Would you tell me how to make a baseline model In MCUNetV2 Paper?

[asdfrv100]

Hello, Thanks for your work.

I want to know how to make baseline model of MCUNetV2 paper's experiment.(NOT MCUNetV2 NAS)

exactly, Exactly, I want to reproduce the mobilenetV2 r144 w0.5 model shown in figure 6 of your paper(MCUNetV2,figure 6).

I generated Mbv2 model by tensorflow library, but it is too big to use in my board(F746gz-disc board)

I know that paper'model be used QAT, but I need other information for model generation.

Here is my model's inforamition. After learning fp32 MBV2 for only 1 epoch on the fp32 Imagenet-1000 dataset,I used post training quantization to quantize the model with int8 weights and activations. (but It was 2.2MB, not under 1MB.)

Could you please tell me what I need to do to make this model less than 1 MB like the MCUNetV2 paper?

*board information

*my model information

and I uploaded my model tfml file and tensorflow code.

*tfml model file mobilenet_v2_quantized.zip

• Here is my model generation code

more

import os import tensorflow as tf import config as c from tqdm import tqdm from tensorflow.keras import optimizers from utils.data_utils import train_iterator from utils.eval_utils import cross_entropy_batch, correct_num_batch, l2_loss from tensorflow.keras.applications.mobilenet_v2 import MobileNetV2 from model.ResNet import ResNet from model.ResNet_v2 import ResNet_v2 from test import test import numpy as np import os import PIL import PIL.Image import tensorflow as tf import pathlib os.environ['CUDA_VISIBLE_DEVICES'] = '0' os.environ['CUDA_VISIBLE_DEVICES'] = '1' checkpoint_path = "training_1/cp.ckpt" checkpoint_dir = os.path.dirname(checkpoint_path) data_dir = pathlib.Path("/home/poweroverwhelming/modelgen/ImageNet_ResNet_Tensorflow2.0/imagenet10/train") image_count = len(list(data_dir.glob('*/*.JPEG'))) #training param batch_size = 32 img_height = 144 img_width = 144 num_classes =1000 #directroy making train_ds = tf.keras.utils.image_dataset_from_directory( data_dir, validation_split=0.2, subset="training", seed=123, image_size=(img_height, img_width), batch_size=batch_size) val_ds = tf.keras.utils.image_dataset_from_directory( data_dir, validation_split=0.2, subset="validation", seed=123, image_size=(img_height, img_width), batch_size=batch_size) #print class name #nomalization normalization_layer = tf.keras.layers.Rescaling(1./255) normalized_ds = train_ds.map(lambda x, y: (normalization_layer(x), y)) image_batch, labels_batch = next(iter(normalized_ds)) first_image = image_batch[0] #Notice the pixel values are now in `[0,1]`. #print(np.min(first_image), np.max(first_image)) AUTOTUNE = tf.data.AUTOTUNE train_ds = train_ds.cache().prefetch(buffer_size=AUTOTUNE) val_ds = val_ds.cache().prefetch(buffer_size=AUTOTUNE) #training model = MobileNetV2(input_shape=(144, 144, 3), alpha=0.5, weights=None, classes=1000) model.compile( optimizer='adam', loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True), metrics=['accuracy']) cp_callback = tf.keras.callbacks.ModelCheckpoint(filepath=checkpoint_path, save_weights_only=True, verbose=1) model.fit(train_ds, validation_data=val_ds, epochs=50) #loss, acc = model.evaluate(test_images, test_labels, verbose=2) #print("Untrained model, accuracy: {:5.2f}%".format(100 * acc)) #Save the entire model as a SavedModel. #!mkdir -p saved_model model.save('saved_model/my_model') def representative_data_gen(): for input_value in tf.data.Dataset.from_tensor_slices(train_ds).batch(1).take(100): # Model has only one input so each data point has one element. yield [input_value] def representative_dataset(): for _ in range(100): data = np.random.rand(1, 144, 144, 3) yield [data.astype(np.float32)] #PTQ part #Convert the model to TensorFlow Lite format converter = tf.lite.TFLiteConverter.from_saved_model('saved_model/my_model') #Perform post-training quantization converter.optimizations = [tf.lite.Optimize.DEFAULT] converter.target_spec.supported_ops = [tf.lite.OpsSet.TFLITE_BUILTINS_INT8] converter.inference_input_type = tf.int8 converter.inference_output_type = tf.int8 converter.representative_dataset = representative_dataset #Save the quantized model tflite_model = converter.convert() open("mobilenet_v2_imgnet10_r144_w05_quantizaed.tflite", "wb").write(tflite_model)

once again, Thank you for your work.