Add tensorflow known issue for inter op parallelism cannot be modified (#5923)
* Add tensorflow known issue for inter op parallelism cannot be modified * try clear session * remove clear session
This commit is contained in:
parent
b9c73c4c20
commit
0d696fcf92
1 changed files with 4 additions and 0 deletions
|
|
@ -52,6 +52,10 @@ spec:
|
|||
|
||||
Nano keras multi-instance training will serialize TensorFlow dataset object into a `graph.pb` file, which does not work with `dataset.from_generators`, `dataset.numpy_function`, `dataset.py_function` due to limitations in TensorFlow.
|
||||
|
||||
### **RuntimeError: Inter op parallelism cannot be modified after initialization**
|
||||
|
||||
If you meet this error when import `bigdl.nano.tf`, it could be that you have already run some TensorFlow code that set the inter/intra op parallelism, such as `tfds.load`. You can try to workaround this issue by trying to import `bigdl.nano.tf` first before running TensorFlow code. See https://github.com/tensorflow/tensorflow/issues/57812 for more information.
|
||||
|
||||
## **Ray Issues**
|
||||
|
||||
### **protobuf version error**
|
||||
|
|
|
|||
Loading…
Reference in a new issue