 Thomas Bouvier
|
3a31d933a1
Compile using C++17 when using PyTorch 2.1 (#3998)
|
1 月之前 |
|
Enrico Minack
|
55a73d4f66
Use pytest<8 so skipif continues to work for test classes (#4060)
|
3 月之前 |
|
Enrico Minack
|
b0a7edad7f
CI: Use specific GPG keyserver to avoid GPG timeout (#4059)
|
3 月之前 |
|
Enrico Minack
|
c8dd86c804
CI: Add protobuf constrait when installing with tensorflow (#4058)
|
3 月之前 |
|
Enrico Minack
|
34c67875c5
Use pip==23.2.1 for pytorch-lightning<1.8.4 (#4057)
|
3 月之前 |
|
Enrico Minack
|
3375443f83
Use Python 3.7 get-pip script where needed (#4056)
|
3 月之前 |
|
Ljubo Nikolić
|
50b7d73bec
CI: Fix warnings & deprecation messages in the workflows (#4055)
|
3 月之前 |
|
Enrico Minack
|
8f70201fbe
Fix mnist download url (#4033)
|
7 月之前 |
|
Enrico Minack
|
9f88e1df14
Move from `docker-compose` to `docker compose` (#4011)
|
9 月之前 |
|
Chen Zhu
|
7e4d99386e
Fix data type in PartialDistributedGradientTape (#3985)
|
1 年之前 |
|
Jinzhe Zeng
|
8724eff14e
detect MPICH via `HYDRA` (#3984)
|
1 年之前 |
|
Max H. Gerlach
|
850bb33f17
Pin cudf-cu11 and dask-cudf-cu11 for horovod-nvtabular docker build (#3981)
|
1 年之前 |
|
Ata Fatahi
|
ca21d8d56d
Add backward_passes_per_step for gradient aggregation with partial distributed optimizer (#3959)
|
1 年之前 |
|
Thomas Bouvier
|
ff4ca09d39
Fix build on gcc 13: add missing headers (#3957)
|
1 年之前 |
|
Max H. Gerlach
|
f9435b8f7e
CI: Bump container images to CUDA 11.6.2 and remove CUDA 10.x container configs (#3977)
|
1 年之前 |
|
Zhong Zhenyu
|
42b2cdf480
Fix missing -iface argument in generated mpirun command (#3946)
|
1 年之前 |
|
Max H. Gerlach
|
1d217b5994
Bump version to v0.28.1 (#3942)
|
1 年之前 |
|
gaopengff
|
3d24900a75
Fix local_rank_op for tensorflow (#3940)
|
1 年之前 |
|
will-HPE
|
c1a8808367
Fix torch build on ROCm by avoiding dynamically loaded cuCtxGetDevice (#3928)
|
1 年之前 |
|
Max H. Gerlach
|
7f875b2ede
Docs: Fix invalid link to local section (#3937)
|
1 年之前 |
|
Max H. Gerlach
|
b93a87a6c7
Update gloo submodule to commit d96897b (#3925)
|
1 年之前 |
|
Max H. Gerlach
|
19093eaebe
Bump version to 0.28.0 (#3910)
|
1 年之前 |
|
romerojosh
|
004fd0d93f
Update with_device functions in MXNet and PyTorch to skip unnecessary cudaSetDevice calls (#3912)
|
1 年之前 |
|
Max H. Gerlach
|
e1e37c1a92
CI: Unpin requests for docker-compose (#3915)
|
1 年之前 |
|
Max H. Gerlach
|
1ea00673d9
CI: Pin requests<2.29.2 for docker-compose, update readthedocs build image (#3913)
|
1 年之前 |
|
Nicolas Castet
|
67ea0427a8
Fix missing Keras variables shim (#3907)
|
1 年之前 |
|
Max H. Gerlach
|
39c8f7cfed
Fix build with tf-nightly (#3906)
|
1 年之前 |
|
pravee2
|
9d61b6d2c6
Fix keras allreduce function (#3905)
|
1 年之前 |
|
Enrico Minack
|
3a9bf1ba1f
Upgrade CI test frameworks and fix linking NCCL 2.12+ (#3846)
|
1 年之前 |
|
i-kosarev
|
438d3e86b9
test: fix failcase exit code requirements (#3895)
|
1 年之前 |
