-
Notifications
You must be signed in to change notification settings - Fork 23
/
Copy pathDockerfile
40 lines (33 loc) · 1.46 KB
/
Dockerfile
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
# syntax=docker.io/docker/dockerfile:1.7-labs
FROM nvcr.io/nvidia/isaac-lab:2.0.0
# Hide conflicting Vulkan files, if needed.
RUN if [ -e "/usr/share/vulkan" ] && [ -e "/etc/vulkan" ]; then \
mv /usr/share/vulkan /usr/share/vulkan_hidden; \
fi
# Install vs code server and file browser to interact with the running workflow.
RUN curl -fsSL https://code-server.dev/install.sh | bash
RUN curl -fsSL https://raw.githubusercontent.com/filebrowser/get/master/get.sh | bash
WORKDIR /workspace/neural_wbc
COPY \
--exclude=Dockerfile \
--exclude=data/ \
--exclude=logs/ \
--exclude=release/ \
--exclude=run.py \
--exclude=workflows/ \
. /workspace/neural_wbc
RUN ./install_deps.sh
# Bake in a resume policy. This could be a teacher or student policy. We set this first since we
# expect it to change less frequently than the policy below.
ARG RESUME_PATH=neural_wbc/data/data/policy/h1:student/
COPY ${RESUME_PATH} /workspace/neural_wbc/policy/resume
ENV RESUME_PATH=/workspace/neural_wbc/policy/resume
ARG RESUME_CHECKPOINT=model_4000.pt
ENV RESUME_CHECKPOINT=${RESUME_CHECKPOINT}
# Bake in the main policy. This could be a teacher or student policy. Depending on the use case this
# is used as the teacher policy or the evaluation policy.
ARG POLICY_PATH=neural_wbc/data/data/policy/h1:teacher/
COPY ${POLICY_PATH} /workspace/neural_wbc/policy/
ENV POLICY_PATH=/workspace/neural_wbc/policy/
ARG POLICY_CHECKPOINT=model_76000.pt
ENV POLICY_CHECKPOINT=${POLICY_CHECKPOINT}