updated docker image sizes

docs/README.md

@@ -50,7 +50,7 @@ Instructions to recreate our implementation are linked below.
 4. [> SiMLInt ML Training Implementation](./training_implementation.md)
 5. [> SiMLInt Simulation](./inference.md)
 
-SiMLInt Docker images have been built for [CPU](https://github.com/orgs/EPCCed/packages/container/package/simlint) and [GPU](https://github.com/orgs/EPCCed/packages/container/package/simlint-gpu). The CPU version can perform run BOUT++ Hasegawa-Wakatani simulations, generate ground-truth data, or run SiMLInt simulations with inference, while the GPU version is intended for use in training ML models (note: The [NVIDIA Container Toolkit](https://github.com/NVIDIA/nvidia-container-toolkit) is required to run CUDA images). For Docker container usage instructions, click [here](docker-images.md).
+SiMLInt Docker images have been built for [CPU](https://github.com/orgs/EPCCed/packages/container/package/simlint) (7.6 GB) and [GPU](https://github.com/orgs/EPCCed/packages/container/package/simlint-gpu) (17.1 GB). The CPU version can perform run BOUT++ Hasegawa-Wakatani simulations, generate ground-truth data, or run SiMLInt simulations with inference, while the GPU version is intended for use in training ML models (note: The [NVIDIA Container Toolkit](https://github.com/NVIDIA/nvidia-container-toolkit) is required to run CUDA images). For Docker container usage instructions, click [here](docker-images.md).
 
 ## References
 

docs/docker-images.md

@@ -1,6 +1,6 @@
 # SiMLInt Docker Images
 
-Containerised version of SiMLInt have been built for [CPU](https://github.com/EPCCed/SiMLInt/pkgs/container/simlint) and [GPU](https://github.com/EPCCed/SiMLInt/pkgs/container/simlint-gpu).
+Containerised version of SiMLInt have been built for [CPU](https://github.com/EPCCed/SiMLInt/pkgs/container/simlint) (7.6 GB) and [GPU](https://github.com/EPCCed/SiMLInt/pkgs/container/simlint-gpu) (17.1 GB).
 
 The CPU image can be used to run simulations with and without LC, and to produce ground-truth data. The GPU image can be used to train ML models.
 
@@ -27,7 +27,7 @@ docker pull ghcr.io/epcced/simlint:v1.0.0
 or
 
 ```shell
-docker pull ghcr.io/epcced/simlint-gpu:v1.0.0dev
+docker pull ghcr.io/epcced/simlint-gpu:v1.0.0
 ```
 
 ## Running the SiMLInt (CPU) Image

files/container/README.md

@@ -1,7 +1,7 @@
 # SiMLInt Docker Containers
 
 ## SiMLInt CPU Container
-A Docker container, [simlint](https://github.com/EPCCed/SiMLInt/pkgs/container/simlint) (13 GB), which can perform run BOUT Hasegawa-Wakatani simulations, generate ground-truth data, or run SiMLInt simulations with inference, has been made available, with the SiMLInt components installed on an Ubuntu 22.04 image using GCC, OpenMPI and Miniconda3. For Docker container usage instructions, click [here](../../docs/docker-images.md).
+A Docker container, [simlint](https://github.com/EPCCed/SiMLInt/pkgs/container/simlint) (7.6 GB), which can perform run BOUT Hasegawa-Wakatani simulations, generate ground-truth data, or run SiMLInt simulations with inference, has been made available, with the SiMLInt components installed on an Ubuntu 22.04 image using GCC, OpenMPI and Miniconda3. For Docker container usage instructions, click [here](../../docs/docker-images.md).
 
 ## SiMLInt GPU Container
-A Docker container, [simlint-gpu]([https://github.com/EPCCed/SiMLInt/pkgs/container/simlint](https://github.com/EPCCed/SiMLInt/pkgs/container/simlint-gpu)) (22 GB), has also been made available. This has been built on an NVIDIA CUDA image, itself built on Ubuntu 22.04. This image can be used on hardware with GPUs and is intended for training of ML models. For Docker container usage instructions, click [here](../../docs/docker-images.md).
+A Docker container, [simlint-gpu]([https://github.com/EPCCed/SiMLInt/pkgs/container/simlint](https://github.com/EPCCed/SiMLInt/pkgs/container/simlint-gpu)) (17.1 GB), has also been made available. This has been built on an NVIDIA CUDA image, itself built on Ubuntu 22.04. This image can be used on hardware with GPUs and is intended for training of ML models. For Docker container usage instructions, click [here](../../docs/docker-images.md).

files/container/gpu/Dockerfile

@@ -80,6 +80,9 @@ RUN $conda init && . $bashrc && conda activate boutsmartsim && pip install numpy
 #     && cp build/examples/hasegawa-wakatani/hasegawa-wakatani /simlint-bin/gt-hw \
 #     && chmod 777 /simlint-bin/gt-hw
 
+ARG CUDNN_LIBRARY_PATH=/usr/lib/x86_64-linux-gnu
+ARG CUDNN_INCLUDE_PATH=/usr/include
+
 RUN $conda init && . $bashrc && conda activate boutsmartsim \
     && git lfs install \
     && python -m pip install smartsim\[ml\] \
@@ -103,7 +106,7 @@ RUN $conda init && . $bashrc && conda activate boutsmartsim \
 
 COPY files/container/entrypoint.sh /entrypoint.sh
 # COPY files/container/smartsim-hw.sh /simlint-bin/smartsim-hw.sh
-# COPY files/6-simulation/start_db.py /start_db.py
+COPY files/5-training/ /training
 
 # RUN mkdir -p /sim/data && chmod -R 777 /sim
 # RUN chmod -R 777 /miniconda3
@@ -112,9 +115,6 @@ COPY files/container/entrypoint.sh /entrypoint.sh
 
 # USER simlint
 
-ARG CUDNN_LIBRARY_PATH=/usr/lib/x86_64-linux-gnu
-ARG CUDNN_INCLUDE_PATH=/usr/include
-
 WORKDIR /
 
 ENTRYPOINT [ "/entrypoint.sh" ]