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

Update test_sagemaker.yaml #80

Closed wanmeihuali closed 1 year ago

github-actions[bot] commented 1 year ago

Running experiment on sagemaker with git sha 87b86763a56ae71d6d21f050c73167e8b4a0a898

github-actions[bot] commented 1 year ago

Training job param-tune-reset-alpha-87b8676-230703-230159-tat-truck created

github-actions[bot] commented 1 year ago

Training job param-tune-reset-alpha-87b8676-230703-230159-tat-truck completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-truck/param-tune-reset-alpha-87b8676-230703-230159-tat-truck/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-truck/param-tune-reset-alpha-87b8676-230703-230159-tat-truck/output/output.tar.gz

github-actions[bot] commented 1 year ago

Training job param-tune-reset-alpha-87b8676-230703-230159-tat-truck final metrics:

Latest Metrics

train:ssimloss train:iteration train:ssim train:num_valid_points val:psnr train:l1loss train:loss val:loss val:ssim train:psnr
0.17281389236450195 30000.0 0.827186107635498 428173.0 22.958433151245117 0.033772438764572144 0.061580728739500046 0.07196833193302155 0.8160181045532227 24.408695220947266

Max Metrics

train:5kpsnr train:7kssim train:ssim val:5kpsnr val:7kssim train:7kpsnr val:psnr train:5kssim val:7kpsnr val:5kssim val:ssim train:psnr
23.631582260131836 0.8626880049705505 0.827186107635498 22.213031768798828 0.7958897352218628 24.561931610107422 22.958433151245117 0.7962536811828613 22.959135055541992 0.7735530138015747 0.8160181045532227 24.408695220947266
github-actions[bot] commented 1 year ago

Running experiment on sagemaker with git sha c069b1122c8002caf07b76fcdeb936ce283a0b01

github-actions[bot] commented 1 year ago

Training job param-tune-reset-alpha-c069b11-230704-062740-tat-truck created

github-actions[bot] commented 1 year ago

Training job param-tune-reset-alpha-c069b11-230704-062740-tat-truck completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-truck/param-tune-reset-alpha-c069b11-230704-062740-tat-truck/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-truck/param-tune-reset-alpha-c069b11-230704-062740-tat-truck/output/output.tar.gz

github-actions[bot] commented 1 year ago

Training job param-tune-reset-alpha-c069b11-230704-062740-tat-truck final metrics:

Latest Metrics

train:ssim val:loss val:ssim train:iteration train:psnr train:loss train:num_valid_points train:ssimloss train:l1loss val:psnr
0.8670576810836792 0.07155710458755493 0.8163155913352966 30000.0 25.803468704223633 0.04825032502412796 428857.0 0.1329423189163208 0.0270773284137249 22.997177124023438

Max Metrics

val:5kpsnr train:ssim train:5kssim val:ssim train:psnr train:7kssim val:7kssim val:5kssim train:5kpsnr train:7kpsnr val:psnr val:7kpsnr
22.333051681518555 0.8670576810836792 0.7833242416381836 0.8163155913352966 25.803468704223633 0.819358766078949 0.7981828451156616 0.7800294756889343 21.871200561523438 24.858718872070312 22.997177124023438 22.957637786865234