gpt-neox

1
# Copyright (c) 2024, EleutherAI
2
#
3
# Licensed under the Apache License, Version 2.0 (the "License");
4
# you may not use this file except in compliance with the License.
5
# You may obtain a copy of the License at
6
#
7
#     http://www.apache.org/licenses/LICENSE-2.0
8
#
9
# Unless required by applicable law or agreed to in writing, software
10
# distributed under the License is distributed on an "AS IS" BASIS,
11
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12
# See the License for the specific language governing permissions and
13
# limitations under the License.
14

15
FROM nvcr.io/nvidia/pytorch:24.02-py3
16

17
ENV DEBIAN_FRONTEND=noninteractive
18

19
# metainformation
20
LABEL org.opencontainers.image.version = "2.0"
21
LABEL org.opencontainers.image.authors = "contact@eleuther.ai"
22
LABEL org.opencontainers.image.source = "https://www.github.com/eleutherai/gpt-neox"
23
LABEL org.opencontainers.image.licenses = " Apache-2.0"
24
LABEL org.opencontainers.image.base.name="nvcr.io/nvidia/pytorch:24.02-py3"
25

26
#### System package (uses default Python 3 version in Ubuntu 20.04)
27
RUN apt-get update -y && \
28
    apt-get install -y \
29
    git python3-dev libpython3-dev python3-pip sudo pdsh \
30
    htop tmux zstd software-properties-common build-essential autotools-dev \
31
    nfs-common pdsh cmake g++ gcc curl wget vim less unzip htop iftop iotop ca-certificates ssh \
32
    rsync iputils-ping net-tools libcupti-dev libmlx4-1 infiniband-diags ibutils ibverbs-utils \
33
    rdmacm-utils perftest rdma-core nano && \
34
    update-alternatives --install /usr/bin/python python /usr/bin/python3 1 && \
35
    update-alternatives --install /usr/bin/pip pip /usr/bin/pip3 1 && \
36
    python -m pip install --upgrade pip && \
37
    python -m pip install gpustat
38

39
### SSH
40
RUN mkdir /var/run/sshd && \
41
    # Prevent user being kicked off after login
42
    sed -i 's@session\s*required\s*pam_loginuid.so@session optional pam_loginuid.so@g' /etc/pam.d/sshd && \
43
    echo 'AuthorizedKeysFile     .ssh/authorized_keys' >> /etc/ssh/sshd_config && \
44
    echo 'PasswordAuthentication yes' >> /etc/ssh/sshd_config && \
45
    # FIX SUDO BUG: https://github.com/sudo-project/sudo/issues/42
46
    echo "Set disable_coredump false" >> /etc/sudo.conf
47

48
# Expose SSH port
49
EXPOSE 22
50

51
# Needs to be in docker PATH if compiling other items & bashrc PATH (later)
52
ENV PATH=/usr/local/mpi/bin:${PATH} \
53
    LD_LIBRARY_PATH=/usr/local/lib:/usr/local/mpi/lib:/usr/local/mpi/lib64:${LD_LIBRARY_PATH}
54

55
# Create a wrapper for OpenMPI to allow running as root by default
56
RUN mv /usr/local/mpi/bin/mpirun /usr/local/mpi/bin/mpirun.real && \
57
    echo '#!/bin/bash' > /usr/local/mpi/bin/mpirun && \
58
    echo 'mpirun.real --allow-run-as-root --prefix /usr/local/mpi "$@"' >> /usr/local/mpi/bin/mpirun && \
59
    chmod a+x /usr/local/mpi/bin/mpirun
60

61
#### User account
62
RUN useradd --create-home --uid 1000 --shell /bin/bash mchorse && \
63
    usermod -aG sudo mchorse && \
64
    echo "mchorse ALL=(ALL) NOPASSWD: ALL" >> /etc/sudoers
65

66
## SSH config and bashrc
67
RUN mkdir -p /home/mchorse/.ssh /job && \
68
    echo 'Host *' > /home/mchorse/.ssh/config && \
69
    echo '    StrictHostKeyChecking no' >> /home/mchorse/.ssh/config && \
70
    echo 'export PDSH_RCMD_TYPE=ssh' >> /home/mchorse/.bashrc && \
71
    echo 'export PATH=/home/mchorse/.local/bin:$PATH' >> /home/mchorse/.bashrc && \
72
    echo 'export PATH=/usr/local/mpi/bin:$PATH' >> /home/mchorse/.bashrc && \
73
    echo 'export LD_LIBRARY_PATH=/usr/local/lib:/usr/local/mpi/lib:/usr/local/mpi/lib64:$LD_LIBRARY_PATH' >> /home/mchorse/.bashrc
74

75
#### Python packages
76
COPY requirements/* ./
77
RUN python -m pip install --no-cache-dir -r requirements.txt && pip install -r requirements-onebitadam.txt
78
RUN python -m pip install -r requirements-sparseattention.txt
79
RUN python -m pip install -r requirements-flashattention.txt
80
RUN python -m pip install -r requirements-wandb.txt
81
RUN python -m pip install protobuf==3.20.*
82

83
COPY megatron/fused_kernels/ /megatron/fused_kernels
84
WORKDIR /megatron/fused_kernels
85
RUN python setup.py install
86

87
# Clear staging
88
RUN mkdir -p /tmp && chmod 0777 /tmp
89

90
#### SWITCH TO mchorse USER
91
USER mchorse
92
WORKDIR /home/mchorse
93