Open eilifsolberg opened 3 years ago
Please answer the following questions for yourself before submitting an issue.
https://github.com/tensorflow/models/tree/master/research/...
The 'freeze_variables' parameter in https://github.com/tensorflow/models/blob/master/research/object_detection/protos/train.proto
does not seem to have any effect when using TensorFlow 2. When setting 'freeze_variables' to '.*', which should freeze all variables, the model variables still show up in detection_model.trainable_variables when performing gradient calculation in https://github.com/tensorflow/models/blob/master/research/object_detection/model_lib_v2.py
detection_model.trainable_variables
It seems like 'freeze_variables' is only used in the old tf1 code https://github.com/tensorflow/models/blob/master/research/object_detection/model_lib.py
I'm looking for something similar to reduce my train time. Any update on this ?
Any update?
Prerequisites
Please answer the following questions for yourself before submitting an issue.
1. The entire URL of the file you are using
https://github.com/tensorflow/models/tree/master/research/...
2. Describe the bug
The 'freeze_variables' parameter in https://github.com/tensorflow/models/blob/master/research/object_detection/protos/train.proto
does not seem to have any effect when using TensorFlow 2. When setting 'freeze_variables' to '.*', which should freeze all variables, the model variables still show up in
detection_model.trainable_variables
when performing gradient calculation in https://github.com/tensorflow/models/blob/master/research/object_detection/model_lib_v2.pyIt seems like 'freeze_variables' is only used in the old tf1 code https://github.com/tensorflow/models/blob/master/research/object_detection/model_lib.py