Add realistic model input for mnist and option for float model input range Added realistic model input image from mnist. As mnist input range is different from others, added a general option for float model input range. Change-Id: Ib67e0fe0d1a4a52f475ad24978db305cd1da0161

diff --git a/build_tools/gen_mlmodel_input.py b/build_tools/gen_mlmodel_input.py
index 5a910a1..117aacb 100755
--- a/build_tools/gen_mlmodel_input.py
+++ b/build_tools/gen_mlmodel_input.py

@@ -19,6 +19,8 @@
                     help='Model input shape (example: "1, 224, 224, 3")', required=True)
 parser.add_argument('--q', dest='is_quant', action='store_true',
                     help='Indicate it is quant model (default: False)')
+parser.add_argument('--r', dest='float_input_range', default="-1.0, 1.0",
+                    help='Float model input range (default: "-1.0, 1.0")')
 args = parser.parse_args()
 
 
@@ -40,12 +42,15 @@
         (input_shape[1], input_shape[2]))
     input = np.array(resized_img).reshape(np.prod(input_shape))
     if not is_quant:
-        input = 2.0 / 255.0 * input - 1
+        low = np.min(float_input_range)
+        high = np.max(float_input_range)
+        input = (high - low) * input / 255.0 + low
     write_binary_file(output_file, input, is_quant)
 
 
 if __name__ == '__main__':
     # convert input shape to a list
     input_shape = [int(x) for x in args.input_shape.split(',')]
+    float_input_range = [float(x) for x in args.float_input_range.split(',')]
     gen_mlmodel_input(args.input_name, args.output_file,
                       input_shape, args.is_quant)

diff --git a/cmake/iree_model_input.cmake b/cmake/iree_model_input.cmake
index f9e09ca..7be9e35 100644
--- a/cmake/iree_model_input.cmake
+++ b/cmake/iree_model_input.cmake

@@ -27,7 +27,7 @@
   cmake_parse_arguments(
     _RULE
     "QUANT"
-    "NAME;SHAPE;SRC"
+    "NAME;SHAPE;SRC;RANGE"
     ""
     ${ARGN}
   )
@@ -56,6 +56,9 @@
   list(APPEND _ARGS "--i=${_INPUT_FILENAME}")
   list(APPEND _ARGS "--o=${_OUTPUT_BINARY}")
   list(APPEND _ARGS "--s=${_RULE_SHAPE}")
+  if(_RULE_RANGE)
+    list(APPEND _ARGS "--r=${_RULE_RANGE}")
+  endif()
   if(_RULE_QUANT)
     list(APPEND _ARGS "--q")
   endif()

diff --git a/samples/float_model_embedding/CMakeLists.txt b/samples/float_model_embedding/CMakeLists.txt
index fbd36b3..029609e 100644
--- a/samples/float_model_embedding/CMakeLists.txt
+++ b/samples/float_model_embedding/CMakeLists.txt

@@ -158,6 +158,18 @@
     example_images/YellowLabradorLooking_new.jpg"
 )
 
+iree_model_input(
+  NAME
+    mnist_input
+  SHAPE
+    "1, 28, 28, 1"
+  SRC
+    "https://raw.githubusercontent.com/google/iree/main/ \
+    iree/samples/vision/mnist_test.png"
+  RANGE
+    "0, 1"
+)
+
 if(${BUILD_INTERNAL_MODELS})
 
 iree_model_input(
@@ -206,6 +218,7 @@
     "mnist.c"
   DEPS
     ::mnist_bytecode_module_dylib_c
+    ::mnist_input_c
     samples::util::util
   LINKOPTS
     "LINKER:--defsym=__stack_size__=100k"

diff --git a/samples/float_model_embedding/mnist.c b/samples/float_model_embedding/mnist.c
index 3e7d121..5c753a3 100644
--- a/samples/float_model_embedding/mnist.c
+++ b/samples/float_model_embedding/mnist.c

@@ -10,6 +10,7 @@
 
 // Compiled module embedded here to avoid file IO:
 #include "samples/float_model_embedding/mnist_bytecode_module_dylib_c.h"
+#include "samples/float_model_embedding/mnist_input_c.h"
 
 const MlModel kModel = {
     .num_input = 1,
@@ -34,14 +35,8 @@
 
 iree_status_t load_input_data(const MlModel *model, void **buffer) {
   iree_status_t result = alloc_input_buffer(model, buffer);
-  // Populate initial value
-  srand(74886738);
-  if (iree_status_is_ok(result)) {
-    for (int i = 0; i < model->input_length[0]; ++i) {
-      int x = rand();
-      ((float *)*buffer)[i] = (float)x / (float)RAND_MAX;
-    }
-  }
+  const struct iree_file_toc_t *input_file_toc = mnist_input_c_create();
+  memcpy(*buffer, input_file_toc->data, input_file_toc->size);
   return result;
 }
 
@@ -56,5 +51,16 @@
   }
   LOG_INFO("Output #%d data length: %d", index_output,
            mapped_memory->contents.data_length / model->output_size_bytes);
+  // find the label index with best prediction
+  float best_out = 0.0;
+  int best_idx = -1;
+  for (int i = 0; i < model->output_length[index_output]; ++i) {
+    float out = ((float *)mapped_memory->contents.data)[i];
+    if (out > best_out) {
+      best_out = out;
+      best_idx = i;
+    }
+  }
+  LOG_INFO("Digit recognition result is: digit: %d", best_idx);
   return result;
 }