INT8 Quantization: how to generate and read calibration table?

AI & Data Science Deep Learning (Training & Inference) TensorRT
1

Hello everyone,

I am using TensorRT in order to quantize a DNN for object detection called “Pixor”.
I am using Python3 + Tensorflow 1.12 + TensorRT 3.0.2

The quantization work fine for me. However, I want to generate and read the calibration table in order to understand if my calibration dataset is good enough or not.

I want to ask also if i can generate the histograms of activation as shown in these slides?

Thank you in advance,
Fares

2

Unfortunately, we have not made public the details of the calibration table. We are always listening to community feedback and please stay tuned for future announcements.

3

Hello. Thanks for your answer,

However, I found the following python code in NVIDIA tutorial to extract the TensorRT calibration table after the calibration is done:

for n in trt_graph.node:
if n.op == “TRTEngineOp”:
print(“Node: %s, %s” % (n.op, n.name.replace(“/”, ““)))
with tf.gfile.GFile(”%s.calib_table" % (n.name.replace(“/”, "”)), ‘wb’) as f:
f.write(n.attr[“calibration_data”].s)

This code generates a file for each node in the trt_graph. However, I can’t read it!
I get something like this:

Up_sample_6/conv2d_25/Relu: 3f556f06
InputPH_1: 3e0324b6
Up_sample_6/conv2d_26/Relu: 3eee7366
OutputPH_0: 3e59daba
Up_sample_6/conv2d_27/Relu: 3dc77c91
InputPH_0: 3f41b751

Can you please explain how to read these files?