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 #79

Closed wanmeihuali closed 1 year ago

github-actions[bot] commented 1 year ago

Running experiment on sagemaker with git sha 36e75986d38a321a2407b30f765247a2a01dd83e

github-actions[bot] commented 1 year ago

Training job densify-policy-8-36e7598-230701-215635-tat-truck created

github-actions[bot] commented 1 year ago

Training job densify-policy-8-36e7598-230701-215635-tat-truck completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-truck/densify-policy-8-36e7598-230701-215635-tat-truck/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-truck/densify-policy-8-36e7598-230701-215635-tat-truck/output/output.tar.gz

github-actions[bot] commented 1 year ago

Training job densify-policy-8-36e7598-230701-215635-tat-truck final metrics:

Latest Metrics

val:psnr train:loss train:num_valid_points train:psnr train:ssimloss val:ssim val:loss train:ssim train:l1loss train:iteration
24.28446388244629 0.04545892775058746 603348.0 26.426097869873047 0.12288200855255127 0.8346765041351318 0.06213441491127014 0.8771179914474487 0.026103157550096512 30000.0

Max Metrics

val:7kssim val:psnr train:7kssim val:7kpsnr train:psnr val:5kssim train:7kpsnr val:ssim train:5kpsnr val:5kpsnr train:ssim train:5kssim
0.8031967878341675 24.28446388244629 0.7998181581497192 22.98328971862793 26.426097869873047 0.783088743686676 22.916439056396484 0.8346765041351318 22.46523094177246 22.391845703125 0.8771179914474487 0.7746769785881042
github-actions[bot] commented 1 year ago

Running experiment on sagemaker with git sha c1f797b1aa1d3e72fb0b5f5c3fec47ef06d3a3bf

github-actions[bot] commented 1 year ago

Training job densify-policy-8-c1f797b-230702-013101-tat-truck created

github-actions[bot] commented 1 year ago

Training job densify-policy-8-c1f797b-230702-013101-tat-truck completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-truck/densify-policy-8-c1f797b-230702-013101-tat-truck/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-truck/densify-policy-8-c1f797b-230702-013101-tat-truck/output/output.tar.gz

github-actions[bot] commented 1 year ago

Training job densify-policy-8-c1f797b-230702-013101-tat-truck final metrics:

Latest Metrics

train:ssim train:psnr val:psnr train:iteration train:loss train:l1loss train:num_valid_points val:loss train:ssimloss val:ssim
0.8575506210327148 25.07217788696289 24.438528060913086 30000.0 0.05411659926176071 0.03203340247273445 434860.0 0.06071334332227707 0.14244937896728516 0.8390138745307922

Max Metrics

val:7kssim train:ssim train:psnr val:psnr train:7kpsnr val:5kssim val:5kpsnr train:5kpsnr train:5kssim train:7kssim val:ssim val:7kpsnr
0.8013823628425598 0.8575506210327148 25.07217788696289 24.438528060913086 13.350324630737305 0.7749874591827393 22.235149383544922 22.22081756591797 0.79558265209198 0.6835116147994995 0.8390138745307922 23.023632049560547