issues
search
facebookresearch
/
FAMBench
Benchmarks to capture important workloads.
Apache License 2.0
28
stars
23
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
BEVT Submodule udpate to support numpy v1.20 and above
#127
itej89
closed
1 month ago
1
Update submodule benchmarks/bevt/ootb/BEVT to new remote URL 'https://github.com/itej89/BEVT.git'
#126
itej89
closed
1 month ago
1
adapt scipy 1.13
#125
liligwu
closed
3 months ago
0
Update oscar
#124
pnunna93
closed
11 months ago
0
Pointing to pinned mmcv version
#123
zstreet87
closed
10 months ago
6
CvT update for torchrun
#122
pnunna93
closed
11 months ago
0
Update Oscar module
#121
pnunna93
closed
11 months ago
0
Revert num-workers patch for rnnt
#120
pnunna93
closed
1 year ago
0
Fix Oscar distributed launch
#119
amathews-amd
closed
1 year ago
0
Change torch.distributed.launch to torchrun for RNNT
#118
pnunna93
closed
1 year ago
0
Set num workers to 0 for rnnt single gpu run
#117
pnunna93
closed
1 year ago
3
updated to torchrun
#116
zstreet87
closed
1 year ago
3
torch.distribute.launch -> torchrun
#115
zstreet87
opened
1 year ago
0
Oscar data is no longer available
#114
amathews-amd
opened
1 year ago
2
Fix werkzeug security issue
#113
xuzhao9
closed
1 year ago
0
fix dlrm/ubench/dlrm_ubench_train_embeddingbag_driver.py
#112
liligwu
closed
1 year ago
0
DLRM ubench failing, possibly due to FBGEMM API update
#111
amathews-amd
closed
1 year ago
1
Disable SCALE_LR in CvT
#110
jataylo
closed
1 year ago
1
Update LR default value in run_cvt_train.sh
#109
jataylo
closed
1 year ago
2
BEVT run script update
#108
zstreet87
closed
1 year ago
0
Update CuDNN Multihead Attention Benchmark
#107
samiwilf
closed
1 year ago
0
Remove redundant logger
#106
samiwilf
closed
1 year ago
0
removed cherry-picks in bevt Dockerfile.rocm
#105
zstreet87
closed
1 year ago
1
Update CvT, MoE, Oscar, and BEVT scripts
#104
zstreet87
closed
1 year ago
0
Fix MOE bash vars EP_WORLD_SIZE & NUM_EXPERTS for under 8 GPUs
#103
samiwilf
closed
1 year ago
0
Add MOE & Oscar scripts to /benchmarks. Update docker commands.
#102
samiwilf
closed
1 year ago
0
update bevt example docker command
#101
samiwilf
closed
1 year ago
0
Updated XLRM with --use-torchtext parameter to choose fairseq or torc…
#100
nrsatish
closed
1 year ago
0
README changes to reflect model changes
#99
nrsatish
closed
1 year ago
0
Bump joblib from 1.1.0 to 1.2.0 in /benchmarks/rnnt/ootb/inference
#98
dependabot[bot]
closed
1 year ago
0
Bump protobuf from 3.20.0 to 3.20.2 in /benchmarks/rnnt/ootb/inference
#97
dependabot[bot]
closed
1 year ago
1
Initial commit for torchtext and use of BetterTransfomer
#96
nrsatish
closed
1 year ago
0
Change memory to pinned
#95
nrsatish
closed
1 year ago
0
Change rnnt's score to be the aggregate of all used GPUs.
#94
samiwilf
closed
1 year ago
0
Add benchmarks for CvT/BEVT/MoE/Oscar
#93
mindest
closed
1 year ago
9
RNN-T Default Updates
#92
jpvillam-amd
closed
1 year ago
1
[ROCm] ROCm Implementation for RNN-T Train
#91
jpvillam-amd
closed
1 year ago
1
XLM-R failing with ValueError: invalid literal for int() with base 10: '0a0'
#90
amathews-amd
opened
2 years ago
2
Update rnnt
#89
samiwilf
closed
1 year ago
1
Add TF32 flag to XLRM
#88
nrsatish
closed
2 years ago
0
TF32 changes
#87
nrsatish
closed
2 years ago
0
XLM-R failing 'omegaconf._utils' has no attribute 'is_primitive_type'
#86
amathews-amd
closed
2 years ago
5
rnnt setup failed bacause of "THC.h" not found
#85
xuzhao9
opened
2 years ago
2
Assert is triggered when execute './run_all.sh'
#84
guoshzhao
opened
2 years ago
0
Added batch size for TRT
#83
nrsatish
closed
2 years ago
0
Add multi-node compatibility to dlrm comms ubenches.
#82
samiwilf
closed
1 year ago
0
Fix dlrm distributed_forward data corruption check.
#81
samiwilf
closed
1 year ago
1
Make CvT and OSCAR WIP
#80
erichan1
closed
2 years ago
1
Add projection (--arch-project-size CLI option). pyfmt dlrm_s_pytorch…
#79
samiwilf
closed
2 years ago
0
Add CvT convolutions ubench.
#78
samiwilf
closed
2 years ago
0
Next