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

optimize backward pass. #105

Closed wanmeihuali closed 1 year ago

github-actions[bot] commented 1 year ago

Running experiment on sagemaker with git sha 541b678da62d40d68a10d46dcda74fc6cf4597af

github-actions[bot] commented 1 year ago

Training job backward-optimization-541b678-230716-215942-tat-truck-baseline created

github-actions[bot] commented 1 year ago

Running experiment on sagemaker with git sha 541b678da62d40d68a10d46dcda74fc6cf4597af

github-actions[bot] commented 1 year ago

Training job backward-optimization-541b678-230716-215945-tat-train-baseline created

wanmeihuali commented 1 year ago

official implementation: image this version: image

github-actions[bot] commented 1 year ago

Training job backward-optimization-541b678-230716-215942-tat-truck-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/backward-optimization-541b678-230716-215942-tat-truck-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/backward-optimization-541b678-230716-215942-tat-truck-baseline/output/output.tar.gz

github-actions[bot] commented 1 year ago

Training job backward-optimization-541b678-230716-215942-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 0.03122282400727272 0.05385543406009674 425831.0 25.504825592041016 0.8556141257286072 0.14438587427139282 0.05396534502506256 25.134395599365234 0.8645569682121277

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
25.937057495117188 0.8579326868057251 23.658187866210938 0.8426594138145447 25.504825592041016 0.8556141257286072 23.0472354888916 0.8143715262413025 23.705787658691406 0.8345955014228821 25.134395599365234 0.8645569682121277
github-actions[bot] commented 1 year ago

Training job backward-optimization-541b678-230716-215945-tat-train-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/backward-optimization-541b678-230716-215945-tat-train-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/backward-optimization-541b678-230716-215945-tat-train-baseline/output/output.tar.gz

github-actions[bot] commented 1 year ago

Training job backward-optimization-541b678-230716-215945-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 0.0451316311955452 0.06997349858283997 653766.0 22.946575164794922 0.830659031867981 0.16934096813201904 0.1010923981666565 20.847064971923828 0.8003285527229309

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
24.149375915527344 0.8617434501647949 19.781614303588867 0.7150112390518188 22.946575164794922 0.830659031867981 19.221424102783203 0.7304475903511047 19.68915367126465 0.7556347846984863 20.847064971923828 0.8003285527229309
github-actions[bot] commented 1 year ago

Running experiment on sagemaker with git sha 966e970023f6494b96686f796cefa7d3739d1ef0

github-actions[bot] commented 1 year ago

Training job backward-optimization-966e970-230717-034945-tat-train-baseline created

github-actions[bot] commented 1 year ago

Running experiment on sagemaker with git sha 966e970023f6494b96686f796cefa7d3739d1ef0

github-actions[bot] commented 1 year ago

Training job backward-optimization-966e970-230717-035001-tat-truck-baseline created

github-actions[bot] commented 1 year ago

Training job backward-optimization-966e970-230717-035001-tat-truck-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/backward-optimization-966e970-230717-035001-tat-truck-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/backward-optimization-966e970-230717-035001-tat-truck-baseline/output/output.tar.gz

github-actions[bot] commented 1 year ago

Training job backward-optimization-966e970-230717-035001-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 0.03276047855615616 0.050863251090049744 424969.0 25.101505279541016 0.8767256736755371 0.12327432632446289 0.05392814427614212 25.156055450439453 0.8647201657295227

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
23.502025604248047 0.8038249015808105 25.385364532470703 0.8695834875106812 25.101505279541016 0.8767256736755371 23.0035400390625 0.8112714290618896 23.75076675415039 0.834479570388794 25.156055450439453 0.8647201657295227
github-actions[bot] commented 1 year ago

Training job backward-optimization-966e970-230717-034945-tat-train-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/backward-optimization-966e970-230717-034945-tat-train-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/backward-optimization-966e970-230717-034945-tat-train-baseline/output/output.tar.gz

github-actions[bot] commented 1 year ago

Training job backward-optimization-966e970-230717-034945-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 0.030602410435676575 0.044951021671295166 650378.0 26.633710861206055 0.8976545333862305 0.10234546661376953 0.10058249533176422 20.877338409423828 0.7995593547821045

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
19.75017738342285 0.7417758703231812 22.90733528137207 0.8184797763824463 26.633710861206055 0.8976545333862305 19.150224685668945 0.7296798229217529 19.73414421081543 0.7561883926391602 20.877338409423828 0.7995593547821045
github-actions[bot] commented 1 year ago

Running experiment on sagemaker with git sha 34143c4f2535bbe4673b759b2d77b230624a140e

github-actions[bot] commented 1 year ago

Training job backward-optimization-34143c4-230723-221306-tat-truck-baseline created

github-actions[bot] commented 1 year ago

Training job backward-optimization-34143c4-230723-221306-tat-truck-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/backward-optimization-34143c4-230723-221306-tat-truck-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/backward-optimization-34143c4-230723-221306-tat-truck-baseline/output/output.tar.gz

github-actions[bot] commented 1 year ago

Training job backward-optimization-34143c4-230723-221306-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 0.031189490109682083 0.04808836430311203 443357.0 25.88581085205078 0.8843161463737488 0.11568385362625122 0.053806550800800323 25.142391204833984 0.8657335042953491

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
22.96419334411621 0.8172551393508911 23.369224548339844 0.8599488139152527 25.88581085205078 0.8843161463737488 22.989625930786133 0.8136014342308044 23.77571678161621 0.8358147144317627 25.142391204833984 0.8657335042953491