Skip to content

AI Edge Quantizer: flexible post training quantization for LiteRT models.

License

Notifications You must be signed in to change notification settings

google-ai-edge/ai-edge-quantizer

Repository files navigation

AI Edge Quantizer

A quantizer for advanced developers to quantize converted LiteRT models. It aims to facilitate advanced users to strive for optimal performance on resource demanding models (e.g., GenAI models).

Build Status

Build Type Status
Unit Tests (Linux)
Nightly Release
Nightly Colab

Installation

Requirements and Dependencies

  • Python versions: 3.9, 3.10, 3.11
  • Operating system: Linux, MacOS
  • TensorFlow: tf-nightly

Install

Nightly PyPi package:

pip install ai-edge-quantizer-nightly

API Usage

The quantizer requires two inputs:

  1. An unquantized source LiteRT model (with FP32 data type in the FlatBuffers format with .tflite extension)
  2. A quantization recipe (details below)

and outputs a quantized LiteRT model that's ready for deployment on edge devices.

Basic Usage

In a nutshell, the quantizer works according to the following steps:

  1. Instantiate a Quantizer class. This is the entry point to the quantizer's functionalities that the user accesses.
  2. Load a desired quantization recipe (details in subsection).
  3. Quantize (and save) the model. This is where most of the quantizer's internal logic works.
qt = quantizer.Quantizer("path/to/input/tflite")
qt.load_quantization_recipe(recipe.dynamic_wi8_afp32())
qt.quantize().export_model("/path/to/output/tflite")

Please see the getting started colab for the simplest quick start guide on those 3 steps, and the selective quantization colab with more details on advanced features.

LiteRT Model

Please refer to the LiteRT documentation for ways to generate LiteRT models from Jax, PyTorch and TensorFlow. The input source model should be an FP32 (unquantized) model in the FlatBuffers format with .tflite extension.

Quantization Recipe

The user needs to specify a quantization recipe using AI Edge Quantizer's API to apply to the source model. The quantization recipe encodes all information on how a model is to be quantized, such as number of bits, data type, symmetry, scope name, etc.

Essentially, a quantization recipe is defined as a collection of the following command:

“Apply Quantization Algorithm X on Operator Y under Scope Z with ConfigN”.

For example:

"Uniformly quantize the FullyConnected op under scope 'dense1/' with INT8 symmetric with Dynamic Quantization".

All the unspecified ops will be kept as FP32 (unquantized). The scope of an operator in TFLite is defined as the output tensor name of the op, which preserves the hierarchical model information from the source model (e.g., scope in TF). The best way to obtain scope name is by visualizing the model with Model Explorer.

The simplest recipe to get started with is using existing recipes from recipe.py. This is demonstrated in the getting started colab example.

Deployment

Please refer to the LiteRT deployment documentation for ways to deploy a quantized LiteRT model.

Advanced Recipes

There are many ways the user can configure and customize the quantization recipe beyond using a template in recipe.py. For example, the user can configure the recipe to achieve these features:

  • Selective quantization (exclude selected ops from being quantized)
  • Flexible mixed scheme quantization (mixture of different precision, compute precision, scope, op, config, etc)
  • 4-bit weight quantization

The selective quantization colab shows some of these more advanced features.

For specifics of the recipe schema, please refer to the OpQuantizationRecipe in [recipe_manager.py].

For advanced usage involving mixed quantization, the following API may be useful:

  • Use Quantizer:load_quantization_recipe() in quantizer.py to load a custom recipe.
  • Use Quantizer:update_quantization_recipe() in quantizer.py to extend or override specific parts of the recipe.

Operator coverage

The table below outlines the allowed configurations for available recipes.

Config DYNAMIC_WI8_AFP32 DYNAMIC_WI4_AFP32 STATIC_WI8_AI16 STATIC_WI4_AI16 STATIC_WI8_AI8 STATIC_WI4_AI8 WEIGHTONLY_WI8_AFP32 WEIGHTONLY_WI4_AFP32
activation num_bits None None 16 16 8 8 None None
symmetric None None TRUE TRUE [TRUE, FALSE] [TRUE, FALSE] None None
granularity None None TENSORWISE TENSORWISE TENSORWISE TENSORWISE None None
dtype None None INT INT INT INT None None
weight num_bits 8 4 8 4 8 4 8 4
symmetric TRUE TRUE TRUE TRUE TRUE TRUE [TRUE, FALSE] [TRUE, FALSE]
granularity [CHANNELWISE, TENSORWISE] [CHANNELWISE, TENSORWISE] [CHANNELWISE, TENSORWISE] [CHANNELWISE, TENSORWISE] [CHANNELWISE, TENSORWISE] [CHANNELWISE, TENSORWISE] [CHANNELWISE, TENSORWISE] [CHANNELWISE, TENSORWISE]
dtype INT INT INT INT INT INT INT INT
explicit_dequantize FALSE FALSE FALSE FALSE FALSE FALSE TRUE TRUE
compute_precision INTEGER INTEGER INTEGER INTEGER INTEGER INTEGER FLOAT FLOAT

Operators Supporting Quantization

Config DYNAMIC_WI8_AFP32 DYNAMIC_WI4_AFP32 STATIC_WI8_AI16 STATIC_WI4_AI16 STATIC_WI8_AI8 STATIC_WI4_AI8 WEIGHTONLY_WI8_AFP32 WEIGHTONLY_WI4_AFP32
FULLY_CONNECTED
CONV_2D
BATCH_MATMUL
EMBEDDING_LOOKUP
DEPTHWISE_CONV_2D
AVERAGE_POOL_2D
RESHAPE
SOFTMAX
TANH
TRANSPOSE
GELU
ADD
CONV_2D_TRANSPOSE
SUB
MUL
MEAN
RSQRT
CONCATENATION
STRIDED_SLICE
SPLIT
LOGISTIC

About

AI Edge Quantizer: flexible post training quantization for LiteRT models.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published