Add docker

README.md

@@ -86,6 +86,20 @@ python src/run.py --eval --n_embd 1536 --n_head 16 --n_layer 48
 
 If your data is not in `/root/downloads/`, set `--ckpt_path` and `--data_path` manually. You should see that the test generative losses are 2.0895, 2.0614, and 2.0466, matching Figure 3 in the paper.
 
+### Docker
+
+Also, there is 'Dockerfile' in 'image-gpt/Docker' directory. To make a docker-container of the model and use it, you should use the 'build' command to create an image (including dot in the end):
+
+```
+docker build -t image-gpt -f docker/Dockerfile .
+```
+
+After, you can run it by typing:
+
+```
+nvidia-docker run image-gpt
+```
+
 ### Citation
 
 Please use the following bibtex entry:

docker/Dockerfile

@@ -0,0 +1,26 @@
+FROM nvidia/cuda:10.0-cudnn7-runtime-ubuntu18.04
+RUN apt-get update && \
+ apt-get install -y python3-pip && \
+ pip3 install
+
+# set directory
+WORKDIR /image-gpt/
+
+# requirements
+COPY requirements.txt .
+RUN pip3 install --no-cache-dir -r requirements.txt
+
+# copy executive files
+COPY ["src/", "download.py", "/src/"]
+
+# download pre-trained model
+RUN python3 /src/download.py --model s --ckpt 1000000
+
+# download dataset
+RUN python3 /src/download.py --model s --ckpt 1000000 --dataset imagenet
+
+# download color clusters
+RUN python3 /src/download.py --model s --ckpt 1000000 --dataset imagenet --clusters
+
+# run iGPT-S
+CMD python3 /src/run.py --eval --n_embd 512 --n_head 8 --n_layer 24

requirements.txt

@@ -1,4 +1,8 @@
+numpy==1.16.3
+tensorflow-gpu==1.13.1
+imageio==2.8.0
 fire>=0.1.3
 regex==2017.4.5
 requests==2.21.0
-tqdm==4.31.1
+tqdm==4.31.1 
+scipy==1.4.1