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
637 stars 58 forks source link

use depth in ray instead of depth to camera ray for sorting key, beca… #137

Open wanmeihuali opened 9 months ago

wanmeihuali commented 9 months ago

…use the occlusion shall be more correct with depth in ray.

github-actions[bot] commented 9 months ago

Running experiment on sagemaker with git sha 637550361281044d57d55c24710ba301d10cde53

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-6375503-231005-232806-tat-truck-baseline created

github-actions[bot] commented 9 months ago

Running experiment on sagemaker with git sha 637550361281044d57d55c24710ba301d10cde53

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-6375503-231005-232815-tat-train-baseline created

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-6375503-231005-232806-tat-truck-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/use-depth-in-ray-6375503-231005-232806-tat-truck-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/use-depth-in-ray-6375503-231005-232806-tat-truck-baseline/output/output.tar.gz

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-6375503-231005-232806-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 0.03345212712883949 0.04938406124711037 443277.0 26.418113708496094 0.8868882060050964 0.11311179399490356 0.05516449734568596 24.924814224243164 0.865136444568634

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
21.80914306640625 0.7733711004257202 24.218690872192383 0.8531867861747742 26.418113708496094 0.8868882060050964 22.97800064086914 0.814297616481781 23.802865982055664 0.837710976600647 24.924814224243164 0.865136444568634
github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-6375503-231005-232815-tat-train-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/use-depth-in-ray-6375503-231005-232815-tat-train-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/use-depth-in-ray-6375503-231005-232815-tat-train-baseline/output/output.tar.gz

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-6375503-231005-232815-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 0.06498153507709503 0.10483353585004807 680547.0 20.77347755432129 0.7357584834098816 0.2642415165901184 0.09922969341278076 21.050630569458008 0.8027860522270203

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
18.892276763916016 0.6808294057846069 20.02995491027832 0.8273240923881531 20.77347755432129 0.7357584834098816 19.14443588256836 0.7345061302185059 19.867244720458984 0.7611867785453796 21.050630569458008 0.8027860522270203
github-actions[bot] commented 9 months ago

Running experiment on sagemaker with git sha 1aa937934f1a82c7dd7ea081b3407d12cb092147

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-1aa9379-231006-015340-tat-truck-baseline created

github-actions[bot] commented 9 months ago

Running experiment on sagemaker with git sha 1aa937934f1a82c7dd7ea081b3407d12cb092147

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-1aa9379-231006-015342-tat-train-baseline created

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-1aa9379-231006-015340-tat-truck-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/use-depth-in-ray-1aa9379-231006-015340-tat-truck-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/use-depth-in-ray-1aa9379-231006-015340-tat-truck-baseline/output/output.tar.gz

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-1aa9379-231006-015340-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 0.04494506120681763 0.0671030580997467 436882.0 22.990821838378906 0.8442649841308594 0.15573501586914062 0.05424079671502113 25.04224395751953 0.8660969734191895

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
23.22745132446289 0.8353778123855591 22.83414649963379 0.8479913473129272 22.990821838378906 0.8442649841308594 22.971342086791992 0.8157235383987427 23.78691291809082 0.8382209539413452 25.04224395751953 0.8660969734191895
github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-1aa9379-231006-015342-tat-train-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/use-depth-in-ray-1aa9379-231006-015342-tat-train-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/use-depth-in-ray-1aa9379-231006-015342-tat-train-baseline/output/output.tar.gz

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-1aa9379-231006-015342-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 0.04397716000676155 0.061862602829933167 679373.0 23.88727378845215 0.8665956258773804 0.13340437412261963 0.09860115498304367 20.963369369506836 0.8028493523597717

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
19.394107818603516 0.7175304889678955 20.012434005737305 0.6925652027130127 23.88727378845215 0.8665956258773804 19.15745735168457 0.7324286103248596 19.83893394470215 0.7598733305931091 20.963369369506836 0.8028493523597717
github-actions[bot] commented 9 months ago

Running experiment on sagemaker with git sha 1aa937934f1a82c7dd7ea081b3407d12cb092147

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-1aa9379-231006-061952-tat-train-baseline created

github-actions[bot] commented 9 months ago

Running experiment on sagemaker with git sha 1aa937934f1a82c7dd7ea081b3407d12cb092147

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-1aa9379-231006-062002-tat-truck-baseline created

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-1aa9379-231006-062002-tat-truck-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/use-depth-in-ray-1aa9379-231006-062002-tat-truck-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/use-depth-in-ray-1aa9379-231006-062002-tat-truck-baseline/output/output.tar.gz

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-1aa9379-231006-062002-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 0.036109477281570435 0.05328259617090225 439431.0 25.218990325927734 0.8780249357223511 0.12197506427764893 0.05479404330253601 25.006031036376953 0.8652077317237854

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
22.467849731445312 0.8326117396354675 24.738550186157227 0.8537129163742065 25.218990325927734 0.8780249357223511 22.916616439819336 0.812911331653595 23.691999435424805 0.8367347121238708 25.006031036376953 0.8652077317237854
github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-1aa9379-231006-061952-tat-train-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/use-depth-in-ray-1aa9379-231006-061952-tat-train-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/use-depth-in-ray-1aa9379-231006-061952-tat-train-baseline/output/output.tar.gz

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-1aa9379-231006-061952-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 0.043533261865377426 0.0540580227971077 683507.0 23.979360580444336 0.9038429260253906 0.09615707397460938 0.10054276883602142 20.893360137939453 0.8000771403312683

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
17.400175094604492 0.6893067359924316 22.41900634765625 0.8071045279502869 23.979360580444336 0.9038429260253906 19.150447845458984 0.7317020893096924 19.664691925048828 0.75634765625 20.893360137939453 0.8000771403312683
github-actions[bot] commented 9 months ago

Running experiment on sagemaker with git sha 8c7d1c12fda0adc75feb225764678d4beea0d5a2

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-8c7d1c1-231017-040207-tat-train-baseline created

github-actions[bot] commented 9 months ago

Running experiment on sagemaker with git sha 8c7d1c12fda0adc75feb225764678d4beea0d5a2

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-8c7d1c1-231017-040312-tat-truck-baseline created

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-8c7d1c1-231017-040312-tat-truck-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/use-depth-in-ray-8c7d1c1-231017-040312-tat-truck-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/use-depth-in-ray-8c7d1c1-231017-040312-tat-truck-baseline/output/output.tar.gz

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-8c7d1c1-231017-040312-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 0.027586055919528008 0.04325316101312637 447685.0 25.437156677246094 0.8940784335136414 0.10592156648635864 0.05411989241838455 25.073688507080078 0.8663197755813599

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
22.94533920288086 0.8416497111320496 25.79204559326172 0.8655107021331787 25.437156677246094 0.8940784335136414 23.094064712524414 0.8150384426116943 23.815950393676758 0.8376243710517883 25.073688507080078 0.8663197755813599
github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-8c7d1c1-231017-040207-tat-train-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/use-depth-in-ray-8c7d1c1-231017-040207-tat-train-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/use-depth-in-ray-8c7d1c1-231017-040207-tat-train-baseline/output/output.tar.gz

github-actions[bot] commented 9 months ago

Training job use-depth-in-ray-8c7d1c1-231017-040207-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 0.04774487391114235 0.05797021836042404 692406.0 23.455257415771484 0.9011284112930298 0.09887158870697021 0.09719205647706985 21.1036319732666 0.8059386014938354

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
25.158458709716797 0.8425201177597046 24.842872619628906 0.923937201499939 23.455257415771484 0.9011284112930298 19.301868438720703 0.735354483127594 19.949430465698242 0.7604627013206482 21.1036319732666 0.8059386014938354