 Abin Shahab
|
d261b5e845
Support resurrecting blacklisted hosts (#3319)
|
2 年之前 |
|
Max H. Gerlach
|
a5edcd02b7
Update to CMake 3.13 for better CUDA support and to enable build concurrency (#3261)
|
2 年之前 |
|
Abin Shahab
|
a729ba75f9
RayExecutor V2: Dynamic executor for elastic and static jobs (#3230)
|
2 年之前 |
|
TJ Xu
|
bf85497282
Call process_set._setup in init() to point to the correct native lib path (#3258)
|
2 年之前 |
|
chongxiaoc
|
dea40c660b
fix the example of pytorch_lightning_mnist.py (#3245)
|
3 年之前 |
|
Max H. Gerlach
|
e5725478d0
Add in-place broadcast for TensorFlow (#3128)
|
3 年之前 |
|
Travis Addair
|
66ad6d5a35
Bump version to 0.23.0 (#3200)
|
3 年之前 |
|
TJ Xu
|
d9afdaebed
Add barrier call to torch module to support easy synchronization for process sets (#3139)
|
3 年之前 |
|
chongxiaoc
|
bf26c9f428
Spark/Keras: remove bare Keras support (#3191)
|
3 年之前 |
|
Haoyang Chen
|
e41fdbca67
Fix the mapping btw pyspark and numpy (#3146)
|
3 年之前 |
|
Antoni Baum
|
adce8fa1d0
Make RayExecutor use the current placement group if one exists (#3134)
|
3 年之前 |
|
chongxiaoc
|
1a672ed08e
Estimator/Lightning: use lightning datamodule (#3084)
|
3 年之前 |
|
Tixxx
|
6c6eee4737
Tixxx/add quit on nan lightning (#3089)
|
3 年之前 |
|
Nicolas Castet
|
cf9f5dde7a
Fix develop/editable install mode (#3074)
|
3 年之前 |
|
Max H. Gerlach
|
097357455d
Concurrently running collective operations on process subsets [TensorFlow] (#2839)
|
3 年之前 |
|
Travis Addair
|
93a2f2583e
Bumped version to v0.22.1 (#2968)
|
3 年之前 |
|
Peng Zhang
|
52d0b27174
add customized data loader (#2923)
|
3 年之前 |
|
Travis Addair
|
4d624206d5
Added back horovod.tensorflow.keras.Compression (#2945)
|
3 年之前 |
|
bharatjUber
|
37f1273646
Enable supplying logger to lightning trainer to enable comet integration (#2926)
|
3 年之前 |
|
Travis Addair
|
3ff94801fb
Bumped version to v0.22.0 (#2916)
|
3 年之前 |
|
herewj
|
b2c00602c5
FP16 support for GPU tensors in mxnet (#2915)
|
3 年之前 |
|
chongxiaoc
|
41af508155
Keras estimator: set reader's epoch = 1 to avoid sample duplication and drop-out in a single epoch (#2896)
|
3 年之前 |
|
chongxiaoc
|
c27159fa2c
Estimator: add petastorm reader_pool_type into constructor (#2903)
|
3 年之前 |
|
Ilia Glazkov
|
99faede6e3
Support sparse gradients and respect `global_step` parameter in gradient aggregation. (#2879)
|
3 年之前 |
|
Richard Liaw
|
6600568718
[ray] use ray node_id as unique id (#2883)
|
3 年之前 |
|
Amog Kamsetty
|
810ba80a67
[Ray] Support client (#2882)
|
3 年之前 |
|
Amog Kamsetty
|
39aa85b7ef
[Ray] Use num_workers API for Horovod Ray (#2870)
|
3 年之前 |
|
Peng Zhang
|
2436af9d1c
Add Pytorch Lightning spark estimator (#2713)
|
3 年之前 |
|
Amog Kamsetty
|
9fdc720687
[Ray] Use Placement Groups for RayExecutor (#2824)
|
3 年之前 |
|
Max H. Gerlach
|
386be429b1
Add NVTX tracing hooks for profiling with Nsight Systems (#2723)
|
3 年之前 |
