Inference optimization

[wanmeihuali]

try use some idea from original implementation.

[github-actions[bot]]

Running experiment on sagemaker with git sha a56205d1f51fcdd59096a584e209096ec42e8d1d

[github-actions[bot]]

Training job cuda-optimization-a56205d-230716-084027-tat-truck-baseline created

[github-actions[bot]]

Running experiment on sagemaker with git sha a56205d1f51fcdd59096a584e209096ec42e8d1d

[github-actions[bot]]

Training job cuda-optimization-a56205d-230716-084135-tat-train-baseline created

[github-actions[bot]]

Training job cuda-optimization-a56205d-230716-084027-tat-truck-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/cuda-optimization-a56205d-230716-084027-tat-truck-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-truck-baseline/cuda-optimization-a56205d-230716-084027-tat-truck-baseline/output/output.tar.gz

[github-actions[bot]]

Training job cuda-optimization-a56205d-230716-084027-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	0.03340969979763031	0.05477404594421387	421312.0	24.220762252807617	0.8597685694694519	0.1402314305305481	0.05353996902704239	25.222740173339844	0.8658005595207214

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
22.08319091796875	0.7829725742340088	23.802860260009766	0.848724365234375	24.220762252807617	0.8597685694694519	22.9982967376709	0.8132348656654358	23.748580932617188	0.8357873558998108	25.222740173339844	0.8658005595207214

[github-actions[bot]]

Training job cuda-optimization-a56205d-230716-084135-tat-train-baseline completed.

Model url: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/cuda-optimization-a56205d-230716-084135-tat-train-baseline/output/model.tar.gz,

tensorboard output path: s3://taichi-3d-gaussian-splatting-log/tat-train-baseline/cuda-optimization-a56205d-230716-084135-tat-train-baseline/output/output.tar.gz

[github-actions[bot]]

Training job cuda-optimization-a56205d-230716-084135-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	0.027324626222252846	0.04322688281536102	652324.0	27.30440902709961	0.8931640982627869	0.10683590173721313	0.10098321735858917	20.809288024902344	0.7996891140937805

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
21.282970428466797	0.8127028942108154	15.480735778808594	0.6046363711357117	27.30440902709961	0.8931640982627869	19.05428695678711	0.7276621460914612	19.60567283630371	0.7535765171051025	20.809288024902344	0.7996891140937805