v0.0.13.post4

update benchmark results

v0.0.13.post3

bump version number to 0.0.13.post3

v0.0.13.post1

Fix T4 compatibility

v0.0.13

What's Changed

• fix aten::to RuntimeError by @chengzeyi in #50
• Dev by @chengzeyi in #51
• closes #53; fix parameters getting released with SDXL and CUDA graphs by @chengzeyi in #54
• Dev by @chengzeyi in #61

Full Changelog: v0.0.12...v0.0.13

v0.0.12.post6

bump version to 0.0.12.post6

v0.0.12.post5

fix

v0.0.12.post4

use tree_copy instead of copy.deepcopy to handle unpickable objects l…

v0.0.12.post3

bump version to 0.0.12.post3

v0.0.12.post2

Dev (#51)

* fix aten::to RuntimeError

* fix Windows build

* test tiny vae and cuda graphify vae

* add cudnn_convolution_bias_add_sigmoid, cudnn_convolution_bias_add_relu, cudnn_convolution_bias_add_tanh

* add TRT comparison

* fix

v0.0.12.post1

bump version to 0.0.12.post1