search

wanmeihuali / taichi_3d_gaussian_splatting

An unofficial implementation of paper 3D Gaussian Splatting for Real-Time Radiance Field Rendering by taichi lang.
Apache License 2.0
648 stars 61 forks source link

do not use shared memory for aggregation. #106

Closed wanmeihuali closed 1 year ago

github-actions[bot] commented 1 year ago

Running experiment on sagemaker with git sha 00c2e2658d321856938a0a93ea4388cda98b71f6

github-actions[bot] commented 1 year ago

Training job backward-no-grad-tile-00c2e26-230717-044916-tat-truck-baseline created

github-actions[bot] commented 1 year ago

Running experiment on sagemaker with git sha 00c2e2658d321856938a0a93ea4388cda98b71f6

github-actions[bot] commented 1 year ago

Training job backward-no-grad-tile-00c2e26-230717-044951-tat-train-baseline created

github-actions[bot] commented 1 year ago

Training job backward-no-grad-tile-00c2e26-230717-044916-tat-truck-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/backward-no-grad-tile-00c2e26-230717-044916-tat-truck-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/backward-no-grad-tile-00c2e26-230717-044916-tat-truck-baseline/output/output.tar.gz

github-actions[bot] commented 1 year ago

Training job backward-no-grad-tile-00c2e26-230717-044916-tat-truck-baseline final metrics:

Latest Metrics

train:iteration train:l1loss train:loss train:num_valid_points train:psnr train:ssim train:ssimloss val:loss val:psnr val:ssim
30000.0 0.036039840430021286 0.054310142993927 427474.0 24.610546112060547 0.8726086616516113 0.12739133834838867 0.05396288260817528 25.12183952331543 0.8649366497993469

Max Metrics

train:5kpsnr train:5kssim train:7kpsnr train:7kssim train:psnr train:ssim val:5kpsnr val:5kssim val:7kpsnr val:7kssim val:psnr val:ssim
23.29336166381836 0.8360162973403931 23.141441345214844 0.837125301361084 24.610546112060547 0.8726086616516113 22.97919464111328 0.8118244409561157 23.810993194580078 0.8342387080192566 25.12183952331543 0.8649366497993469
github-actions[bot] commented 1 year ago

Running experiment on sagemaker with git sha 809eeaa3117d1f86dc5dd70190a03abaeee3240a

github-actions[bot] commented 1 year ago

Training job backward-no-grad-tile-809eeaa-230717-064552-tat-train-baseline created

github-actions[bot] commented 1 year ago

Running experiment on sagemaker with git sha 809eeaa3117d1f86dc5dd70190a03abaeee3240a

github-actions[bot] commented 1 year ago

Training job backward-no-grad-tile-809eeaa-230717-064601-tat-truck-baseline created

github-actions[bot] commented 1 year ago

Training job backward-no-grad-tile-00c2e26-230717-044951-tat-train-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/backward-no-grad-tile-00c2e26-230717-044951-tat-train-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/backward-no-grad-tile-00c2e26-230717-044951-tat-train-baseline/output/output.tar.gz

github-actions[bot] commented 1 year ago

Training job backward-no-grad-tile-00c2e26-230717-044951-tat-train-baseline final metrics:

Latest Metrics

train:iteration train:l1loss train:loss train:num_valid_points train:psnr train:ssim train:ssimloss val:loss val:psnr val:ssim
30000.0 0.03747538477182388 0.05534902215003967 642307.0 24.239452362060547 0.8731564283370972 0.12684357166290283 0.10029135644435883 20.879302978515625 0.8007410168647766

Max Metrics

train:5kpsnr train:5kssim train:7kpsnr train:7kssim train:psnr train:ssim val:5kpsnr val:5kssim val:7kpsnr val:7kssim val:psnr val:ssim
21.408409118652344 0.7158409357070923 21.83965492248535 0.7794806957244873 24.239452362060547 0.8731564283370972 19.197694778442383 0.7316948771476746 19.684951782226562 0.7565571665763855 20.879302978515625 0.8007410168647766
github-actions[bot] commented 1 year ago

Training job backward-no-grad-tile-809eeaa-230717-064601-tat-truck-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/backward-no-grad-tile-809eeaa-230717-064601-tat-truck-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/backward-no-grad-tile-809eeaa-230717-064601-tat-truck-baseline/output/output.tar.gz

github-actions[bot] commented 1 year ago

Training job backward-no-grad-tile-809eeaa-230717-064601-tat-truck-baseline final metrics:

Latest Metrics

train:iteration train:l1loss train:loss train:num_valid_points train:psnr train:ssim train:ssimloss val:loss val:psnr val:ssim
30000.0 0.02405722625553608 0.03889888525009155 442177.0 28.250160217285156 0.901734471321106 0.09826552867889404 0.054630592465400696 25.067495346069336 0.8641621470451355

Max Metrics

train:5kpsnr train:5kssim train:7kpsnr train:7kssim train:psnr train:ssim val:5kpsnr val:5kssim val:7kpsnr val:7kssim val:psnr val:ssim
24.186708450317383 0.826861560344696 23.955533981323242 0.8279774188995361 28.250160217285156 0.901734471321106 23.105283737182617 0.8156219124794006 23.76471710205078 0.8365946412086487 25.067495346069336 0.8641621470451355
github-actions[bot] commented 1 year ago

Training job backward-no-grad-tile-809eeaa-230717-064552-tat-train-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/backward-no-grad-tile-809eeaa-230717-064552-tat-train-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/backward-no-grad-tile-809eeaa-230717-064552-tat-train-baseline/output/output.tar.gz

github-actions[bot] commented 1 year ago

Training job backward-no-grad-tile-809eeaa-230717-064552-tat-train-baseline final metrics:

Latest Metrics

train:iteration train:l1loss train:loss train:num_valid_points train:psnr train:ssim train:ssimloss val:loss val:psnr val:ssim
30000.0 0.033265840262174606 0.04680868238210678 668672.0 25.117252349853516 0.8990199565887451 0.10098004341125488 0.09787896275520325 21.07177734375 0.8040274977684021

Max Metrics

train:5kpsnr train:5kssim train:7kpsnr train:7kssim train:psnr train:ssim val:5kpsnr val:5kssim val:7kpsnr val:7kssim val:psnr val:ssim
20.064132690429688 0.7424206137657166 20.371191024780273 0.7225081920623779 25.117252349853516 0.8990199565887451 19.17981719970703 0.7327414155006409 19.66835594177246 0.7553011775016785 21.07177734375 0.8040274977684021
github-actions[bot] commented 1 year ago

Running experiment on sagemaker with git sha 10e704d1838c380a024942f98a937574feb74634

github-actions[bot] commented 1 year ago

Training job backward-no-grad-tile-10e704d-230718-030927-tat-train-baseline created

github-actions[bot] commented 1 year ago

Running experiment on sagemaker with git sha 10e704d1838c380a024942f98a937574feb74634

github-actions[bot] commented 1 year ago

Training job backward-no-grad-tile-10e704d-230718-030934-tat-truck-baseline created

github-actions[bot] commented 1 year ago

Training job backward-no-grad-tile-10e704d-230718-030934-tat-truck-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/backward-no-grad-tile-10e704d-230718-030934-tat-truck-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/backward-no-grad-tile-10e704d-230718-030934-tat-truck-baseline/output/output.tar.gz

github-actions[bot] commented 1 year ago

Training job backward-no-grad-tile-10e704d-230718-030934-tat-truck-baseline final metrics:

Latest Metrics

train:iteration train:l1loss train:loss train:num_valid_points train:psnr train:ssim train:ssimloss val:loss val:psnr val:ssim
30000.0 0.02999674342572689 0.048221103847026825 444644.0 25.360095977783203 0.8788814544677734 0.12111854553222656 0.05415721982717514 25.098535537719727 0.8658057451248169

Max Metrics

train:5kpsnr train:5kssim train:7kpsnr train:7kssim train:psnr train:ssim val:5kpsnr val:5kssim val:7kpsnr val:7kssim val:psnr val:ssim
23.065759658813477 0.8152251243591309 24.677345275878906 0.8701910376548767 25.360095977783203 0.8788814544677734 23.136329650878906 0.8158812522888184 23.86459732055664 0.8363640904426575 25.098535537719727 0.8658057451248169
github-actions[bot] commented 1 year ago

Training job backward-no-grad-tile-10e704d-230718-030927-tat-train-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/backward-no-grad-tile-10e704d-230718-030927-tat-train-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/backward-no-grad-tile-10e704d-230718-030927-tat-train-baseline/output/output.tar.gz

github-actions[bot] commented 1 year ago

Training job backward-no-grad-tile-10e704d-230718-030927-tat-train-baseline final metrics:

Latest Metrics

train:iteration train:l1loss train:loss train:num_valid_points train:psnr train:ssim train:ssimloss val:loss val:psnr val:ssim
30000.0 0.03943511098623276 0.06264723092317581 678083.0 24.2143611907959 0.8445042967796326 0.15549570322036743 0.10062317550182343 20.964387893676758 0.7995827198028564

Max Metrics

train:5kpsnr train:5kssim train:7kpsnr train:7kssim train:psnr train:ssim val:5kpsnr val:5kssim val:7kpsnr val:7kssim val:psnr val:ssim
23.097301483154297 0.844767689704895 24.095457077026367 0.8336064219474792 24.2143611907959 0.8445042967796326 19.094995498657227 0.7282618284225464 19.77592658996582 0.7559015154838562 20.964387893676758 0.7995827198028564