onnxruntime
Add Benchmark Script for Bert Models
#3829
Merged

Add Benchmark Script for Bert Models #3829

tianleiwu merged 20 commits into master from tlwu/benchmark_bert
tianleiwu
tianleiwu Add benchmark script for BERT models
c3e33aaf
tianleiwu Add license
37a04931
tianleiwu tianleiwu requested a review from liuziyue liuziyue 5 years ago
tianleiwu tianleiwu requested a review from yufenglee yufenglee 5 years ago
tianleiwu tianleiwu requested a review 5 years ago
tianleiwu tianleiwu added model:transformer
liuziyue
liuziyue commented on 2020-05-06
tianleiwu Set intra_op_num_threads=1 for cpu (version <= 1.2.0)
add1de35
tianleiwu Use comma instead of space to split models string
66e2ef7b
tianleiwu format
ad2909cf
tianleiwu Add percentiles for latency
10dd103d
tianleiwu Constraint models to predefined list.
44426c1a
tianleiwu torch.set_num_threads (for intra op) to get fair comparison
4fef83e9
tianleiwu Allow export ONNX model with specified number of inputs
70b7093c
tianleiwu Add warning that--num_inputs is not implemented for torch/torchscript
02d69397
tianleiwu change default test_times to 1000
31144a98
tianleiwu Add --optimize_onnx and --num_inputs options
514d592e
tianleiwu rename file
bf8a3f24
tianleiwu Rename file to benchmark.py
e53f6b60
tianleiwu Output a summary file
cbbe3a06
tianleiwu update comments
e9d9d945
tianleiwu Add xlnet-base-cased
936aeb2c
tianleiwu Add openai-gpt
30dc0945
tianleiwu Update default batch_sizes and sequence_lengths
7c73bb96
tianleiwu Install transformers from source
d9586448
liuziyue
liuziyue approved these changes on 2020-05-12
tianleiwu tianleiwu merged 29496170 into master 5 years ago
tianleiwu tianleiwu deleted the tlwu/benchmark_bert branch 5 years ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone