Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

TRT support for MAISI #701

Open
wants to merge 16 commits into
base: dev
Choose a base branch
from
Open

TRT support for MAISI #701

Show file tree
Hide file tree
Changes from 14 commits
Commits
Show all changes
16 commits
Select commit Hold shift + click to select a range
93a4dc5
Added trt_compile configs for vista2d and vista3d
borisfom Aug 28, 2024
c401a63
Merge branch 'dev' of github.com:Project-MONAI/model-zoo into vista_trt
borisfom Sep 4, 2024
89dfac1
Stash
borisfom Sep 6, 2024
c1ab420
Merge branch 'dev' into maisi-trt
borisfom Oct 9, 2024
027f71b
Merge remote-tracking branch 'origin/dev' into maisi-trt
borisfom Oct 9, 2024
b6628fc
Working MAISI
borisfom Oct 12, 2024
860b932
Merge remote-tracking branch 'origin/dev' into maisi-trt
borisfom Oct 16, 2024
e91ad57
Adding TRT support
borisfom Oct 16, 2024
b336e06
cleanup
borisfom Oct 16, 2024
97d2b0e
fixing condition
borisfom Oct 16, 2024
60000df
Added output_lists option
borisfom Oct 21, 2024
ebb50bc
update pre commit config
yiheng-wang-nv Oct 22, 2024
74a6ecc
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Oct 22, 2024
a9a8b03
update metadata
yiheng-wang-nv Oct 22, 2024
e827f5c
Addressing code review comments
borisfom Oct 22, 2024
76f8fb6
Merge remote-tracking branch 'origin/dev' into maisi-trt
borisfom Oct 30, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion .pre-commit-config.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,5 @@
default_language_version:
python: python3.8
python: python3

ci:
autofix_prs: true
Expand Down
4 changes: 2 additions & 2 deletions models/maisi_ct_generative/configs/inference.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -34,8 +34,8 @@
"spatial_dims": 3,
"image_channels": 1,
"latent_channels": 4,
"output_size_xy": 512,
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hi @borisfom , could you revert this size change? The default values are set by MAISI author

"output_size_z": 512,
"output_size_xy": 256,
"output_size_z": 256,
"output_size": [
"@output_size_xy",
"@output_size_xy",
Expand Down
17 changes: 17 additions & 0 deletions models/maisi_ct_generative/configs/inference_trt.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,17 @@
{
"+imports": [
"$from monai.networks import trt_compile"
],
"c_trt_args": {
"output_lists": [
[
-1
],
[
0
]
]
},
"controlnet": "$trt_compile(@controlnet_def.to(@device), @trained_controlnet_path, args=@c_trt_args)",
"diffusion_unet": "$trt_compile(@diffusion_unet_def.to(@device), @trained_diffusion_path)"
}
3 changes: 2 additions & 1 deletion models/maisi_ct_generative/configs/metadata.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,8 @@
{
"schema": "https://github.com/Project-MONAI/MONAI-extra-test-data/releases/download/0.8.1/meta_schema_generator_ldm_20240318.json",
"version": "0.4.5",
"version": "0.4.6",
"changelog": {
"0.4.6": "add TensorRT support",
"0.4.5": "update README",
"0.4.4": "update issue for IgniteInfo",
"0.4.3": "remove download large files, add weights_only when loading weights and add label_dict to large files",
Expand Down
1 change: 0 additions & 1 deletion models/maisi_ct_generative/scripts/sample.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -241,7 +241,6 @@ def ldm_conditional_sample_one_image(
# synthesize latents
noise_scheduler.set_timesteps(num_inference_steps=num_inference_steps)
for t in tqdm(noise_scheduler.timesteps, ncols=110):
# Get controlnet output
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

same here, the comment may be useful, could you revert the changes? Thanks!

down_block_res_samples, mid_block_res_sample = controlnet(
x=latents, timesteps=torch.Tensor((t,)).to(device), controlnet_cond=controlnet_cond_vis
)
Expand Down
Loading