From e90562e89a0b2d4ae20da60cf54d59001b8b3a1c Mon Sep 17 00:00:00 2001 From: Francesco Conti Date: Thu, 9 Apr 2020 14:25:10 +0200 Subject: [PATCH] Update README.md (#5) --- README.md | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-) diff --git a/README.md b/README.md index f13d1f7..8238e05 100644 --- a/README.md +++ b/README.md @@ -17,15 +17,23 @@ All the quantized representations support mixed-precision weights (signed and as NEMO is organized as a Python library that can be applied with relatively small changes to an existing PyTorch based script or training framework. +# Installation and requirements +The NEMO library currently supports PyTorch >= 1.3 and runs on Python >= 3.5. +To install it from PyPI, just run +``` +pip install pytorch-nemo +``` +Then, you can import it in your script using +``` +import nemo +``` + # Example - MNIST post-training quantization: https://colab.research.google.com/drive/1AmcITfN2ELQe07WKQ9szaxq-WSu4hdQb # License NEMO is released under Apache 2.0, see the LICENSE file in the root of this repository for details. -# Requirements -The NEMO library (NEural Minimizer for tOrch) currently supports PyTorch >= 1.3. - # Acknowledgements ![ALOHA Logo](/var/aloha.png)