype(np.float32)
yield [inp1]
converter.build(input_fn=my_input_fn) # Generate corresponding TRT engines
output_saved_model_dir = "trt_graph2"
converter.save(output_saved_model_dir) # Generated engines will be saved.
print("------------------------freezing the graph---------------------")
from tensorflow.python.framework.convert_to_constants import convert_variables_to_constants_v2
saved_model_loaded = tf.saved_model.load(
output_saved_model_dir, tags=[tf.compat.v1.saved_model.SERVING])
graph_func = saved_model_loaded.signatures[
tf.compat.v1.saved_model.signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY]
frozen_func = convert_variables_to_constants_v2(
graph_func)
frozen_func.graph.as_graph_def()
tf.io.write_graph(graph_or_graph_def=frozen_func.graph,
logdir="./",
name="unet_frozen_graphTensorRt.pb",
as_text=False)
I downloaded the repository, that was used for Jetson NX benchmarking ( https://github.com/NVIDIA-AI-IOT/jetson_benchmarks ), and the speed of unet256x256 really is ~146FPS. But there is no pipeline to optimize the model.
How can I get the similar results? I am looking for the solutions to get speed of my model(unet-mobilenet-512x512) close to 30FPSMaybe I should run inference in other way(without tensorflow) or change some converting parameters?
Any suggestions, thanks
As far as I can see, the repository you linked to uses command line tools that use TensorRT (TRT) under the hood. Note that TensorRT is not the same as "TensorRT in TensorFlow" aka TensorFlow-TensorRT (TF-TRT) which is what you are using in your code. Both TF-TRT and TRT models run faster than regular TF models on a Jetson device but TF-TRT models still tend to be slower than TRT ones (source 1, source 2).
The downside of TRT is that the conversion to TRT needs to be done on the target device and that it can be quite difficult to implement it successfully as there are various TensorFlow operations that TRT does not support (in which case you need to write a custom plugin or pray to God that someone on the internet has already done so. …or use TensorRT only for part of your model and do pre-/postprocessing in TensorFlow).
There are basically two ways to convert models from TensorFlow models to TensorRT "engines" aka "plan files", both of which use intermediate formats:
TF -> UFF -> TRT
TF -> ONNX -> TRT
In both cases, the graphsurgeon/onnx-graphsurgeon libraries can be used to modify the TF/ONNX graph to achieve compatibility of graph operations. Unsupported operations can be added by means of TensorRT plugins, as mentioned above. (This is really the main challenge here: Different graph file formats and different target GPUs support different graph operations.)
There's also a third way where you do TF -> Caffe -> TRT and apparently a fourth one where you use Nvidia's Transfer Learning Toolkit (TLT) (based upon TF/Keras) and a tool called tlt-converter but I'm not familiar with it. The latter link does mention converting a UNet model, though.
Note that the paths involving UFF and Caffe are now deprecated and support will be removed in TensorRT 9.0, so if you want something future-proof, you should probably go for ONNX. That being said, most sample code online I've come across online still uses UFF and TensorRT 9.0 is still some time away.
Anyway, I haven't tried converting a UNet to TensorRT yet, but the following repositories provide sample code which might give you an idea of how it works in principle:
TF -> UFF -> TRT: jkjung-avt/tensorrt_demos, NVIDIA-AI-IOT/tf_to_trt_image_classification (the latter using a bit of C++)
TF -> ONNX -> TRT: tensorflow-onnx, onnx-tensorrt
Keras -> ONNX -> TRT: Nvidia blog post (This one mentions converting a Unet to TRT!)
Note that even if you don't manage to pull off the conversion from ONNX to TRT for your model, using the ONNX runtime for inference could potentially still give you a performance gain, especially when you're using the CUDA or the TensorRT execution provider which will be enabled automatically provided you're on a Jetson device and running the correct ONNXRuntime build. (I'm not sure how it compares to TF-TRT or TRT, though, but it might still be worth a shot.)
Finally, for completeness's sake let me also mention that at least my team has been dabbling with the idea of switching from TF to PyTorch, partly because the Nvidia support has been getting a lot better lately and Nvidia employees seem to gravitate towards PyTorch, too. In particular, there are now two separate ways to convert models to TRT:
PyTorch -> ONNX -> TRT (used by dusty_nv)
PyTorch -> TRT (direct conversion via torch2trt). It seems that quite a few Nvidia repositories use this.
Hi can you share the errors you are getting? Its should work with the following steps:
Convert the TensorFlow/Keras model to a .pb file.
Convert the .pb file to ONNX format.
Create a TensorRT engine.
Run inference from the TensorRT engine.
I am not sure about Unet (I will check) but you may have some operations not supported by onnx (please share your errors).
Here is an example with Resnet-50.
Conversion to .pb:
import tensorflow as tf
import keras
from tensorflow.keras.models import Model
import keras.backend as K
K.set_learning_phase(0)
def keras_to_pb(model, output_filename, output_node_names):
"""
This is the function to convert the Keras model to pb.
Args:
model: The Keras model.
output_filename: The output .pb file name.
output_node_names: The output nodes of the network. If None, then
the function gets the last layer name as the output node.
"""
# Get the names of the input and output nodes.
in_name = model.layers[0].get_output_at(0).name.split(':')[0]
if output_node_names is None:
output_node_names = [model.layers[-1].get_output_at(0).name.split(':')[0]]
sess = keras.backend.get_session()
# The TensorFlow freeze_graph expects a comma-separated string of output node names.
output_node_names_tf = ','.join(output_node_names)
frozen_graph_def = tf.graph_util.convert_variables_to_constants(
sess,
sess.graph_def,
output_node_names)
sess.close()
wkdir = ''
tf.train.write_graph(frozen_graph_def, wkdir, output_filename, as_text=False)
return in_name, output_node_names
# load the ResNet-50 model pretrained on imagenet
model = keras.applications.resnet.ResNet50(include_top=True, weights='imagenet', input_tensor=None, input_shape=None, pooling=None, classes=1000)
# Convert the Keras ResNet-50 model to a .pb file
in_tensor_name, out_tensor_names = keras_to_pb(model, "models/resnet50.pb", None)
Then you need to convert the .pb model to the ONNX format. To do this, you will need to install tf2onnx.
Example:
python -m tf2onnx.convert --input /Path/to/resnet50.pb --inputs input_1:0 --outputs probs/Softmax:0 --output resnet50.onnx
Last step create the TensorRT engine from the ONNX file:
import tensorrt as trt
TRT_LOGGER = trt.Logger(trt.Logger.WARNING)
trt_runtime = trt.Runtime(TRT_LOGGER)
def build_engine(onnx_path, shape = [1,224,224,3]):
"""
This is the function to create the TensorRT engine
Args:
onnx_path : Path to onnx_file.
shape : Shape of the input of the ONNX file.
"""
with trt.Builder(TRT_LOGGER) as builder, builder.create_network(1) as network, trt.OnnxParser(network, TRT_LOGGER) as parser:
builder.max_workspace_size = (256 << 20)
with open(onnx_path, 'rb') as model:
parser.parse(model.read())
network.get_input(0).shape = shape
engine = builder.build_cuda_engine(network)
return engine
def save_engine(engine, file_name):
buf = engine.serialize()
with open(file_name, 'wb') as f:
f.write(buf)
def load_engine(trt_runtime, plan_path):
with open(engine_path, 'rb') as f:
engine_data = f.read()
engine = trt_runtime.deserialize_cuda_engine(engine_data)
return engine
I suggest you check this Pytorch TRT Unet implementation
I am converting a frozen pb model to onnx, but onnx does not support uint8. How do I go about replacing uint8 with int8 in the model file? I can also retrain if I have to, but I am not sure which file to modify and how. I would have guessed that I needed to modify the file which contains the neural net architecture:
https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet/mobilenet_v2.py
but that does not seem to be the case...
There's a MobileNet v2 ONNX Model in the ONNX model zoo that has already been converted (although the source model was originally from mxnet). This might be useful.
ONNX operators do support uint8 datatypes, as described in the ONNX IR documentation.
I trained an ANN model as saved it as .h5 file.Then I converted the model into tensorflow model and got 'savedmodel.pb' and 'variables' folder.
Then I used model optimiser openvino to generate IR files using:
python3 mo_tf.py --input_model saved_model.pb
But I get the following error:
[ FRAMEWORK ERROR ] Error parsing message
TensorFlow cannot read the model file: "/home/user/Downloads/OpenVino/dldt-2019/model-optimizer/saved_model.pb" is incorrect TensorFlow model file
Can openVINO be used to convert ANN models in the first place?
Unfortunately, OpenVINO does not support Keras model format.
I guess you got an error from model optimizer because your model is not frozen.
There are a lot of scripts which can be used to convert a Keras model to a frozen Tensorflow graph. I can recommend this one for example.
Hope it will help.
i tried to follow this tutorial on how to convert a Keras H5 Model zu ProtoBuff and serving it using Tensorflow Serve:
https://towardsdatascience.com/deploying-keras-models-using-tensorflow-serving-and-flask-508ba00f1037
That tutorial among many other resources on the web use "tf.saved_model.simple_save", which is deprecated and removed by now (March 2019).
Converting the h5 into pb using freeze_session as shown here:
How to export Keras .h5 to tensorflow .pb?
Seems to miss a "serve" Tag, as the tensorflow_model_server outputs:
Loading servable: {name: ImageClassifier version: 1} failed: Not found: Could not find meta graph def matching supplied tags: { serve }. To inspect available tag-sets in the SavedModel, please use the SavedModel CLI: saved_model_cli
checked it with saved_model_cli, there are no tags.
What is the way to make a h5 model serveable in tensorflow_server nowadays?
NOTE: This applies to TF 2.0+
I'm assuming you have your Keras model in model.h5.
Firstly, just load the model with tensorflow's implementation of Keras:
from tensorflow import keras
model = keras.models.load_model('model.h5')
Then, simply export a SavedModel
keras.experimental.export_saved_model(model, 'path_to_saved_model')
Finally, apply any transformation you nomally d to go from SavedModel to the .pb inference file (e.g.: freezing, optimizing for inference, etc)
You can hve more details and a full example in TF's official guide for saving and serializing models in TF 2.0
I have a Problem converting Keras Models into Layers API format models to use with tensorflowjs
I use the command:
$ tensorflowjs_converter --input_format keras kerasModels/vgg16_weights_tf_dim_ordering_tf_kernels.h5 convertedModels/
I get an error "KeyError: Can't open attribute (can't locate attribute 'keras version')"
Here is an image of the error log:
I assume you are trying to convert the model downloaded from here, which is possibly outdated now.
You can download the VGG16 model fresh from keras-applications using the following python script:
from keras.applications.vgg16 import VGG16
model = VGG16(include_top=True, weights='imagenet')
model.save("VGG16.h5")