pytorch/benchmark

Pull Requests Commits

try dcgm standalone mode first

FindHao committed 3 years ago

1f8750e2

add workaround to dcgm embedded mode

FindHao committed 3 years ago

f448ed10

Enable E2E hf_t5 dynamo+inductor train/eval (#1323)

yanboliang committed 3 years ago

6d6e3204

Fix install.py and dependencies (#1333)

xuzhao9 committed 3 years ago

e5f3ae29

Add Bert_large (#1321)

davidberard98 committed 3 years ago

6b127f1c

Add FSDP to ddp/dynamo benchmarks (#1318)

davidberard98 committed 3 years ago

f40c66b2

Fix dependencies for the distributed benchmark (#1327)

xuzhao9 committed 3 years ago

cf15b127

Increase the default batch size for cmf_10x, ads_dhen_5x, and blue_reels_vdd_v2 models

xuzhao9 committed 3 years ago

6cd28b51

Add cuda 116-117 comparison (#1307)

xuzhao9 committed 3 years ago

795a5d4e

README.md fix (#1316)

Yanbo Zhao committed 3 years ago

ab3ce83d

Add --amp option to torchbench run.py (#1317)

jansel committed 3 years ago

c1385f20

Fix hf_t5 with run_e2e.py (#1313)

xuzhao9 committed 3 years ago

2e5d7237

Add slurm_exclude arg to distributed userbenchmarks (#1309)

davidberard98 committed 3 years ago

269c6ad3

fix typo for model.get_flops (#1312)

FindHao committed 3 years ago

ba8cfca2

Enable E2E hf_bert dynamo+inductor train/eval (#1302)

yanboliang committed 3 years ago

f74244e7

reorganize arguments (#1299)

FindHao committed 3 years ago

272a2457

Dynamo option to disable() optimizer.step (#1308)

davidberard98 committed 3 years ago

4cb79557

Add --profile-options to support fine toggling of profile options (#1306)

aaronenyeshi committed 3 years ago

9b1ac22b

update model list and batch sizes (#1304)

davidberard98 committed 3 years ago

8ff0ea1f

Add distributed correctness checks (#1294)

davidberard98 committed 3 years ago

dc955607

Support setting up extra environments in ModelTask

xuzhao9 committed 3 years ago

38196605

set triton.cudagraphs only in inductor backend (#1298)

yuguo68 committed 3 years ago

96644b4e

add nvml monitor (#1288)

FindHao committed 3 years ago

e59ee218

Handle run.py model_analyzer is none (#1297)

aaronenyeshi committed 3 years ago

7ad47ee0

Fix a bug in run.py related to model_analyzer

xuzhao9 committed 3 years ago

7b932305

Align run.py profile_one_step and run_one_step (#1296)

aaronenyeshi committed 3 years ago

022dfe3d

Only check the internal userbenchmarks when the directory exists (#1290)

xuzhao9 committed 3 years ago

f28c72d8

Fix e2e hf_bert model (#1291)

xuzhao9 committed 3 years ago

b35e3d86

Suppress `pretrained` warnings from torchvision (#1280)

davidberard98 committed 3 years ago

2bb181ce

Add resnet152 (#1279)

davidberard98 committed 3 years ago

1168045b

Older