Skip to main content

AIMET torch Package

Project description

#==============================================================================
# @@-COPYRIGHT-START-@@
#
# Copyright (c) 2021-2023, Qualcomm Innovation Center, Inc. All rights reserved.
#
# SPDX-License-Identifier: BSD-3-Clause
#
# @@-COPYRIGHT-END-@@
#==============================================================================

========
Overview
========
AI Model Efficiency Toolkit (AIMET) is a library that provides advanced model
quantization and model compression techniques for trained neural network
models. It provides features that have been proven to improve run-time
performance of deep learning neural network models with lower compute and
memory requirements and minimal impact to task accuracy.

Features
========
AIMET supports the following features

- Model Quantization
- Quantization simulation: Simulates on-target quantized inference.
Specifically simulates Qualcomm SnapDragon DSP accelerators.
- Quantization-aware training: Fine-tune models to improve on-target
quantized accuracy
- Data Free quantization: Post-training technique to improve quantized
accuracy by equalizing model weights (Cross-Layer Equalization) and
correcting shifts in layer outputs due to quantization (Bias Correction)

- Model Compression
- Spatial SVD: Tensor decomposition technique to split a large layer
into two smaller ones
- Channel Pruning: Removes redundant input channels of convolutional
layers and modifies the model graph accordingly
- Compression-ratio Selection: Automatically selects per-layer compression
ratios

============
Dependencies
============
See the https://quic.github.io/aimet-pages/releases/latest/install/index.html for details.

=============
Documentation
=============
Please refer to the Documentation at https://quic.github.io/aimet-pages/index.html
for the user guide and API documentation.

=================
Using the Package
=================
Please see https://github.com/quic/aimet#getting-started for package requirements
and usage.


Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distribution

aimet_torch-1.30.1-cp38-cp38-manylinux_2_29_x86_64.whl (10.7 MB view hashes)

Uploaded CPython 3.8 manylinux: glibc 2.29+ x86-64

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page