text stringlengths 0 89 |
|---|
step lr_h_ lr_z_ loss_n loss_raw |
0 1.00e-04 1.00e-03 9.899999e-01 3.373188e-04 (0.4s) |
200 1.00e-04 1.00e-03 4.137788e-01 1.052848e-05 (3.1s) |
400 9.99e-05 9.99e-04 3.288589e-01 7.403866e-06 (5.8s) |
600 9.98e-05 9.98e-04 1.976465e-01 5.062388e-06 (8.5s) |
800 9.96e-05 9.96e-04 1.012213e-01 2.913257e-06 (11.2s) |
1000 9.94e-05 9.94e-04 4.944959e-02 1.512166e-06 (13.9s) |
1200 9.91e-05 9.91e-04 2.111490e-02 7.917419e-07 (16.6s) |
1400 9.88e-05 9.88e-04 8.711814e-03 4.729739e-07 (19.3s) |
1600 9.84e-05 9.84e-04 3.305168e-03 2.832524e-07 (22.0s) |
1800 9.80e-05 9.80e-04 1.212854e-03 1.596487e-07 (24.6s) |
2000 9.76e-05 9.76e-04 6.498378e-04 9.049983e-08 (27.3s) |
2200 9.70e-05 9.70e-04 4.106582e-04 5.247339e-08 (30.1s) |
2400 9.65e-05 9.65e-04 2.513851e-04 3.072650e-08 (32.8s) |
2600 9.59e-05 9.59e-04 1.161344e-04 1.789874e-08 (35.5s) |
2800 9.52e-05 9.52e-04 3.884140e-05 1.084195e-08 (38.2s) |
3000 9.45e-05 9.45e-04 1.439986e-05 7.015490e-09 (40.9s) |
3200 9.38e-05 9.38e-04 7.394677e-06 4.783196e-09 (43.6s) |
3400 9.30e-05 9.30e-04 4.461457e-06 3.387385e-09 (46.3s) |
3600 9.22e-05 9.22e-04 2.807948e-06 2.401918e-09 (48.9s) |
3800 9.13e-05 9.13e-04 1.843382e-06 1.727472e-09 (51.5s) |
4000 9.04e-05 9.04e-04 1.234968e-06 1.244131e-09 (54.1s) |
4200 8.95e-05 8.95e-04 8.830935e-07 9.190770e-10 (56.7s) |
4400 8.85e-05 8.85e-04 5.519699e-07 6.303793e-10 (59.3s) |
4600 8.75e-05 8.75e-04 5.243079e-07 4.509174e-10 (61.9s) |
4800 8.64e-05 8.64e-04 1.614547e-06 1.016686e-09 (64.5s) |
5000 8.53e-05 8.53e-04 1.261206e-06 7.670667e-10 (67.1s) |
5200 8.42e-05 8.42e-04 4.127501e-07 3.040752e-10 (69.7s) |
5400 8.31e-05 8.31e-04 2.427853e-07 1.944843e-10 (72.4s) |
5600 8.19e-05 8.19e-04 1.182562e-07 1.101810e-10 (75.0s) |
5800 8.06e-05 8.06e-04 5.559014e-08 6.512272e-11 (77.6s) |
6000 7.94e-05 7.94e-04 5.220014e-06 2.599490e-09 (80.2s) |
6200 7.81e-05 7.81e-04 8.069476e-08 5.781436e-11 (82.8s) |
6400 7.68e-05 7.68e-04 4.942267e-06 2.463083e-09 (85.4s) |
6600 7.54e-05 7.54e-04 4.327044e-08 3.290569e-11 (88.0s) |
6800 7.41e-05 7.41e-04 2.208285e-06 1.188761e-09 (90.6s) |
7000 7.27e-05 7.27e-04 1.718795e-08 1.320014e-11 (93.3s) |
7200 7.13e-05 7.13e-04 2.477446e-08 1.527550e-11 (95.9s) |
7400 6.99e-05 6.99e-04 1.596595e-06 8.209299e-10 (98.5s) |
7600 6.84e-05 6.84e-04 1.666000e-07 8.027988e-11 (101.1s) |
7800 6.69e-05 6.69e-04 1.939004e-07 1.011138e-10 (103.7s) |
8000 6.54e-05 6.54e-04 2.236593e-06 1.160069e-09 (106.3s) |
8200 6.39e-05 6.39e-04 9.002082e-10 1.035532e-12 (109.0s) |
8400 6.24e-05 6.24e-04 3.600618e-10 6.364629e-13 (111.6s) |
8600 6.09e-05 6.09e-04 2.657786e-10 4.558121e-13 (114.2s) |
8800 5.94e-05 5.94e-04 1.677942e-10 3.083431e-13 (116.8s) |
9000 5.78e-05 5.78e-04 1.116643e-10 2.081768e-13 (119.4s) |
9200 5.63e-05 5.63e-04 7.259565e-11 1.371304e-13 (122.0s) |
9400 5.47e-05 5.47e-04 7.449564e-11 1.197476e-13 (124.7s) |
9600 5.31e-05 5.31e-04 4.271805e-11 8.051700e-14 (127.3s) |
9800 5.16e-05 5.16e-04 2.944148e-11 5.686489e-14 (129.9s) |
10000 5.00e-05 5.00e-04 2.038757e-11 3.981455e-14 (132.5s) |
10200 4.84e-05 4.84e-04 1.390902e-11 2.738762e-14 (135.1s) |
10400 4.69e-05 4.69e-04 9.307127e-12 1.847055e-14 (137.7s) |
10600 4.53e-05 4.53e-04 6.103900e-12 1.217573e-14 (140.3s) |
10800 4.37e-05 4.37e-04 3.909524e-12 7.861472e-15 (142.9s) |
11000 4.22e-05 4.22e-04 2.487353e-12 4.965374e-15 (145.5s) |
11200 4.06e-05 4.06e-04 1.502118e-12 3.051524e-15 (148.2s) |
11400 3.91e-05 3.91e-04 9.856621e-13 1.839018e-15 (150.8s) |
11600 3.76e-05 3.76e-04 5.327559e-13 1.077988e-15 (153.4s) |
11800 3.60e-05 3.60e-04 3.108704e-13 6.225594e-16 (156.0s) |
12000 3.45e-05 3.45e-04 1.811145e-13 3.509059e-16 (158.6s) |
12200 3.31e-05 3.31e-04 1.070902e-13 1.954636e-16 (161.2s) |
12400 3.16e-05 3.16e-04 1.917674e-09 5.998317e-13 (163.8s) |
12600 3.01e-05 3.01e-04 3.140196e-11 2.052744e-14 (166.3s) |
12800 2.87e-05 2.87e-04 9.514572e-12 8.416558e-15 (168.8s) |
13000 2.73e-05 2.73e-04 4.137469e-12 4.316568e-15 (171.5s) |
13200 2.59e-05 2.59e-04 2.183356e-12 2.499800e-15 (174.1s) |
13400 2.45e-05 2.45e-04 1.304119e-12 1.571343e-15 (176.7s) |
13600 2.32e-05 2.32e-04 8.489776e-13 1.042979e-15 (179.3s) |
13800 2.19e-05 2.19e-04 5.911657e-13 7.254407e-16 (181.9s) |
14000 2.06e-05 2.06e-04 4.348660e-13 5.296722e-16 (184.5s) |
14200 1.93e-05 1.93e-04 3.334696e-13 3.988176e-16 (187.1s) |
14400 1.81e-05 1.81e-04 2.655885e-13 3.115493e-16 (189.8s) |
14600 1.69e-05 1.69e-04 2.178250e-13 2.477708e-16 (192.5s) |
14800 1.58e-05 1.58e-04 1.828971e-13 2.034260e-16 (195.1s) |
15000 1.46e-05 1.46e-04 1.573605e-13 1.699378e-16 (197.7s) |
15200 1.35e-05 1.35e-04 1.377947e-13 1.450615e-16 (200.3s) |
15400 1.25e-05 1.25e-04 1.232633e-13 1.272182e-16 (202.9s) |
15600 1.15e-05 1.15e-04 1.118347e-13 1.131620e-16 (205.5s) |
15800 1.05e-05 1.05e-04 1.032605e-13 1.022059e-16 (208.1s) |
16000 9.54e-06 9.54e-05 9.618966e-14 9.332877e-17 (210.8s) |
16200 8.64e-06 8.64e-05 9.073574e-14 8.648817e-17 (213.4s) |
16400 7.78e-06 7.78e-05 8.660675e-14 8.176718e-17 (216.0s) |
16600 6.96e-06 6.96e-05 8.335737e-14 7.850351e-17 (218.6s) |
16800 6.18e-06 6.18e-05 8.092426e-14 7.627139e-17 (221.2s) |
17000 5.45e-06 5.45e-05 7.924940e-14 7.456801e-17 (223.8s) |
17200 4.76e-06 4.76e-05 7.790726e-14 7.271501e-17 (226.4s) |
17400 4.11e-06 4.11e-05 7.702928e-14 7.247145e-17 (229.0s) |
17600 3.51e-06 3.51e-05 7.649096e-14 7.230193e-17 (231.7s) |
17800 2.95e-06 2.95e-05 7.601301e-14 7.226037e-17 (234.3s) |
18000 2.44e-06 2.44e-05 7.577821e-14 7.188597e-17 (236.9s) |
18200 1.98e-06 1.98e-05 7.561422e-14 7.176419e-17 (239.5s) |
18400 1.57e-06 1.57e-05 7.541634e-14 7.111519e-17 (242.1s) |
18600 1.20e-06 1.20e-05 7.525163e-14 7.169881e-17 (244.7s) |
18800 8.84e-07 8.84e-06 7.521820e-14 7.134831e-17 (247.3s) |
19000 6.14e-07 6.14e-06 7.522505e-14 7.148845e-17 (249.9s) |
19200 3.93e-07 3.93e-06 7.531087e-14 7.144312e-17 (252.5s) |
19400 2.21e-07 2.21e-06 7.528377e-14 7.110751e-17 (255.1s) |
19600 9.82e-08 9.82e-07 7.528926e-14 7.130197e-17 (257.8s) |
End of preview. Expand in Data Studio
Hypernet — Prior Topic 03 Archive
Complete archive of the prior topic 03 research thread: per-shape SIREN decoders, per-layer hypernetwork architectures, mapper experiments, and ancillary checkpoints. This work predates the current image-to-3D pipeline documented in hypernet-image-to-3d and the main dataset.
100 shapes, naming convention obj_NN for NN in [0..99].
Contents
| Path | Size | Description |
|---|---|---|
watertight/ |
~5.6 GB | 100 watertight .obj meshes (filename obj_NN.obj) |
sdf_samples/ |
~591 MB | 100 .npz files with point/SDF pairs |
shape_sirens/ |
~505 MB | 100 trained per-shape SIREN decoders + reconstructions in meshes_w30/ |
hypernets/ |
~6.7 GB | 100 trained per-shape hypernetworks (full weight space) |
image_sirens/ |
~629 MB | 2400 image-SIREN files (24 views × 100 shapes) |
checkpoints/ |
~220 MB | VAE, direct-mapper, and h90 variant checkpoints |
diagnostics/ |
~1 MB | Training diagnostic plots and metric dumps |
hypernet_meshes/ |
~188 MB | Sample reconstructions from trained hypernets |
Note on naming
The corresponding folder in the source repo was named sdf_samples_old10;
it has been renamed to sdf_samples here for clarity. The _old10 suffix
was a local distinguisher when this codebase ran alongside other variants.
Citation
Jain, A. (2026). Hypernet research archive — prior topic 03 (per-shape hypernetworks
and SIREN decoders). HuggingFace: bobthebuilderinternational/hypernet-prior-topic03.
License
Apache 2.0. Underlying meshes are derived from publicly available sources under their own licenses.
- Downloads last month
- 12