Merge commit for internal changes

3 files changed, 30 insertions, 44 deletions

diff --git a/tensorflow/contrib/lite/python/convert.py b/tensorflow/contrib/lite/python/convert.py
index c038c88945..0ea2630f71 100644
--- a/tensorflow/contrib/lite/python/convert.py
+++ b/tensorflow/contrib/lite/python/convert.py
@@ -25,7 +25,6 @@ import tempfile as _tempfile
 from tensorflow.contrib.lite.python import lite_constants
 from tensorflow.contrib.lite.toco import model_flags_pb2 as _model_flags_pb2
 from tensorflow.contrib.lite.toco import toco_flags_pb2 as _toco_flags_pb2
-from tensorflow.python.framework import dtypes as _dtypes
 from tensorflow.python.platform import resource_loader as _resource_loader
 from tensorflow.python.util.lazy_loader import LazyLoader
 
@@ -135,11 +134,11 @@ def build_toco_convert_protos(input_tensors,
     input_tensors: List of input tensors. Type and shape are computed using
       `foo.get_shape()` and `foo.dtype`.
     output_tensors: List of output tensors (only .name is used from this).
-    inference_type: Target data type of arrays in the output file. Currently
-      must be `{FLOAT, QUANTIZED_UINT8}`.  (default FLOAT)
-    inference_input_type: Target data type of input arrays. Allows for a
-      different type for input arrays in the case of quantization. Currently
-      must be `{FLOAT, QUANTIZED_UINT8}`. (default `inference_type`)
+    inference_type: Target data type of real-number arrays in the output file.
+      Must be `{FLOAT, QUANTIZED_UINT8}`.  (default FLOAT)
+    inference_input_type: Target data type of real-number input arrays. Allows
+      for a different type for input arrays in the case of quantization.
+      Must be `{FLOAT, QUANTIZED_UINT8}`. (default `inference_type`)
     input_format: Type of data to read Currently must be
       `{TENSORFLOW_GRAPHDEF}`. (default TENSORFLOW_GRAPHDEF)
     output_format: Output file format. Currently must be `{TFLITE,
@@ -202,29 +201,13 @@ def build_toco_convert_protos(input_tensors,
   if dump_graphviz_dir:
     toco.dump_graphviz_dir = dump_graphviz_dir
   toco.dump_graphviz_include_video = dump_graphviz_video
+
   model = _model_flags_pb2.ModelFlags()
   model.change_concat_input_ranges = change_concat_input_ranges
   for idx, input_tensor in enumerate(input_tensors):
-    if input_tensor.dtype == _dtypes.float32:
-      tflite_input_type = lite_constants.FLOAT
-    elif input_tensor.dtype == _dtypes.int32:
-      tflite_input_type = lite_constants.INT32
-    elif input_tensor.dtype == _dtypes.int64:
-      tflite_input_type = lite_constants.INT64
-    elif input_tensor.dtype == _dtypes.uint8:
-      tflite_input_type = lite_constants.QUANTIZED_UINT8
-    # TODO(aselle): Insert strings when they are available
-    else:
-      raise ValueError("Tensors %s not known type %r" % (input_tensor.name,
-                                                         input_tensor.dtype))
-
     input_array = model.input_arrays.add()
-
     if inference_type == lite_constants.QUANTIZED_UINT8:
-      if tflite_input_type == lite_constants.FLOAT:
-        tflite_input_type = lite_constants.QUANTIZED_UINT8
       input_array.mean_value, input_array.std_value = quantized_input_stats[idx]
-
     input_array.name = tensor_name(input_tensor)
     input_array.shape.dims.extend(map(int, input_tensor.get_shape()))
 
diff --git a/tensorflow/contrib/lite/python/lite.py b/tensorflow/contrib/lite/python/lite.py
index 88dda7290b..69a2f638af 100644
--- a/tensorflow/contrib/lite/python/lite.py
+++ b/tensorflow/contrib/lite/python/lite.py
@@ -67,11 +67,11 @@ class TocoConverter(object):
 
   Attributes:
 
-    inference_type: Target data type of arrays in the output file. Currently
-      must be `{FLOAT, QUANTIZED_UINT8}`.  (default FLOAT)
-    inference_input_type: Target data type of input arrays. Allows for a
-      different type for input arrays in the case of quantization. Currently
-      must be `{FLOAT, QUANTIZED_UINT8}`. (default `inference_type`)
+    inference_type: Target data type of real-number arrays in the output file.
+      Must be `{FLOAT, QUANTIZED_UINT8}`.  (default FLOAT)
+    inference_input_type: Target data type of real-number input arrays. Allows
+      for a different type for input arrays in the case of quantization.
+      Must be `{FLOAT, QUANTIZED_UINT8}`. (default `inference_type`)
     output_format: Output file format. Currently must be `{TFLITE,
       GRAPHVIZ_DOT}`. (default TFLITE)
     quantized_input_stats: Dict of strings representing input tensor names
diff --git a/tensorflow/contrib/lite/python/tflite_convert.py b/tensorflow/contrib/lite/python/tflite_convert.py
index f497533bed..d18a29834b 100644
--- a/tensorflow/contrib/lite/python/tflite_convert.py
+++ b/tensorflow/contrib/lite/python/tflite_convert.py
@@ -23,19 +23,15 @@ import os
 import sys
 
 from tensorflow.contrib.lite.python import lite
+from tensorflow.contrib.lite.python import lite_constants
 from tensorflow.contrib.lite.toco import toco_flags_pb2 as _toco_flags_pb2
 from tensorflow.contrib.lite.toco import types_pb2 as _types_pb2
 from tensorflow.python.platform import app
 
 
-def _parse_array(values):
+def _parse_array(values, type_fn=str):
   if values:
-    return values.split(",")
-
-
-def _parse_int_array(values):
-  if values:
-    return [int(val) for val in values.split(",")]
+    return [type_fn(val) for val in values.split(",") if val]
 
 
 def _parse_set(values):
@@ -57,7 +53,8 @@ def _get_toco_converter(flags):
   input_shapes = None
   if flags.input_shapes:
     input_shapes_list = [
-        _parse_int_array(shape) for shape in flags.input_shapes.split(":")
+        _parse_array(shape, type_fn=int)
+        for shape in flags.input_shapes.split(":")
     ]
     input_shapes = dict(zip(input_arrays, input_shapes_list))
   output_arrays = _parse_array(flags.output_arrays)
@@ -103,8 +100,8 @@ def _convert_model(flags):
 
   if flags.mean_values and flags.std_dev_values:
     input_arrays = converter.get_input_arrays()
-    std_dev_values = _parse_int_array(flags.std_dev_values)
-    mean_values = _parse_int_array(flags.mean_values)
+    std_dev_values = _parse_array(flags.std_dev_values, type_fn=int)
+    mean_values = _parse_array(flags.mean_values, type_fn=int)
     quant_stats = zip(mean_values, std_dev_values)
     if ((not flags.input_arrays and len(input_arrays) > 1) or
         (len(input_arrays) != len(quant_stats))):
@@ -130,6 +127,9 @@ def _convert_model(flags):
   if flags.allow_custom_ops:
     converter.allow_custom_ops = flags.allow_custom_ops
   if flags.quantize_weights:
+    if flags.inference_type == lite_constants.QUANTIZED_UINT8:
+      raise ValueError("--quantized_weights is not supported with "
+                       "--inference_type=QUANTIZED_UINT8")
     converter.quantize_weights = flags.quantize_weights
   if flags.dump_graphviz_dir:
     converter.dump_graphviz_dir = flags.dump_graphviz_dir
@@ -200,6 +200,9 @@ def _check_flags(flags, unparsed):
     raise ValueError("--default_ranges_min and --default_ranges_max must be "
                      "used together")
 
+  if flags.dump_graphviz_video and not flags.dump_graphviz:
+    raise ValueError("--dump_graphviz_video must be used with --dump_graphviz")
+
 
 def run_main(_):
   """Main in toco_convert.py."""
@@ -235,13 +238,13 @@ def run_main(_):
       "--inference_type",
       type=str.upper,
       choices=["FLOAT", "QUANTIZED_UINT8"],
-      help="Target data type of arrays in the output file.")
+      help="Target data type of real-number arrays in the output file.")
   parser.add_argument(
       "--inference_input_type",
       type=str.upper,
       choices=["FLOAT", "QUANTIZED_UINT8"],
-      help=("Target data type of input arrays. Allows for a different type for "
-            "input arrays in the case of quantization."))
+      help=("Target data type of real-number input arrays. Allows for a "
+            "different type for input arrays in the case of quantization."))
 
   # Input and output arrays flags.
   parser.add_argument(
@@ -275,12 +278,12 @@ def run_main(_):
       "--std_dev_values",
       type=str,
       help=("Standard deviation of training data for each input tensor, "
-            "comma-separated. Used for quantization. (default None)"))
+            "comma-separated integers. Used for quantization. (default None)"))
   parser.add_argument(
       "--mean_values",
       type=str,
-      help=("Mean of training data for each input tensor, comma-separated. "
-            "Used for quantization. (default None)"))
+      help=("Mean of training data for each input tensor, comma-separated "
+            "integers. Used for quantization. (default None)"))
   parser.add_argument(
       "--default_ranges_min",
       type=int,