pytorch/benchmark

Pull Requests Commits

[optim] add new model hf_clip to DENSE_MODELS

janeyx99 committed 2 years ago

d9e87c76

Consistent accuracy results with dynamobench (#1941)

xuzhao9 committed 2 years ago

34805901

Delete the model after checking accuracy in-process

xuzhao9 committed 2 years ago

aaa8baf5

Add support for Torch-TensorRT Nightly Install in Docker (#1909)

gs-olive committed 2 years ago

160bcfe8

Fix PyTorch CI HUD dashboard missing perf numbers: hf_Whisper (#1935)

xmfan committed 2 years ago

3f1c3ebe

Checkout repo submodules recursively (#1937)

xuzhao9 committed 2 years ago

64409d57

hf and multimodal clip (#1921)

msaroufim committed 2 years ago

6fef32dd

Rename aot_runtime to aoti_runtime (#1930)

desertfire committed 2 years ago

b7e94047

torchbench is now a library (#1933)

msaroufim committed 2 years ago

f2d615b7

disable PT2 benchmarks until we investigate the noisy reports (#1932)

janeyx99 committed 2 years ago

75889b44

Train step for cm3leon (#1922)

msaroufim committed 2 years ago

4ea3bba3

Support llamav27b training (#1924)

msaroufim committed 2 years ago

64dec593

Instructions to use torchbench as a library (#1923)

msaroufim committed 2 years ago

8a4ba7f0

Add stable diffusion models to dense to skip on sparseadam (#1931)

janeyx99 committed 2 years ago

e1185ffb

Update README.md (#1920)

msaroufim committed 2 years ago

2bd1aa13

Add interface to support multiple batches (#1914)

xuzhao9 committed 2 years ago

842f11db

Add nanogpt_train to torchbench, and merge with nanogpt_generate into joint folder (#1911)

xmfan committed 2 years ago

d00b9421

Explain `test.py`'s `NotImplementedError`s (#1913)

hmellor committed 2 years ago

1f70eae8

Add warmup and benchamrk iterations for cpu userbenchmark (#1908)

chuanqi129 committed 2 years ago

2d05303a

Rename the `enable_fp16_half()` function to `enable_fp16` (#1900)

xuzhao9 committed 2 years ago

644a9ceb

Pin beartype for torchbench (#1907)

malfet committed 2 years ago

8360dba3

Fix ig_ctr model failure on AIBench

xuzhao9 committed 2 years ago

cccaacee

Stable Diffusion XL (#1882)

msaroufim committed 2 years ago

95207dfd

Skip the model test by the device name in its metadata. (#1906)

xuzhao9 committed 2 years ago

568f9475

Upgrade version of transformers (#1905)

xuzhao9 committed 2 years ago

8d159fda

`norm()` -> `linalg.norm()` in `lennard_jones` (#1898)

hmellor committed 2 years ago

22cb901a

Remove `.cuda()` from `cm3leon`'s `model.py` (#1897)

hmellor committed 2 years ago

6532c61f

Disable Deepcopy on stable_diffusion_unet. (#1895)

xuzhao9 committed 2 years ago

e1b51f52

Always load the model within the same process in sandcastle test.

xuzhao9 committed 2 years ago

0f33c9b7

Exclude timm_vision_transformer pt2, fix runtime errors in 1888 (#1890)

janeyx99 committed 2 years ago

fd78a760

Older