From a91771a1327a5519b7133be785927edd6fdc733b Mon Sep 17 00:00:00 2001 From: Ajay Jain Date: Sat, 15 Feb 2020 01:07:51 -0800 Subject: [PATCH] Update README with CPU instructions --- README.md | 18 ++++++++++++++++-- setup.py | 1 - 2 files changed, 16 insertions(+), 3 deletions(-) diff --git a/README.md b/README.md index c556f0e..43b3602 100644 --- a/README.md +++ b/README.md @@ -24,19 +24,33 @@ $ conda activate checkmate-mlsys-artifact ``` ### Step 2: Install the `remat` package and dependencies +Clone this repository and check out the `mlsys20_artifact` branch: +``` +$ git clone git@github.com:parasj/checkmate.git +$ cd checkmate +$ git checkput mlsys20_artifact +``` From this directory, ``` $ conda install -c conda-forge python-graphviz $ pip install -e . ``` +If you are running setup on a machine without a GPU, run, +``` +$ pip install tensorflow-cpu +``` +CPU-only is enough to use this artifact. If you have a GPU-enabled machine with CUDA installed, you can instead run +``` +$ pip install tensorflow>=2.0.0 +``` ### Step 3: Install Gurobi Checkmate uses the Gurobi optimziation library to solve an integer linear program that chooses a recomputation schedule for a given neural network architecture. This requires a license to Gurobi, which is free for academic use. The `grbgetkey` command used below must be run on a computer connected to a university network directly or via a VPN. -1. Please follow these instructions to install Gurobi on your system: https://www.gurobi.com/documentation/quickstart.html +1. Please follow these instructions to install Gurobi on your system: https://www.gurobi.com/documentation/quickstart.html. For example, on Linux, follow `https://www.gurobi.com/documentation/9.0/quickstart_linux/software_installation_guid.html`. 2. Make an academic account with Gurobi at: https://pages.gurobi.com/registration 3. Request an acadmic license at: https://www.gurobi.com/downloads/end-user-license-agreement-academic/ -4. Install the license by running the `grbgetkey` command at the end of the page. Ensure you are on an academic network like Airbears2. +4. Install the license by running the `grbgetkey` command at the end of the page. Ensure you are on an academic network like Airbears2 for UC Berkeley. If you save the license to a non-default location (outside your home directory), you will need to export the `GRB_LICENSE_FILE` variable with the path to the licence. 5. Set up the gurobipy Anaconda channel by running `conda config --add channels http://conda.anaconda.org/gurobi` 6. Install gurobipy by running: `conda install gurobi` diff --git a/setup.py b/setup.py index 8d184cb..8774c63 100644 --- a/setup.py +++ b/setup.py @@ -16,7 +16,6 @@ "ray>=0.7.5", "graphviz", "python-dotenv", - "tensorflow>=2.0.0", "pytest", "pillow", "gym",