Predicting a tfestimators model after exporting without session restart fails

[javierluraschi]

Run the following, notice that this is fixed if the r session is restarted after running export_savedmodel().

library(tfestimators)

mtcars_input_fn <- function(data, num_epochs = 1) {
  input_fn(data,
           features = c("disp", "cyl"),
           response = "mpg",
           batch_size = 32,
           num_epochs = num_epochs)
}

cols <- feature_columns(column_numeric("disp"), column_numeric("cyl"))

model <- linear_regressor(feature_columns = cols)

indices <- sample(1:nrow(mtcars), size = 0.80 * nrow(mtcars))
train <- mtcars[indices, ]
test  <- mtcars[-indices, ]

model %>% train(mtcars_input_fn(train, num_epochs = 10))

export_savedmodel(model, "savedmodel")

# transform data frame records into list of named lists
test_records <- mtcars[1:5, c("disp", "cyl")]
test_records <- lapply(1:nrow(test_records), function(i) as.list(test_records[i,]))

# generate predictions
library(tfdeploy)
predict_savedmodel(
  test_records,
  "savedmodel",  
  type = "export",
  signature_name = "predict"
)

Error in py_call_impl(callable, dots$args, dots$keywords) : 
  InvalidArgumentError: You must feed a value for placeholder tensor 'Placeholder_2' with dtype float and shape [?,1]
	 [[Node: Placeholder_2 = Placeholder[dtype=DT_FLOAT, shape=[?,1], _device="/job:localhost/replica:0/task:0/device:CPU:0"]()]]

Caused by op u'Placeholder_2', defined at:
  File "/Users/javierluraschi/.virtualenvs/r-tensorflow/lib/python2.7/site-packages/tensorflow/python/saved_model/loader_impl.py", line 216, in load
    saver = tf_saver.import_meta_graph(meta_graph_def_to_load, **saver_kwargs)
  File "/Users/javierluraschi/.virtualenvs/r-tensorflow/lib/python2.7/site-packages/tensorflow/python/training/saver.py", line 1810, in import_meta_graph
    **kwargs)
  File "/Users/javierluraschi/.virtualenvs/r-tensorflow/lib/python2.7/site-packages/tensorflow/python/framework/meta_graph.py", line 660, in import_scoped_meta_graph
    producer_op_list=producer_op_list)
  File "/Users/javierluraschi/.virtualenvs/r-tensorflow/lib/python2.7/site-packages/tensorflow/python/ 

[javierluraschi]

The problem is that under export_savedmodel, we create a placeholder which seems to affect subsequent use of the TF session...

[javierluraschi]

See https://github.com/rstudio/tfestimators/compare/bugfix/export-savedmodel-default?expand=1