SOFTMAX : 1 (instrs: 4,003)
Number of all operators : 126 (total instrs: 11,484,469,630)
```
+
+## Convert a TensorFlow model into TfLite model
+
+### TensorFlow
+
+TensorFlow provides some kinds of converting guideline. In Python, the [TFLiteConverter](https://www.tensorflow.org/api_docs/python/tf/contrib/lite/TFLiteConverter) class will help you to convert a TensorFlow GraphDef or SavedModel into `output_format` using TOCO. The `output_format` can be `TFLITE` or `GRAPHVIZ_DOT` format. The default `output_format` is `TFLITE`. And there is a Python command line interface for running TOCO, and its name is [tflite_convert](https://github.com/tensorflow/tensorflow/blob/master/tensorflow/lite/python/tflite_convert.py). This converts a TensorFlow GraphDef or SavedModel into `TFLITE` or `GRAPHVIZ_DOT` format like [TFLiteConverter](https://www.tensorflow.org/api_docs/python/tf/contrib/lite/TFLiteConverter). These two way also supports to convert a TensorFlow Keras model into `output_format`. Both functions are implemented using a tool called TOCO.
+
+### with tflkit
+
+The tflkit uses the [tflite_convert](https://github.com/tensorflow/tensorflow/blob/master/tensorflow/lite/python/tflite_convert.py) python command line interface to convert a TensorFlow model into TfLite model. It only supports to convert a TensorFlow GraphDef file into `TFLITE` format file. This tool supports the creation of individual `TFLITE` files for different input shapes. When converting to multiple `TFLITE` files, it needs to put a string called `NAME` in `TFLITE_PATH`. The string `NAME` will be replaced by what is listed in teh `NAME` environment. This tool requires an information file as a parameter. There is an [example file](info/convert.template) for a convert information. The `--tensorflow_path` and `--tensorflow_version` can change the TensorFlow location. By default, it uses `externals/tensorflow` directory.
+
+Convert information:
+ * GRAPHDEF_PATH : Full filepath of file containing frozen TensorFlow GraphDef.
+ * TFLITE_PATH : Full filepath of the output TfLite model. If `NAME` optional environment is used, it must include `NAME` string in the file name. (ex. `[...]/inception_v3_NAME.tflite`)
+ * INPUT : Names of the input arrays, comma-separated.
+ * INPUT_SHAPE : Shapes correspoding to `INPUT`, colon-separated.
+ For the creation of individual `TFLITE` files for different input shapes, space-separated.
+ * OUTPUT : Names of the output arrays, comma-seperated.
+ * NAME(Optional) : Names of the individual `TFLITE` files for different input shapes, space-seperated.
+
+Usage (for example, [InceptionV3](https://storage.googleapis.com/download.tensorflow.org/models/tflite/model_zoo/upload_20180427/inception_v3_2018_04_27.tgz)):
+```
+$ wget https://storage.googleapis.com/download.tensorflow.org/models/tflite/model_zoo/upload_20180427/inception_v3_2018_04_27.tgz
+$ tar xzvf inception_v3_2018_04_27.tgz ./inception_v3.pb
+$ cat > convert.info << END
+GRAPHDEF_PATH="${PWD}/inception_v3.pb"
+TFLITE_PATH="${PWD}/inception_v3.tflite"
+INPUT="input"
+INPUT_SHAPE="1,299,299,3"
+OUTPUT="InceptionV3/Predictions/Reshape_1"
+#NAME=""
+END
+$ ./tflite_convert.sh --info=./convert.info --tensorflow_version=1.12
+$ ls *.tflite
+inception_v3.tflite
+```
+
+Usage (for example, multiple `TFLITE` files):
+```
+$ cat > convert_multiple.info << END
+GRAPHDEF_PATH="${PWD}/inception_v3.pb"
+TFLITE_PATH="${PWD}/inception_v3_NAME.tflite"
+INPUT="input"
+INPUT_SHAPE="1,299,299,3 3,299,299,3"
+OUTPUT="InceptionV3/Predictions/Reshape_1"
+NAME="batch1 batch3"
+END
+$ ./tflite_convert.sh --info=./convert_multiple.info --tensorflow_version=1.12
+$ ls *.tflite
+inception_v3_batch1.tflite
+inception_v3_batch3.tflite
+```
--- /dev/null
+#!/bin/bash
+
+usage()
+{
+ echo "usage : $0"
+ echo " --info=Information file"
+ echo " --tensorflow_path=TensorFlow path (Use externals/tensorflow by default)"
+ echo " --tensorflow_version=TensorFlow version (Must be entered)"
+}
+
+SCRIPT_PATH="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+
+TF_DIR="${SCRIPT_PATH}/../../externals/tensorflow"
+
+for i in "$@"
+do
+ case $i in
+ --info=*)
+ INFO=${i#*=}
+ ;;
+ --tensorflow_path=*)
+ TF_DIR=${i#*=}
+ ;;
+ --tensorflow_version=*)
+ TF_VERSION=${i#*=}
+ ;;
+ -h|--help)
+ usage
+ exit 0
+ ;;
+ *)
+ usage
+ exit 1
+ ;;
+ esac
+ shift
+done
+
+if [ -z "$INFO" ]; then
+ echo "INFO is unset or set to the empty string"
+ usage
+ exit 1
+fi
+if [ -z "$TF_DIR" ]; then
+ echo "tensorflow_path is unset or set to the empty string"
+ usage
+ exit 1
+fi
+if [ -z "$TF_VERSION" ]; then
+ echo "tensorflow_version is unset or set to the empty string"
+ usage
+ exit 1
+fi
+
+if [ ! -x "$(command -v bazel)" ]; then
+ echo "Cannot find bazel. Please install bazel."
+ exit 1
+fi
+
+source $INFO
+
+if [ -z "$GRAPHDEF_PATH" ]; then
+ echo "GRAPHDEF_PATH is unset or set to the empty string"
+ echo "Update the $INFO file"
+ exit 1
+fi
+if [ -z "$TFLITE_PATH" ]; then
+ echo "TFLITE_PATH is unset or set to the empty string"
+ echo "Update the $INFO file"
+ exit 1
+fi
+if [ -z "$INPUT" ]; then
+ echo "INPUT is unset or set to the empty string"
+ echo "Update the $INFO file"
+ exit 1
+fi
+if [ -z "$INPUT_SHAPE" ]; then
+ echo "INPUT_SHAPE is unset or set to the empty string"
+ echo "Update the $INFO file"
+ exit 1
+fi
+if [ -z "$OUTPUT" ]; then
+ echo "OUTPUT is unset or set to the empty string"
+ echo "Update the $INFO file"
+ exit 1
+fi
+
+CUR_DIR=$(pwd)
+{
+ echo "Enter $TF_DIR"
+ pushd $TF_DIR > /dev/null
+
+ NAME_LIST=()
+ INPUT_SHAPE_LIST=()
+ if [ -z "$NAME" ]; then
+ NAME_LIST+=("$TFLITE_PATH")
+ INPUT_SHAPE_LIST+=("$INPUT_SHAPE")
+ else
+ for name in $NAME; do
+ NAME_LIST[${#NAME_LIST[*]}]="${TFLITE_PATH/NAME/$name}"
+ done
+ for shape in $INPUT_SHAPE; do
+ INPUT_SHAPE_LIST[${#INPUT_SHAPE_LIST[*]}]="$shape"
+ done
+ if (( ${#NAME_LIST[*]} != ${#INPUT_SHAPE_LIST[*]} )); then
+ echo "The number of NAME and INPUT_SHAPE are different"
+ echo "Update the $INFO file"
+ exit 1
+ fi
+ fi
+
+ for (( i=0; i < ${#NAME_LIST[@]}; ++i )); do
+ if [ "${TF_VERSION%%.*}" = "2" ]; then
+ bazel run tensorflow/lite/python:tflite_convert -- \
+ --output_file="${NAME_LIST[$i]}" \
+ --graph_def_file="$GRAPHDEF_PATH" \
+ --input_arrays="$INPUT" \
+ --input_shapes="${INPUT_SHAPE_LIST[$i]}" \
+ --output_arrays="$OUTPUT" \
+ --allow_custom_ops=true
+ else
+ bazel run tensorflow/contrib/lite/python:tflite_convert -- \
+ --output_file="${NAME_LIST[$i]}" \
+ --graph_def_file="$GRAPHDEF_PATH" \
+ --input_arrays="$INPUT" \
+ --input_shapes="${INPUT_SHAPE_LIST[$i]}" \
+ --output_arrays="$OUTPUT" \
+ --allow_custom_ops=true
+ fi
+ done
+ popd
+
+ for (( i=0; i < ${#NAME_LIST[@]}; ++i )); do
+ echo "OUTPUT FILE : ${NAME_LIST[$i]}"
+ done
+}
projects / platform / core / ml / nnfw.git / commitdiff