search

google-research / deeplab2

DeepLab2 is a TensorFlow library for deep labeling, aiming to provide a unified and state-of-the-art TensorFlow codebase for dense pixel labeling tasks.
Apache License 2.0
1.01k stars 159 forks source link

AttributeError: module 'tensorflow._api.v2.lookup.experimental' has no attribute 'MutableHashTable' #91

Closed AndreiBaraian closed 2 years ago

AndreiBaraian commented 2 years ago

First of all, thanks a lot for the repo!

I was trying to fine-tune a semantic segmentation model on my own custom dataset, and I came across this error:

python trainer/train.py --mode=train --config_file=configs/cityscapes/panoptic_deeplab/resnet50_os32_semseg.textproto --model_dir=../model_zoo/resnet50_os32_panoptic_deeplab_cityscapes_crowd_trainfine_saved_model --num_gpus=0 2022-03-20 04:45:21.366821: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library libcudart.so.11.0 Traceback (most recent call last): File "trainer/train.py", line 26, in from deeplab2.trainer import train_lib File "/home/abandrei/work/deeplab2_workspace/deeplab2/trainer/train_lib.py", line 35, in from deeplab2.video import vip_deeplab File "/home/abandrei/work/deeplab2_workspace/deeplab2/video/vip_deeplab.py", line 30, in from deeplab2.model.post_processor import vip_deeplab File "/home/abandrei/work/deeplab2_workspace/deeplab2/model/post_processor/vip_deeplab.py", line 110, in class VideoPanopticPredictionStitcher(tf.keras.Model): File "/home/abandrei/work/deeplab2_workspace/deeplab2/model/post_processor/vip_deeplab.py", line 143, in VideoPanopticPredictionStitcher self, id_array: tf.Tensor) -> tf.lookup.experimental.MutableHashTable: AttributeError: module 'tensorflow._api.v2.lookup.experimental' has no attribute 'MutableHashTable'

I am using Tensorflow 2.5.0 Any ideas?

Nikolai10 commented 2 years ago

Hello @AndreiBaraian, it seems like MutableHashTable is only supported starting from tensorflow==2.6.0.

So you might wanna consider updating to tf 2.6+.