423A35C7/pytorch3d
1
0
Fork 0
mirror of https://github.com/facebookresearch/pytorch3d.git synced 2025-08-02 03:42:50 +08:00
Code Issues Packages Projects Releases Wiki Activity

Builds with pytorch 1.5

Browse Source 
Summary:
Add conda packages for pytorch 1.5. Make wheels be only pytorch 1.5.

Note that pytorch 1.4 has conda packages for cuda 9.2, 10.0 and 10.1, whilst pytorch 1.5 has packages for cuda 9.2, 10.1 and 10.2. We mirror these choices.

Reviewed By: nikhilaravi

Differential Revision: D21157392

fbshipit-source-id: 2f7311e6a83774a6d6c8afb8110b8bd9f37f1454
This commit is contained in:
Jeremy Reizenstein 2020-04-21 12:12:24 -07:00 committed by Facebook GitHub Bot
parent f2b229c1d1
commit a53a2d3731
4 changed files with 147 additions and 76 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

19
.circleci/config.in.yml
View File

@ -207,22 +207,27 @@ workflows:
- main
{{workflows()}}
- binary_linux_conda_cuda:
name: testrun_conda_cuda_py3.7_cu100
name: testrun_conda_cuda_py37_cu100_pyt14
python_version: "3.7"
pytorch_version: "1.4"
cu_version: "cu100"
- binary_linux_conda_cuda:
name: testrun_conda_cuda_py37_cu102_pyt15
python_version: "3.7"
pytorch_version: "1.5"
cu_version: "cu102"
- binary_macos_wheel:
cu_version: cpu
name: macos_wheel_py3.6_cpu
name: macos_wheel_py36_cpu
python_version: '3.6'
pytorch_version: '1.4'
pytorch_version: '1.5'
- binary_macos_wheel:
cu_version: cpu
name: macos_wheel_py3.7_cpu
name: macos_wheel_py37_cpu
python_version: '3.7'
pytorch_version: '1.4'
pytorch_version: '1.5'
- binary_macos_wheel:
cu_version: cpu
name: macos_wheel_py3.8_cpu
name: macos_wheel_py38_cpu
python_version: '3.8'
pytorch_version: '1.4'
pytorch_version: '1.5'

140
.circleci/config.yml
View File

@ -207,87 +207,131 @@ workflows:
- main
- binary_linux_conda:
cu_version: cu92
name: binary_linux_conda_py3.6_cu92
name: linux_conda_py36_cu92_pyt14
python_version: '3.6'
pytorch_version: '1.4'
wheel_docker_image: pytorch/manylinux-cuda92
- binary_linux_conda:
cu_version: cu100
name: binary_linux_conda_py3.6_cu100
name: linux_conda_py36_cu100_pyt14
python_version: '3.6'
pytorch_version: '1.4'
wheel_docker_image: pytorch/manylinux-cuda100
- binary_linux_conda:
cu_version: cu101
name: binary_linux_conda_py3.6_cu101
name: linux_conda_py36_cu101_pyt14
python_version: '3.6'
pytorch_version: '1.4'
- binary_linux_conda:
cu_version: cu92
name: binary_linux_conda_py3.7_cu92
python_version: '3.7'
pytorch_version: '1.4'
wheel_docker_image: pytorch/manylinux-cuda92
- binary_linux_conda:
cu_version: cu100
name: binary_linux_conda_py3.7_cu100
python_version: '3.7'
pytorch_version: '1.4'
wheel_docker_image: pytorch/manylinux-cuda100
name: linux_conda_py36_cu92_pyt15
python_version: '3.6'
pytorch_version: '1.5'
- binary_linux_conda:
cu_version: cu101
name: binary_linux_conda_py3.7_cu101
name: linux_conda_py36_cu101_pyt15
python_version: '3.6'
pytorch_version: '1.5'
- binary_linux_conda:
cu_version: cu102
name: linux_conda_py36_cu102_pyt15
python_version: '3.6'
pytorch_version: '1.5'
- binary_linux_conda:
cu_version: cu92
name: linux_conda_py37_cu92_pyt14
python_version: '3.7'
pytorch_version: '1.4'
- binary_linux_conda:
cu_version: cu100
name: linux_conda_py37_cu100_pyt14
python_version: '3.7'
pytorch_version: '1.4'
- binary_linux_conda:
cu_version: cu101
name: linux_conda_py37_cu101_pyt14
python_version: '3.7'
pytorch_version: '1.4'
- binary_linux_conda:
cu_version: cu92
name: binary_linux_conda_py3.8_cu92
python_version: '3.8'
pytorch_version: '1.4'
wheel_docker_image: pytorch/manylinux-cuda92
- binary_linux_conda:
cu_version: cu100
name: binary_linux_conda_py3.8_cu100
python_version: '3.8'
pytorch_version: '1.4'
wheel_docker_image: pytorch/manylinux-cuda100
name: linux_conda_py37_cu92_pyt15
python_version: '3.7'
pytorch_version: '1.5'
- binary_linux_conda:
cu_version: cu101
name: binary_linux_conda_py3.8_cu101
python_version: '3.8'
pytorch_version: '1.4'
- binary_linux_wheel:
cu_version: cpu
name: binary_linux_wheel_py3.6_cpu
python_version: '3.6'
pytorch_version: '1.4'
- binary_linux_wheel:
cu_version: cpu
name: binary_linux_wheel_py3.7_cpu
name: linux_conda_py37_cu101_pyt15
python_version: '3.7'
pytorch_version: '1.4'
- binary_linux_wheel:
cu_version: cpu
name: binary_linux_wheel_py3.8_cpu
pytorch_version: '1.5'
- binary_linux_conda:
cu_version: cu102
name: linux_conda_py37_cu102_pyt15
python_version: '3.7'
pytorch_version: '1.5'
- binary_linux_conda:
cu_version: cu92
name: linux_conda_py38_cu92_pyt14
python_version: '3.8'
pytorch_version: '1.4'
- binary_linux_conda:
cu_version: cu100
name: linux_conda_py38_cu100_pyt14
python_version: '3.8'
pytorch_version: '1.4'
- binary_linux_conda:
cu_version: cu101
name: linux_conda_py38_cu101_pyt14
python_version: '3.8'
pytorch_version: '1.4'
- binary_linux_conda:
cu_version: cu92
name: linux_conda_py38_cu92_pyt15
python_version: '3.8'
pytorch_version: '1.5'
- binary_linux_conda:
cu_version: cu101
name: linux_conda_py38_cu101_pyt15
python_version: '3.8'
pytorch_version: '1.5'
- binary_linux_conda:
cu_version: cu102
name: linux_conda_py38_cu102_pyt15
python_version: '3.8'
pytorch_version: '1.5'
- binary_linux_wheel:
cu_version: cpu
name: linux_wheel_py36_cpu_pyt15
python_version: '3.6'
pytorch_version: '1.5'
- binary_linux_wheel:
cu_version: cpu
name: linux_wheel_py37_cpu_pyt15
python_version: '3.7'
pytorch_version: '1.5'
- binary_linux_wheel:
cu_version: cpu
name: linux_wheel_py38_cpu_pyt15
python_version: '3.8'
pytorch_version: '1.5'
- binary_linux_conda_cuda:
name: testrun_conda_cuda_py3.7_cu100
name: testrun_conda_cuda_py37_cu100_pyt14
python_version: "3.7"
pytorch_version: "1.4"
cu_version: "cu100"
- binary_linux_conda_cuda:
name: testrun_conda_cuda_py37_cu102_pyt15
python_version: "3.7"
pytorch_version: "1.5"
cu_version: "cu102"
- binary_macos_wheel:
cu_version: cpu
name: macos_wheel_py3.6_cpu
name: macos_wheel_py36_cpu
python_version: '3.6'
pytorch_version: '1.4'
pytorch_version: '1.5'
- binary_macos_wheel:
cu_version: cpu
name: macos_wheel_py3.7_cpu
name: macos_wheel_py37_cpu
python_version: '3.7'
pytorch_version: '1.4'
pytorch_version: '1.5'
- binary_macos_wheel:
cu_version: cpu
name: macos_wheel_py3.8_cpu
name: macos_wheel_py38_cpu
python_version: '3.8'
pytorch_version: '1.4'
pytorch_version: '1.5'

42
.circleci/regenerate.py
View File

@ -11,14 +11,24 @@ import jinja2
import yaml
# The CUDA versions which have pytorch conda packages available for linux for each
# version of pytorch.
CONDA_CUDA_VERSIONS = {
"1.4": ["cu92", "cu100", "cu101"],
"1.5": ["cu92", "cu101", "cu102"],
}
def workflows(prefix="", filter_branch=None, upload=False, indentation=6):
w = []
for btype in ["conda"]:
for python_version in ["3.6", "3.7", "3.8"]:
for cu_version in ["cu92", "cu100", "cu101"]:
for pytorch_version in ["1.4", "1.5"]:
for cu_version in CONDA_CUDA_VERSIONS[pytorch_version]:
w += workflow_pair(
btype=btype,
python_version=python_version,
pytorch_version=pytorch_version,
cu_version=cu_version,
prefix=prefix,
upload=upload,
@ -30,6 +40,7 @@ def workflows(prefix="", filter_branch=None, upload=False, indentation=6):
w += workflow_pair(
btype=btype,
python_version=python_version,
pytorch_version="1.5",
cu_version=cu_version,
prefix=prefix,
upload=upload,
@ -40,16 +51,26 @@ def workflows(prefix="", filter_branch=None, upload=False, indentation=6):
def workflow_pair(
*, btype, python_version, cu_version, prefix="", upload=False, filter_branch
*,
btype,
python_version,
pytorch_version,
cu_version,
prefix="",
upload=False,
filter_branch,
):
w = []
base_workflow_name = f"{prefix}binary_linux_{btype}_py{python_version}_{cu_version}"
py = python_version.replace(".", "")
pyt = pytorch_version.replace(".", "")
base_workflow_name = f"{prefix}linux_{btype}_py{py}_{cu_version}_pyt{pyt}"
w.append(
generate_base_workflow(
base_workflow_name=base_workflow_name,
python_version=python_version,
pytorch_version=pytorch_version,
cu_version=cu_version,
btype=btype,
filter_branch=filter_branch,
@ -70,21 +91,22 @@ def workflow_pair(
def generate_base_workflow(
*, base_workflow_name, python_version, cu_version, btype, filter_branch=None
*,
base_workflow_name,
python_version,
cu_version,
pytorch_version,
btype,
filter_branch=None,
):
d = {
"name": base_workflow_name,
"python_version": python_version,
"cu_version": cu_version,
"pytorch_version": "1.4",
"pytorch_version": pytorch_version,
}
if cu_version == "cu92":
d["wheel_docker_image"] = "pytorch/manylinux-cuda92"
elif cu_version == "cu100":
d["wheel_docker_image"] = "pytorch/manylinux-cuda100"
if filter_branch is not None:
d["filters"] = {"branches": {"only": filter_branch}}

6
INSTALL.md
View File

@ -9,17 +9,17 @@ The core library is written in PyTorch. Several components have underlying imple
- Linux or macOS or Windows
- Python ≥ 3.6
- PyTorch 1.4
- PyTorch 1.4 or 1.5
- torchvision that matches the PyTorch installation. You can install them together at pytorch.org to make sure of this.
- gcc & g++ ≥ 4.9
- CUDA 9.2 or 10.0 or 10.1
- [fvcore](https://github.com/facebookresearch/fvcore)
- If CUDA is to be used, use at least version 9.2.
These can be installed by running:
```
conda create -n pytorch3d python=3.8
conda activate pytorch3d
conda install -c pytorch pytorch torchvision cudatoolkit=10.0
conda install -c pytorch pytorch torchvision cudatoolkit=10.2
conda install -c conda-forge -c fvcore fvcore
```