Models: TypeError: x and y must have the same dtype, got tf.float32 != tf.int32

[jimo123]

i am learning the tensorflow object detection api.i can train model normally using the train.py script,but when i use the eval.py script to evaluate it,some wrong happend.
python eval.py --logtostderr \

--checkpoint_dir=train_log \
--eval_dir=eval_log \
--pipeline_config_path=ssd_mobilenet_v1_coco.config
INFO:tensorflow:depth of additional conv before box predictor: 0
INFO:tensorflow:depth of additional conv before box predictor: 0
INFO:tensorflow:depth of additional conv before box predictor: 0
INFO:tensorflow:depth of additional conv before box predictor: 0
INFO:tensorflow:depth of additional conv before box predictor: 0
INFO:tensorflow:depth of additional conv before box predictor: 0
Traceback (most recent call last):
File "eval.py", line 168, in
tf.app.run()
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/platform/app.py", line 48, in run
_sys.exit(main(_sys.argv[:1] + flags_passthrough))
File "eval.py", line 164, in main
FLAGS.checkpoint_dir, FLAGS.eval_dir)
File "/home/yj/tensorflow3/models/research/object_detection/evaluator.py", line 142, in evaluate
ignore_groundtruth=eval_config.ignore_groundtruth)
File "/home/yj/tensorflow3/models/research/object_detection/evaluator.py", line 61, in _extract_prediction_tensors
detections = model.postprocess(prediction_dict)
File "/home/yj/tensorflow3/models/research/object_detection/meta_architectures/ssd_meta_arch.py", line 405, in postprocess
class_predictions_without_background)
File "/home/yj/tensorflow3/models/research/object_detection/builders/post_processing_builder.py", line 94, in score_converter_fn
scaled_logits = tf.divide(logits, logit_scale, name='scale_logits')
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/ops/math_ops.py", line 309, in divide
return DivideDelegateWithName(x, name) / y
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/ops/math_ops.py", line 294, in __truediv__
return _truediv_python3(self.x, y, self.name)
File "/usr/local/lib/python3.4/dist-packages/tensorflow/python/ops/math_ops.py", line 981, in _truediv_python3
(x_dtype, y_dtype))
TypeError: x and y must have the same dtype, got tf.float32 != tf.int32

my tf version is 1.4.0,python3.4,cpu,thanks

[aselle]

Can you print the value of logit_scale ... it seems like it may be integer.

[jimo123]

I have solved it, because I changed python version to 3.5.but I still don't know the reason.thank you all the same.

[marcbelmont]

@aselle I'm also having the issue (with python 3.6).
logit_scale is 1 (type int)

[Kishwar]

@jimo123 you saved my day...

[Kishwar]

@marcbelmont change python to 3.5

[wt-huang]

Closing as this is resolved

[ian-cannon]

Running with Python 3.5 does not work because it causes conflicts with other packages. Is there any other way around this error?

@jimo123 @Kishwar Are you able to send me the conda or virtual env you are working in?

[ian-cannon]

This can be solved by this post: https://stackoverflow.com/questions/56544619/how-to-fix-typeerror-x-and-y-must-have-the-same-dtype-got-tf-uint8-tf-floa