checkpoint file causing errors.

[kharanshu2]

Checkpoint file is causing errors while running inference in command line(as given in the inference example): (I have tried it on 2L and 1L models and same error occurs):

1. When I keep model path and not checkpoint path in ckpt_name the inference command:
   (base) D:\Repos\covid\COVIDNet-CT>python run_covidnet_ct.py infer --model_dir D:\Repos\covid\COVID-Net_CT-1_S --meta_name D:\Repos\covid\COVID-Net_CT-1_S\model.meta --ckpt_name D:\Repos\covid\COVID-Net_CT-1_S --image_file D:\Repos\covid\data\covid1.png
   WARNING:tensorflow:From run_covidnet_ct.py:361: The name tf.logging.set_verbosity is deprecated. Please use tf.compat.v1.logging.set_verbosity instead.

WARNING:tensorflow:From run_covidnet_ct.py:361: The name tf.logging.ERROR is deprecated. Please use tf.compat.v1.logging.ERROR instead.

2023-03-16 10:19:18.597935: I tensorflow/core/platform/cpu_feature_guard.cc:142] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
Loading meta graph from D:\Repos\covid\COVID-Net_CT-1_S\model.meta
Loading weights from D:\Repos\covid\COVID-Net_CT-1_S
2023-03-16 10:19:19.828731: W tensorflow/core/util/tensor_slice_reader.cc:95] Could not open D:\Repos\covid\COVID-Net_CT-1_S: Unknown: NewRandomAccessFile failed to Create/Open: D:\Repos\covid\COVID-Net_CT-1_S : Access is denied.
; Input/output error
2023-03-16 10:19:19.830576: W tensorflow/core/util/tensor_slice_reader.cc:95] Could not open D:\Repos\covid\COVID-Net_CT-1_S: Unknown: NewRandomAccessFile failed to Create/Open: D:\Repos\covid\COVID-Net_CT-1_S : Access is denied.
; Input/output error
2023-03-16 10:19:19.830663: W tensorflow/core/framework/op_kernel.cc:1651] OP_REQUIRES failed at save_restore_tensor.cc:175 : Data loss: Unable to open table file D:\Repos\covid\COVID-Net_CT-1_S: Unknown: NewRandomAccessFile failed to Create/Open: D:\Repos\covid\COVID-Net_CT-1_S : Access is denied.
; Input/output error
Traceback (most recent call last):
File "C:\Users\KharanshuNaghera\miniconda3\lib\site-packages\tensorflow_core\python\client\session.py", line 1365, in _do_call
return fn(*args)
File "C:\Users\KharanshuNaghera\miniconda3\lib\site-packages\tensorflow_core\python\client\session.py", line 1350, in _run_fn
target_list, run_metadata)
File "C:\Users\KharanshuNaghera\miniconda3\lib\site-packages\tensorflow_core\python\client\session.py", line 1443, in _call_tf_sessionrun
run_metadata)
tensorflow.python.framework.errors_impl.DataLossError: Unable to open table file D:\Repos\covid\COVID-Net_CT-1_S: Unknown: NewRandomAccessFile failed to Create/Open: D:\Repos\covid\COVID-Net_CT-1_S : Access is denied.
; Input/output error
[[{{node save/RestoreV2}}]]

2. When I keep checkpoint file in ckpt_name:
   (base) D:\Repos\covid\COVIDNet-CT>python run_covidnet_ct.py infer --model_dir D:\Repos\covid\COVID-Net_CT-1_S --meta_name D:\Repos\covid\COVID-Net_CT-1_S\model.meta --ckpt_name D:\Repos\covid\COVID-Net_CT-1_S\checkpoint --image_file D:\Repos\covid\data\covid1.png
   WARNING:tensorflow:From run_covidnet_ct.py:361: The name tf.logging.set_verbosity is deprecated. Please use tf.compat.v1.logging.set_verbosity instead.

WARNING:tensorflow:From run_covidnet_ct.py:361: The name tf.logging.ERROR is deprecated. Please use tf.compat.v1.logging.ERROR instead.

2023-03-16 10:22:38.860381: I tensorflow/core/platform/cpu_feature_guard.cc:142] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
Loading meta graph from D:\Repos\covid\COVID-Net_CT-1_S\model.meta
Loading weights from D:\Repos\covid\COVID-Net_CT-1_S\checkpoint
2023-03-16 10:22:40.122130: W tensorflow/core/util/tensor_slice_reader.cc:95] Could not open D:\Repos\covid\COVID-Net_CT-1_S\checkpoint: Data loss: not an sstable (bad magic number): perhaps your file is in a different file format and you need to use a different restore operator?
2023-03-16 10:22:40.123169: W tensorflow/core/util/tensor_slice_reader.cc:95] Could not open D:\Repos\covid\COVID-Net_CT-1_S\checkpoint: Data loss: not an sstable (bad magic number): perhaps your file is in a different file format and you need to use a different restore operator?
2023-03-16 10:22:40.123238: W tensorflow/core/framework/op_kernel.cc:1651] OP_REQUIRES failed at save_restore_tensor.cc:175 : Data loss: Unable to open table file D:\Repos\covid\COVID-Net_CT-1_S\checkpoint: Data loss: not an sstable (bad magic number): perhaps your file is in a different file format and you need to use a different restore operator?
Traceback (most recent call last):
File "C:\Users\KharanshuNaghera\miniconda3\lib\site-packages\tensorflow_core\python\client\session.py", line 1365, in _do_call
return fn(*args)
File "C:\Users\KharanshuNaghera\miniconda3\lib\site-packages\tensorflow_core\python\client\session.py", line 1350, in _run_fn
target_list, run_metadata)
File "C:\Users\KharanshuNaghera\miniconda3\lib\site-packages\tensorflow_core\python\client\session.py", line 1443, in _call_tf_sessionrun
run_metadata)
tensorflow.python.framework.errors_impl.DataLossError: Unable to open table file D:\Repos\covid\COVID-Net_CT-1_S\checkpoint: Data loss: not an sstable (bad magic number): perhaps your file is in a different file format and you need to use a different restore operator?
[[{{node save/RestoreV2}}]]

--Model does work when no path is to ckpt file is provided.

Further more while running the ipynb file provided for visualization during execution of this block:

Create full paths

meta_file = os.path.join(MODEL_DIR, META_NAME)
ckpt = os.path.join(MODEL_DIR, CKPT_NAME)

Load metagraph and create session

graph, sess, saver = load_graph(meta_file)

Load checkpoint

with graph.as_default():
load_ckpt(ckpt, sess, saver)
final_conv, pooled_grads = make_gradcam_graph(graph)

this error occurs:

Loading meta graph from D:/Repos/covid/COVID-Net_CT-1_S/model.meta
Loading weights from D:/Repos/covid/COVID-Net_CT-1_S/checkpoint
INFO:tensorflow:Restoring parameters from D:/Repos/covid/COVID-Net_CT-1_S/checkpoint

---

DataLossError Traceback (most recent call last)
~\miniconda3\envs\covidnet\lib\site-packages\tensorflow_core\python\client\session.py in _do_call(self, fn, *args)
1364 try:
-> 1365 return fn(*args)
1366 except errors.OpError as e:

~\miniconda3\envs\covidnet\lib\site-packages\tensorflow_core\python\client\session.py in _run_fn(feed_dict, fetch_list, target_list, options, run_metadata)
1349 return self._call_tf_sessionrun(options, feed_dict, fetch_list,
-> 1350 target_list, run_metadata)
1351

~\miniconda3\envs\covidnet\lib\site-packages\tensorflow_core\python\client\session.py in _call_tf_sessionrun(self, options, feed_dict, fetch_list, target_list, run_metadata)
1442 fetch_list, target_list,
-> 1443 run_metadata)
1444

DataLossError: Unable to open table file D:\Repos\covid\COVID-Net_CT-1_S\checkpoint: Data loss: not an sstable (bad magic number): perhaps your file is in a different file format and you need to use a different restore operator?
[[{{node save/RestoreV2}}]]

[kharanshu2]

tensorflow/models#2676 (comment)

this link helped really, apparently you are supposed to load model.data-00000-of-00001 and not checkpoint and use just the name "model" instead of full "model.data-00000-of-00001"