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
662 stars 62 forks source link

Update test_sagemaker.yaml #78

Closed wanmeihuali closed 1 year ago

github-actions[bot] commented 1 year ago

Running experiment on sagemaker with git sha fccb8e469c12020d86e93dee120f4ea7cff35189

github-actions[bot] commented 1 year ago

Training job densify-policy-7-fccb8e4-230701-215228-tat-truck created

github-actions[bot] commented 1 year ago

Training job densify-policy-7-fccb8e4-230701-215228-tat-truck completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-truck/densify-policy-7-fccb8e4-230701-215228-tat-truck/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-truck/densify-policy-7-fccb8e4-230701-215228-tat-truck/output/output.tar.gz

github-actions[bot] commented 1 year ago

Training job densify-policy-7-fccb8e4-230701-215228-tat-truck final metrics:

Latest Metrics

val:ssim train:loss train:iteration train:num_valid_points val:psnr train:ssimloss train:l1loss train:psnr val:loss train:ssim
0.8390111327171326 0.0610487125813961 30000.0 436897.0 24.441465377807617 0.1783214807510376 0.031730517745018005 24.428129196166992 0.06082937866449356 0.8216785192489624

Max Metrics

val:ssim val:7kpsnr val:5kssim train:5kpsnr train:7kssim val:7kssim val:psnr train:7kpsnr train:5kssim train:psnr val:5kpsnr train:ssim
0.8390111327171326 22.98468780517578 0.77360600233078 21.705167770385742 0.8238461017608643 0.7993574142456055 24.441465377807617 23.286149978637695 0.7435254454612732 24.428129196166992 22.183349609375 0.8216785192489624
github-actions[bot] commented 1 year ago

Running experiment on sagemaker with git sha 017ba1f2a3b552de9dcb30d9ca0694b97933160d

github-actions[bot] commented 1 year ago

Training job densify-policy-7-017ba1f-230702-012821-tat-truck created

github-actions[bot] commented 1 year ago

Training job densify-policy-7-017ba1f-230702-012821-tat-truck completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-truck/densify-policy-7-017ba1f-230702-012821-tat-truck/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-truck/densify-policy-7-017ba1f-230702-012821-tat-truck/output/output.tar.gz

github-actions[bot] commented 1 year ago

Training job densify-policy-7-017ba1f-230702-012821-tat-truck final metrics:

Latest Metrics

train:ssim train:loss train:num_valid_points train:iteration val:ssim val:loss train:psnr train:ssimloss train:l1loss val:psnr
0.8267943859100342 0.060944944620132446 436860.0 30000.0 0.840575098991394 0.06041005626320839 24.45001220703125 0.17320561408996582 0.032879773527383804 24.492000579833984

Max Metrics

train:ssim val:7kpsnr train:7kssim val:7kssim train:5kssim val:ssim train:psnr train:5kpsnr val:5kssim val:5kpsnr val:psnr train:7kpsnr
0.8267943859100342 22.945880889892578 0.8090329170227051 0.7986046671867371 0.773406982421875 0.840575098991394 24.45001220703125 22.444698333740234 0.7730562090873718 22.234243392944336 24.492000579833984 22.549219131469727