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 #76

Closed wanmeihuali closed 1 year ago

github-actions[bot] commented 1 year ago

Running experiment on sagemaker with git sha 6494947b6411ab30297e47aa66bec7768b998135

github-actions[bot] commented 1 year ago

Training job densify-policy-5-6494947-230701-215308-tat-truck created

github-actions[bot] commented 1 year ago

Training job densify-policy-5-6494947-230701-215308-tat-truck completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-truck/densify-policy-5-6494947-230701-215308-tat-truck/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-truck/densify-policy-5-6494947-230701-215308-tat-truck/output/output.tar.gz

github-actions[bot] commented 1 year ago

Training job densify-policy-5-6494947-230701-215308-tat-truck final metrics:

Latest Metrics

train:num_valid_points train:l1loss train:ssim train:loss train:psnr val:loss val:psnr train:iteration train:ssimloss val:ssim
195749.0 0.03322650492191315 0.8143660426139832 0.06370799243450165 25.14641761779785 0.06642979383468628 23.99769401550293 30000.0 0.18563395738601685 0.81866455078125

Max Metrics

val:5kssim train:5kpsnr train:ssim train:5kssim val:ssim train:7kssim val:7kpsnr val:psnr val:5kpsnr train:7kpsnr val:7kssim train:psnr
0.7713650465011597 21.866548538208008 0.8143660426139832 0.7596760988235474 0.81866455078125 0.8172940611839294 23.025487899780273 23.99769401550293 22.21259117126465 24.421344757080078 0.7957154512405396 25.14641761779785
github-actions[bot] commented 1 year ago

Running experiment on sagemaker with git sha ffdbaf0fc6ba9a541eb652efef32a141c47153a0

github-actions[bot] commented 1 year ago

Training job densify-policy-5-ffdbaf0-230701-234530-tat-truck created

github-actions[bot] commented 1 year ago

Training job densify-policy-5-ffdbaf0-230701-234530-tat-truck completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-truck/densify-policy-5-ffdbaf0-230701-234530-tat-truck/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-truck/densify-policy-5-ffdbaf0-230701-234530-tat-truck/output/output.tar.gz

github-actions[bot] commented 1 year ago

Training job densify-policy-5-ffdbaf0-230701-234530-tat-truck final metrics:

Latest Metrics

train:ssimloss train:iteration val:loss train:loss train:l1loss train:num_valid_points train:ssim train:psnr val:psnr val:ssim
0.12951147556304932 30000.0 0.05932069942355156 0.04826009273529053 0.02794724889099598 860610.0 0.8704885244369507 26.812551498413086 24.520112991333008 0.8428623676300049

Max Metrics

val:ssim val:7kssim train:ssim train:7kssim train:psnr train:5kssim val:5kssim val:psnr train:5kpsnr val:7kpsnr val:5kpsnr train:7kpsnr
0.8428623676300049 0.8109146952629089 0.8704885244369507 0.8355647325515747 26.812551498413086 0.8434890508651733 0.7853209972381592 24.520112991333008 24.209484100341797 23.14113426208496 22.35540771484375 24.280298233032227