Compressed pruned model is the same size as compressed baseline model #47838
Labels
ModelOptimizationToolkit
TF Model Optimization Toolkit
stat:awaiting tensorflower
Status - Awaiting response from tensorflower
TF 2.3
Issues related to TF 2.3
type:bug
Bug
System information
Have I written custom code (as opposed to using a stock example script
provided in TensorFlow):
Code from pruning documentation:
https://www.tensorflow.org/model_optimization/guide/pruning/comprehensive_guide
OS Platform and Distribution (e.g., Linux Ubuntu 16.04):
Windows 10
TensorFlow installed from (source or binary):
Anaconda Navigator
TensorFlow version (use command below):
tensorflow-gpu 2.3.0
Python version:
3.7.9
CUDA/cuDNN version:
cudatoolkit 10.1.243
cudnn 7.6.5
GPU model and memory:
GTX 1070 8GB
Describe the problem
I followed the code provided in the documentation and pruned a model. I expected the compressed pruned model to be of smaller size than the baseline. Here, the baseline is compared to the pruned model, and is definitely smaller.
In my case, they are both the same size.
Source code / logs
Output:
Size of gzipped baseline model: 2935.00 bytes
Size of gzipped pruned model without stripping: 3360.00 bytes
Size of gzipped pruned model with stripping: 2935.00 bytes
The text was updated successfully, but these errors were encountered: