Commits · 31ca3b97ebc1ca37b1d4db6ff3bf062fcbf16b5d · ModelZoo / ResNet50_tensorflow

15 Jul, 2020 1 commit
- Internal changes. · 8b510f95
  Allen Wang authored Jul 15, 2020
```
PiperOrigin-RevId: 321382083
```
  8b510f95
14 Jul, 2020 1 commit
- Internal change · 900b1e07
  Allen Wang authored Jul 14, 2020
```
PiperOrigin-RevId: 321261790
```
  900b1e07
13 Jul, 2020 2 commits
- Add in MLIR benchmarking for ResNet. · be8c2556
  Allen Wang authored Jul 13, 2020
```
PiperOrigin-RevId: 321039601
```
  be8c2556
- Internal change · 5df3d2e3
  Hongkun Yu authored Jul 13, 2020
```
PiperOrigin-RevId: 321025013
```
  5df3d2e3
01 May, 2020 1 commit
- Internal change · 10b38209
  A. Unique TensorFlower authored May 01, 2020
```
PiperOrigin-RevId: 309479930
```
  10b38209
21 Apr, 2020 1 commit
- Internal change · 338a0fc2
  Zongwei Zhou authored Apr 21, 2020
```
PiperOrigin-RevId: 307641075
```
  338a0fc2
14 Apr, 2020 2 commits

Move benchmark_wrappers to benchmark folder. · d70eca30
Hongkun Yu authored Apr 14, 2020
```
PiperOrigin-RevId: 306521269
```
d70eca30

Set dataset private threads and GPU private threads together for better Resnet... · 0672bdcf

Zongwei Zhou authored Apr 13, 2020

Set dataset private threads and GPU private threads together for better Resnet performance; remove outdated TF1 benchmarks; remove redundant classifier param runtime.gpu_threads_enabled

PiperOrigin-RevId: 306341022

0672bdcf

13 Apr, 2020 1 commit
- Reduce Resnet50 pure eager FP32 test batch size · 1025682f
  Zongwei Zhou authored Apr 13, 2020
```
PiperOrigin-RevId: 306287724
```
  1025682f
12 Apr, 2020 1 commit
- Internal change · 684bddcd
  Hongkun Yu authored Apr 12, 2020
```
PiperOrigin-RevId: 306136273
```
  684bddcd
11 Apr, 2020 1 commit
- Reduce Resnet50 pure eager FP16 test batch size · 7194a5cc
  Zongwei Zhou authored Apr 11, 2020
```
PiperOrigin-RevId: 306064486
```
  7194a5cc
09 Apr, 2020 1 commit
- move perfzero benchmark to benchmark/ · 7351c164
  Hongkun Yu authored Apr 09, 2020
```
PiperOrigin-RevId: 305709689
```
  7351c164
03 Apr, 2020 1 commit
- Internal change · b7c0e5c6
  A. Unique TensorFlower authored Apr 03, 2020
```
PiperOrigin-RevId: 304715282
```
  b7c0e5c6
16 Mar, 2020 1 commit
- Move ResNet related files to the ResNet folder (e.g. utils, CTL, CFit, etc.) · ad710aa1
  Allen Wang authored Mar 16, 2020
```
PiperOrigin-RevId: 301185629
```
  ad710aa1
13 Mar, 2020 1 commit
- Internal change · 225eda71
  Will Cromar authored Mar 13, 2020
```
PiperOrigin-RevId: 300850465
```
  225eda71
11 Mar, 2020 1 commit
- Set `use_tf_while_loop=True` for Resnet eager benchmarks. · 93fc3c9f
  Ruoxin Sang authored Mar 11, 2020
```
PiperOrigin-RevId: 300351756
```
  93fc3c9f
09 Mar, 2020 1 commit
- Correct Resnet CTL data dir · 801c522d
  Zongwei Zhou authored Mar 08, 2020
```
PiperOrigin-RevId: 299764679
```
  801c522d
03 Mar, 2020 1 commit
- Internal change · b31b1982
  Ruoxin Sang authored Mar 02, 2020
```
PiperOrigin-RevId: 298493612
```
  b31b1982
06 Feb, 2020 1 commit
- Increase batch size for Resnet pure eager tests due to recent memory optimizations · faf5ad73
  Zongwei Zhou authored Feb 05, 2020
```
PiperOrigin-RevId: 293498210
```
  faf5ad73
01 Feb, 2020 1 commit
- Enable Resnet50 and BERT 1GPU pure eager throughput test · f4147e22
  Zongwei Zhou authored Jan 31, 2020
```
PiperOrigin-RevId: 292671694
```
  f4147e22
28 Jan, 2020 1 commit
- Internal change · ddc40272
  Sai Ganesh Bandiatmakuri authored Jan 28, 2020
```
PiperOrigin-RevId: 292030392
```
  ddc40272
14 Dec, 2019 1 commit
- Add Resnet-50 CTL FP16 benchmarks · 4b06a97a
  Zongwei Zhou authored Dec 14, 2019
```
PiperOrigin-RevId: 285568490
```
  4b06a97a
27 Nov, 2019 1 commit
- Remove 'default' in get_distribution_strategy which is complex and error-prone · 04256053
  Hongkun Yu authored Nov 26, 2019
```
PiperOrigin-RevId: 282669615
```
  04256053
25 Nov, 2019 1 commit

Inject enable_runtime_flags into benchmarks. · bcce419a

Sai Ganesh Bandiatmakuri authored Nov 25, 2019

This will help general debugging by enabling custom execution with --benchmark_method_steps.

E.g --benchmark_method_steps=train_steps=7 will run the benchmark for only 7 steps without modifying benchmark code.

PiperOrigin-RevId: 282396875

bcce419a

01 Nov, 2019 1 commit

Support fp16 using tf.keras.mixed_precision in CTL resnet. · ffa522ea

Reed Wanderman-Milne authored Nov 01, 2019

To test, I ran the following command:

python resnet_ctl_imagenet_main.py --batch_size=2048 --data_dir ~/imagenet --datasets_num_private_threads=14 --epochs_between_evals=10 --model_dir ~/tmp_model_dir --clean --num_gpus=8 --train_epochs=90 --dtype=fp16

I got 76.15% final evaluation accuracy.

PiperOrigin-RevId: 278010061

ffa522ea

18 Oct, 2019 1 commit
- Move CTL resnet example. · f2c61881
  Hongkun Yu authored Oct 18, 2019
```
PiperOrigin-RevId: 275417626
```
  f2c61881
17 Oct, 2019 1 commit
- Move ResNet CTL benchmark to benchmark folder. · d2f86342
  Hongkun Yu authored Oct 17, 2019
```
PiperOrigin-RevId: 275288636
```
  d2f86342
20 Sep, 2019 1 commit

additional fix for resnet ctl benchamrk. As copybara needs exact spacing in... · 4fa82ae1

A. Unique TensorFlower authored Sep 19, 2019

additional fix for resnet ctl benchamrk. As copybara needs exact spacing in the file, adding back the missing space.

PiperOrigin-RevId: 270167733

4fa82ae1

19 Sep, 2019 1 commit
- Fix for resnet data_dir name in copybara. · 0a79175b
  A. Unique TensorFlower authored Sep 19, 2019
```
PiperOrigin-RevId: 270079001
```
  0a79175b
17 Sep, 2019 2 commits
- auto formatting: remove tailing white space etc. · 9fd17d8c
  Hongkun Yu authored Sep 17, 2019
```
PiperOrigin-RevId: 269653296
```
  9fd17d8c
- Double batch size for fp16 benchmarks · c3db5a9f
  Kaixi Hou authored Sep 17, 2019
  
  c3db5a9f
06 Sep, 2019 1 commit
- minor changes · fcb8590b
  Kaixi Hou authored Sep 06, 2019
  
  fcb8590b
05 Sep, 2019 1 commit
- Enable graph rewrite for resnet 50 ctl · 8fe2729e
  Kaixi Hou authored Sep 04, 2019
```
Add checks on if use_tf_function is true

benchmarks for the ctl with AMP
```
  8fe2729e
22 Aug, 2019 2 commits
- Enable single l2 loss op in Resnet50 CTL 1gpu eager benchmark · 211bf1c0
  Zongwei Zhou authored Aug 22, 2019
```
PiperOrigin-RevId: 264937518
```
  211bf1c0
- Internal change · 9af989ce
  Hongkun Yu authored Aug 22, 2019
```
PiperOrigin-RevId: 264895439
```
  9af989ce
02 Aug, 2019 1 commit

Merged commit includes the following changes: (#7368) · 23c0017f

Haoyu Zhang authored Aug 02, 2019

* Merged commit includes the following changes:
261380794  by haoyuzhang<haoyuzhang@google.com>:

    Internal change

261374439  by haoyuzhang<haoyuzhang@google.com>:

    Change Keras CTL dependencies from r1 ResNet to Keras ResNet.

--

PiperOrigin-RevId: 261380794

* Revert unintentional change

* Revert unintentional change

23c0017f

25 Jul, 2019 1 commit
- Add Resnet50 CTL benchmark (pure eager w/ distribution strategy) · b59420dd
  Zongwei Zhou authored Jul 25, 2019
  
  b59420dd
19 Jun, 2019 2 commits

Add flags info when reporting benchmarks (#7056) · 1e527fb5

anj-s authored Jun 19, 2019

* first version of ctl

* fix indent

* remove monkey patching for core

* add dtype arg

* fix dtype arg

* add logging lib

* remove compat.v1.logging

* add datetime import

* fix FLAGS import

* add constant vals

* move to using as tf import

* move to using as tf import

* remove steps per epoch = 1

* test train and test for one step

* test train and test for one step

* test train and test for one step

* test train and test for the entire dataset

* use an iterator for test

* pass tensors instead of an iterator

* add stats dict

* fix list declaration

* fix list declaration

* fix elapsed time calc

* print lr at epoch boundary alone

* Use regular tf import instead of compat

* remove tensorboard chkpts

* add correct logging import

* add correct logging import

* add benchmark configs

* add tests and configs

* add tests and configs

* add keras flags import

* add keras flags import

* fix eval ds creation cond

* return numpy value of train_loss

* return numpy value of loss and acc values

* add option for full eager mode

* fix lint errors

* add ctl flags

* add ctl import

* add the xla flag

* enable v2 behavior in unit tests

* rename dataset var

* add synthetic dataset without monkey patching

* add ctl local constants

* add ctl local constants

* change to using v2 imports

* change to using v2 imports

* change to using v2 imports

* change to using keras synthetic input fn

* remove enable_eager flag from benchmarks

* remove enable_eager flag from benchmarks

* remove enable_eager flag from benchmarks

* add option for no distrat

* add lambda for flags

* remove no_func benchmarks due to OOM error

* remove README

* remove unused comments

* remove unchanged file

* remove unchanged file

* remove unused drop_remainder_arg

* use keras.common lr function

* address PR comments

* remove reference to deleted file

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* fix lint errors

* .

* add flags info

1e527fb5

Add benchmarks for custom training loops + tf.distribute (#6980) · 65636099

anj-s authored Jun 19, 2019

* first version of ctl

* fix indent

* remove monkey patching for core

* add dtype arg

* fix dtype arg

* add logging lib

* remove compat.v1.logging

* add datetime import

* fix FLAGS import

* add constant vals

* move to using as tf import

* move to using as tf import

* remove steps per epoch = 1

* test train and test for one step

* test train and test for one step

* test train and test for one step

* test train and test for the entire dataset

* use an iterator for test

* pass tensors instead of an iterator

* add stats dict

* fix list declaration

* fix list declaration

* fix elapsed time calc

* print lr at epoch boundary alone

* Use regular tf import instead of compat

* remove tensorboard chkpts

* add correct logging import

* add correct logging import

* add benchmark configs

* add tests and configs

* add tests and configs

* add keras flags import

* add keras flags import

* fix eval ds creation cond

* return numpy value of train_loss

* return numpy value of loss and acc values

* add option for full eager mode

* fix lint errors

* add ctl flags

* add ctl import

* add the xla flag

* enable v2 behavior in unit tests

* rename dataset var

* add synthetic dataset without monkey patching

* add ctl local constants

* add ctl local constants

* change to using v2 imports

* change to using v2 imports

* change to using v2 imports

* change to using keras synthetic input fn

* remove enable_eager flag from benchmarks

* remove enable_eager flag from benchmarks

* remove enable_eager flag from benchmarks

* add option for no distrat

* add lambda for flags

* remove no_func benchmarks due to OOM error

* remove README

* remove unused comments

* remove unchanged file

* remove unchanged file

* remove unused drop_remainder_arg

* use keras.common lr function

* address PR comments

* remove reference to deleted file

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* .

* fix lint errors

* .

65636099