Jax Model Conversion For TFLite

Overview

This CodeLab demonstrates how to build a model for MNIST recognition using Jax, and how to convert it to TensorFlow Lite. This codelab will also demonstrate how to optimize the Jax-converted TFLite model with post-training quantiztion.

View on TensorFlow.org Run in Google Colab View source on GitHub Download notebook

Prerequisites

It's recommended to try this feature with the newest TensorFlow nightly pip build.

pip install tf-nightly --upgrade
pip install jax --upgrade
pip install jaxlib --upgrade

Data Preparation

Download the MNIST data with Keras dataset and pre-process.

import numpy as np
import tensorflow as tf
import functools

import time
import itertools

import numpy.random as npr

import jax.numpy as jnp
from jax import jit, grad, random
from jax.experimental import optimizers
from jax.experimental import stax
def _one_hot(x, k, dtype=np.float32):
  """Create a one-hot encoding of x of size k."""
  return np.array(x[:, None] == np.arange(k), dtype)

(train_images, train_labels), (test_images, test_labels) = tf.keras.datasets.mnist.load_data()
train_images, test_images = train_images / 255.0, test_images / 255.0
train_images = train_images.astype(np.float32)
test_images = test_images.astype(np.float32)

train_labels = _one_hot(train_labels, 10)
test_labels = _one_hot(test_labels, 10)
Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz
11493376/11490434 [==============================] - 0s 0us/step
11501568/11490434 [==============================] - 0s 0us/step

Build the MNIST model with Jax

def loss(params, batch):
  inputs, targets = batch
  preds = predict(params, inputs)
  return -jnp.mean(jnp.sum(preds * targets, axis=1))

def accuracy(params, batch):
  inputs, targets = batch
  target_class = jnp.argmax(targets, axis=1)
  predicted_class = jnp.argmax(predict(params, inputs), axis=1)
  return jnp.mean(predicted_class == target_class)

init_random_params, predict = stax.serial(
    stax.Flatten,
    stax.Dense(1024), stax.Relu,
    stax.Dense(1024), stax.Relu,
    stax.Dense(10), stax.LogSoftmax)

rng = random.PRNGKey(0)
WARNING:absl:No GPU/TPU found, falling back to CPU. (Set TF_CPP_MIN_LOG_LEVEL=0 and rerun for more info.)

Train & Evaluate the model

step_size = 0.001
num_epochs = 10
batch_size = 128
momentum_mass = 0.9


num_train = train_images.shape[0]
num_complete_batches, leftover = divmod(num_train, batch_size)
num_batches = num_complete_batches + bool(leftover)

def data_stream():
  rng = npr.RandomState(0)
  while True:
    perm = rng.permutation(num_train)
    for i in range(num_batches):
      batch_idx = perm[i * batch_size:(i + 1) * batch_size]
      yield train_images[batch_idx], train_labels[batch_idx]
batches = data_stream()

opt_init, opt_update, get_params = optimizers.momentum(step_size, mass=momentum_mass)

@jit
def update(i, opt_state, batch):
  params = get_params(opt_state)
  return opt_update(i, grad(loss)(params, batch), opt_state)

_, init_params = init_random_params(rng, (-1, 28 * 28))
opt_state = opt_init(init_params)
itercount = itertools.count()

print("\nStarting training...")
for epoch in range(num_epochs):
  start_time = time.time()
  for _ in range(num_batches):
    opt_state = update(next(itercount), opt_state, next(batches))
  epoch_time = time.time() - start_time

  params = get_params(opt_state)
  train_acc = accuracy(params, (train_images, train_labels))
  test_acc = accuracy(params, (test_images, test_labels))
  print("Epoch {} in {:0.2f} sec".format(epoch, epoch_time))
  print("Training set accuracy {}".format(train_acc))
  print("Test set accuracy {}".format(test_acc))
Starting training...
Epoch 0 in 4.26 sec
Training set accuracy 0.8729000091552734
Test set accuracy 0.880299985408783
Epoch 1 in 3.55 sec
Training set accuracy 0.8983666896820068
Test set accuracy 0.9047999978065491
Epoch 2 in 3.79 sec
Training set accuracy 0.9102166891098022
Test set accuracy 0.9138000011444092
Epoch 3 in 3.63 sec
Training set accuracy 0.9172499775886536
Test set accuracy 0.9218999743461609
Epoch 4 in 3.72 sec
Training set accuracy 0.9224500060081482
Test set accuracy 0.9254000186920166
Epoch 5 in 3.62 sec
Training set accuracy 0.9272000193595886
Test set accuracy 0.930899977684021
Epoch 6 in 3.74 sec
Training set accuracy 0.9327666759490967
Test set accuracy 0.9334999918937683
Epoch 7 in 3.55 sec
Training set accuracy 0.9360166788101196
Test set accuracy 0.9370999932289124
Epoch 8 in 3.76 sec
Training set accuracy 0.9390000104904175
Test set accuracy 0.939300000667572
Epoch 9 in 3.60 sec
Training set accuracy 0.9425666928291321
Test set accuracy 0.9430000185966492

Convert to TFLite model.

Note here, we

  1. Inline the params to the Jax predict func with functools.partial.
  2. Build a jnp.zeros, this is a "placeholder" tensor used for Jax to trace the model.
  3. Call experimental_from_jax: > * The serving_func is wrapped in a list. > * The input is associated with a given name and passed in as an array wrapped in a list.
serving_func = functools.partial(predict, params)
x_input = jnp.zeros((1, 28, 28))
converter = tf.lite.TFLiteConverter.experimental_from_jax(
    [serving_func], [[('input1', x_input)]])
tflite_model = converter.convert()
with open('jax_mnist.tflite', 'wb') as f:
  f.write(tflite_model)
2021-09-08 11:23:09.594301: W tensorflow/compiler/mlir/lite/python/tf_tfl_flatbuffer_helpers.cc:363] Ignored output_format.
2021-09-08 11:23:09.594350: W tensorflow/compiler/mlir/lite/python/tf_tfl_flatbuffer_helpers.cc:366] Ignored drop_control_dependency.
WARNING:absl:Buffer deduplication procedure will be skipped when flatbuffer library is not properly loaded
2021-09-08 11:23:09.594359: W tensorflow/compiler/mlir/lite/python/tf_tfl_flatbuffer_helpers.cc:372] Ignored change_concat_input_ranges.

Check the Converted TFLite Model

Compare the converted model's results with the Jax model.

expected = serving_func(train_images[0:1])

# Run the model with TensorFlow Lite
interpreter = tf.lite.Interpreter(model_content=tflite_model)
interpreter.allocate_tensors()
input_details = interpreter.get_input_details()
output_details = interpreter.get_output_details()
interpreter.set_tensor(input_details[0]["index"], train_images[0:1, :, :])
interpreter.invoke()
result = interpreter.get_tensor(output_details[0]["index"])

# Assert if the result of TFLite model is consistent with the JAX model.
np.testing.assert_almost_equal(expected, result, 1e-5)

Optimize the Model

We will provide a representative_dataset to do post-training quantiztion to optimize the model.

def representative_dataset():
  for i in range(1000):
    x = train_images[i:i+1]
    yield [x]

converter = tf.lite.TFLiteConverter.experimental_from_jax(
    [serving_func], [[('x', x_input)]])
tflite_model = converter.convert()
converter.optimizations = [tf.lite.Optimize.DEFAULT]
converter.representative_dataset = representative_dataset
converter.target_spec.supported_ops = [tf.lite.OpsSet.TFLITE_BUILTINS_INT8]
tflite_quant_model = converter.convert()
with open('jax_mnist_quant.tflite', 'wb') as f:
  f.write(tflite_quant_model)
2021-09-08 11:23:11.502205: W tensorflow/compiler/mlir/lite/python/tf_tfl_flatbuffer_helpers.cc:363] Ignored output_format.
2021-09-08 11:23:11.502258: W tensorflow/compiler/mlir/lite/python/tf_tfl_flatbuffer_helpers.cc:366] Ignored drop_control_dependency.
2021-09-08 11:23:11.502266: W tensorflow/compiler/mlir/lite/python/tf_tfl_flatbuffer_helpers.cc:372] Ignored change_concat_input_ranges.
WARNING:absl:Buffer deduplication procedure will be skipped when flatbuffer library is not properly loaded
2021-09-08 11:23:11.580031: W tensorflow/compiler/mlir/lite/python/tf_tfl_flatbuffer_helpers.cc:363] Ignored output_format.
2021-09-08 11:23:11.580077: W tensorflow/compiler/mlir/lite/python/tf_tfl_flatbuffer_helpers.cc:366] Ignored drop_control_dependency.
2021-09-08 11:23:11.580084: W tensorflow/compiler/mlir/lite/python/tf_tfl_flatbuffer_helpers.cc:372] Ignored change_concat_input_ranges.
fully_quantize: 0, inference_type: 6, input_inference_type: 0, output_inference_type: 0
WARNING:absl:Buffer deduplication procedure will be skipped when flatbuffer library is not properly loaded

Evaluate the Optimized Model

expected = serving_func(train_images[0:1])

# Run the model with TensorFlow Lite
interpreter = tf.lite.Interpreter(model_content=tflite_quant_model)
interpreter.allocate_tensors()
input_details = interpreter.get_input_details()
output_details = interpreter.get_output_details()
interpreter.set_tensor(input_details[0]["index"], train_images[0:1, :, :])
interpreter.invoke()
result = interpreter.get_tensor(output_details[0]["index"])

# Assert if the result of TFLite model is consistent with the Jax model.
np.testing.assert_almost_equal(expected, result, 1e-5)

Compare the Quantized Model size

We should be able to see the quantized model is four times smaller than the original model.

du -h jax_mnist.tflite
du -h jax_mnist_quant.tflite
7.2M    jax_mnist.tflite
1.8M    jax_mnist_quant.tflite