Upload folder using huggingface_hub
Browse files- .gitattributes +2 -0
- README.md +44 -0
- config.json +11 -0
- config/config.yaml +58 -0
- model/SirenFNO.py +139 -0
- model/__init__.py +3 -0
- results/budget_verification.json +19 -0
- results/darcy_fields.pdf +0 -0
- results/darcy_fields.png +3 -0
- results/data_manifest.json +90 -0
- results/error_distribution.png +0 -0
- results/execution_provenance.json +69 -0
- results/independent_metric_verification.json +17 -0
- results/metrics.json +24 -0
- results/normalization.json +6 -0
- results/per_sample_metrics.csv +201 -0
- results/predictions.pt +3 -0
- results/report.md +48 -0
- results/run_metadata.json +12 -0
- results/smoke/checks.json +22 -0
- results/training_curves.png +3 -0
- results/training_history.csv +501 -0
- results/training_status.json +18 -0
- results/verification.json +31 -0
- scripts/common.py +150 -0
- scripts/inference.py +72 -0
- scripts/result.py +132 -0
- scripts/smoke.py +60 -0
- scripts/train.py +139 -0
- weight/best_model.pth +3 -0
- weight/last_model.pth +3 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
results/darcy_fields.png filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
results/training_curves.png filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,44 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
library_name: pytorch
|
| 4 |
+
tags:
|
| 5 |
+
- OneScience
|
| 6 |
+
- neural-operator
|
| 7 |
+
- darcy-flow
|
| 8 |
+
---
|
| 9 |
+
|
| 10 |
+
# SirenFNO: Darcy 32x32
|
| 11 |
+
|
| 12 |
+
This package reproduces the full-spectrum SirenFNO experiment for two-dimensional Darcy flow in Table 2 of the [SirenFNO paper](https://ijcai.org/proceedings/2026/0535.pdf). The model maps a permeability field to its pressure solution. It has four spectral blocks, width 32, and was trained for 500 epochs on 1,000 samples.
|
| 13 |
+
|
| 14 |
+
## Results
|
| 15 |
+
|
| 16 |
+
The reloaded checkpoint achieves a mean per-sample relative L2 error of **0.04524468** on 200 test fields. The paper reports **0.0351**; this run is about 28.9% higher. The loss definition and some unspecified implementation details may differ, so this is not a strict same-condition comparison. The original training report and plots are in `results/`.
|
| 17 |
+
|
| 18 |
+
## Download
|
| 19 |
+
|
| 20 |
+
```bash
|
| 21 |
+
hf download OneScience-Group/SirenFNO --local-dir ./SirenFNO
|
| 22 |
+
cd SirenFNO
|
| 23 |
+
```
|
| 24 |
+
|
| 25 |
+
The original 32x32 Darcy tensors are published by the [NeuralOperator team on Zenodo](https://zenodo.org/records/12784353). The same two files are mirrored for this experiment in the [OneScience/SirenFNO-Darcy32 ModelScope dataset](https://www.modelscope.cn/datasets/OneScience/SirenFNO-Darcy32). No dataset files are included in this Hugging Face model repository.
|
| 26 |
+
|
| 27 |
+
```bash
|
| 28 |
+
modelscope download --dataset OneScience/SirenFNO-Darcy32 --local_dir ./data
|
| 29 |
+
```
|
| 30 |
+
|
| 31 |
+
Run the commands below from the model directory. The published configuration reads `data/darcy_train_32.pt` and `data/darcy_test_32.pt` and writes to `results/`. Each file contains float32 `x` (permeability) and `y` (pressure) tensors on a 32x32 grid. The inference script checks the data SHA256 against the checkpoint.
|
| 32 |
+
|
| 33 |
+
## Environment and usage
|
| 34 |
+
|
| 35 |
+
Use Python 3.11, PyTorch with a supported GPU or DCU, `onescience[cfd-gpu]` or `onescience[cfd-dcu]`, PyYAML, and Matplotlib. On a cluster, run model computation on an allocated accelerator node.
|
| 36 |
+
|
| 37 |
+
```bash
|
| 38 |
+
python scripts/inference.py --config config/config.yaml
|
| 39 |
+
python scripts/result.py --config config/config.yaml
|
| 40 |
+
```
|
| 41 |
+
|
| 42 |
+
To start a new 500-epoch training run, move the included `weight/` and `results/` outside the model directory first, then run `python scripts/train.py --config config/config.yaml`. The published inference accepts a different data/output location but retains the model-configuration and data-hash checks. The weights are unchanged from the original run.
|
| 43 |
+
|
| 44 |
+
Checkpoint: `weight/best_model.pth` (SHA256 `ceb2392d5ba68e386f08e0f117ea1705df1f3a0185d3cef0971f07951113ce1e`). The dataset's original use and redistribution terms should be checked at the Zenodo source.
|
config.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_type": "sirenfno_darcy32",
|
| 3 |
+
"architectures": ["SirenFNO"],
|
| 4 |
+
"framework": "pytorch",
|
| 5 |
+
"experiment": "Darcy_32_Table2",
|
| 6 |
+
"input_channels": 1,
|
| 7 |
+
"output_channels": 1,
|
| 8 |
+
"resolution": [32, 32],
|
| 9 |
+
"checkpoint": "weight/best_model.pth",
|
| 10 |
+
"dataset": "OneScience/SirenFNO-Darcy32"
|
| 11 |
+
}
|
config/config.yaml
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
experiment: SirenFNO_Table2_Darcy_32
|
| 2 |
+
data:
|
| 3 |
+
root: data
|
| 4 |
+
train_file: darcy_train_32.pt
|
| 5 |
+
test_file: darcy_test_32.pt
|
| 6 |
+
n_train: 1000
|
| 7 |
+
n_test: 200
|
| 8 |
+
resolution: [32, 32]
|
| 9 |
+
selection: first_n
|
| 10 |
+
normalize_input: false
|
| 11 |
+
normalize_output: true
|
| 12 |
+
normalization_epsilon: 1.0e-7
|
| 13 |
+
model:
|
| 14 |
+
in_channels: 1
|
| 15 |
+
out_channels: 1
|
| 16 |
+
width: 32
|
| 17 |
+
layers: 4
|
| 18 |
+
lifting_channels: 64
|
| 19 |
+
projection_channels: 64
|
| 20 |
+
residual_expansion: 4
|
| 21 |
+
siren_hidden: 32
|
| 22 |
+
siren_layers: 1
|
| 23 |
+
rff_features: 16
|
| 24 |
+
rff_sigma: 512.0
|
| 25 |
+
omega: 30.0
|
| 26 |
+
learnable_omega: true
|
| 27 |
+
add_grid: true
|
| 28 |
+
truncation: false
|
| 29 |
+
factorization: dense
|
| 30 |
+
siren_activation: sin
|
| 31 |
+
training:
|
| 32 |
+
epochs: 500
|
| 33 |
+
batch_size: 32
|
| 34 |
+
optimizer: AdamW
|
| 35 |
+
lr: 0.001
|
| 36 |
+
weight_decay: 0.0001
|
| 37 |
+
betas: [0.9, 0.999]
|
| 38 |
+
optimizer_epsilon: 1.0e-8
|
| 39 |
+
scheduler: CosineAnnealingLR
|
| 40 |
+
t_max: 500
|
| 41 |
+
eta_min: 0.0
|
| 42 |
+
loss: mean_squared_relative_l2_original_scale
|
| 43 |
+
seed: 42
|
| 44 |
+
shuffle: false
|
| 45 |
+
drop_last: false
|
| 46 |
+
early_stopping: false
|
| 47 |
+
precision: float32
|
| 48 |
+
checkpoint_selection: minimum_full_training_original_scale_squared_relative_l2
|
| 49 |
+
print_batch_interval: 16
|
| 50 |
+
runtime:
|
| 51 |
+
device: cuda
|
| 52 |
+
cpu_threads: 8
|
| 53 |
+
output:
|
| 54 |
+
root: .
|
| 55 |
+
checkpoint: weight/best_model.pth
|
| 56 |
+
last_checkpoint: weight/last_model.pth
|
| 57 |
+
results: results
|
| 58 |
+
paper_reference: 0.0351
|
model/SirenFNO.py
ADDED
|
@@ -0,0 +1,139 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Independent implementation of SirenFNO, paper equations (5)-(7), (14).
|
| 2 |
+
|
| 3 |
+
Dense complex kernels cover every rFFT bin. The RFF and initialization
|
| 4 |
+
hyperparameters fill paper omissions using read-only official implementation
|
| 5 |
+
observations. Paper-specified sine activations and 64-channel lifting/projection
|
| 6 |
+
take precedence over conflicting repository choices. See reproduction_spec.md.
|
| 7 |
+
"""
|
| 8 |
+
|
| 9 |
+
import math
|
| 10 |
+
|
| 11 |
+
import torch
|
| 12 |
+
from torch import nn
|
| 13 |
+
from torch.nn import functional as F
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
class PointwiseMLP(nn.Module):
|
| 17 |
+
def __init__(self, input_channels, hidden_channels, output_channels):
|
| 18 |
+
super().__init__()
|
| 19 |
+
self.layers = nn.Sequential(
|
| 20 |
+
nn.Conv2d(input_channels, hidden_channels, 1),
|
| 21 |
+
nn.GELU(),
|
| 22 |
+
nn.Conv2d(hidden_channels, output_channels, 1),
|
| 23 |
+
)
|
| 24 |
+
|
| 25 |
+
def forward(self, x):
|
| 26 |
+
return self.layers(x)
|
| 27 |
+
|
| 28 |
+
|
| 29 |
+
class KernelNetwork(nn.Module):
|
| 30 |
+
"""One real component of a complex mode-to-channel-matrix hypernetwork."""
|
| 31 |
+
|
| 32 |
+
def __init__(self, width, hidden, depth, features, sigma, omega, learnable_omega):
|
| 33 |
+
super().__init__()
|
| 34 |
+
self.frequencies = nn.Parameter(torch.randn(2, features // 2) * sigma)
|
| 35 |
+
self.affine = nn.ModuleList()
|
| 36 |
+
self.scales = nn.ParameterList()
|
| 37 |
+
for index in range(depth):
|
| 38 |
+
inputs = features if index == 0 else hidden
|
| 39 |
+
layer = nn.Linear(inputs, hidden, bias=True)
|
| 40 |
+
bound = 1.0 / inputs if index == 0 else math.sqrt(6.0 / inputs) / omega
|
| 41 |
+
nn.init.uniform_(layer.weight, -bound, bound)
|
| 42 |
+
nn.init.zeros_(layer.bias)
|
| 43 |
+
self.affine.append(layer)
|
| 44 |
+
self.scales.append(nn.Parameter(torch.rand(hidden) * omega,
|
| 45 |
+
requires_grad=learnable_omega))
|
| 46 |
+
self.readout = nn.Linear(hidden, width * width, bias=False)
|
| 47 |
+
bound = math.sqrt(6.0 / hidden) / omega
|
| 48 |
+
nn.init.uniform_(self.readout.weight, -bound, bound)
|
| 49 |
+
|
| 50 |
+
def forward(self, coordinates):
|
| 51 |
+
phase = math.pi * (coordinates @ self.frequencies)
|
| 52 |
+
hidden = torch.cat((phase.cos(), phase.sin()), dim=-1)
|
| 53 |
+
for affine, scale in zip(self.affine, self.scales):
|
| 54 |
+
# Paper eq. (7): sin(w W x + b); b starts at zero and is learned.
|
| 55 |
+
hidden = torch.sin(F.linear(hidden, affine.weight) * scale + affine.bias)
|
| 56 |
+
return self.readout(hidden)
|
| 57 |
+
|
| 58 |
+
|
| 59 |
+
class FullSpectrumOperator(nn.Module):
|
| 60 |
+
def __init__(self, width, **kernel_args):
|
| 61 |
+
super().__init__()
|
| 62 |
+
self.width = width
|
| 63 |
+
self.real_kernel = KernelNetwork(width=width, **kernel_args)
|
| 64 |
+
self.imag_kernel = KernelNetwork(width=width, **kernel_args)
|
| 65 |
+
self.register_buffer("coordinates", torch.empty(0), persistent=False)
|
| 66 |
+
self.coordinate_shape = None
|
| 67 |
+
|
| 68 |
+
def mode_coordinates(self, height, half_width, x):
|
| 69 |
+
shape = (height, half_width)
|
| 70 |
+
if self.coordinate_shape != shape or self.coordinates.device != x.device:
|
| 71 |
+
# Midpoints of normalized array-index coordinates, supplement from
|
| 72 |
+
# official dense kernel. No frequency mask or zero-filled high bins.
|
| 73 |
+
u = 2 * (torch.arange(height, device=x.device, dtype=x.dtype) + .5) / height - 1
|
| 74 |
+
v = 2 * (torch.arange(half_width, device=x.device, dtype=x.dtype) + .5) / half_width - 1
|
| 75 |
+
self.coordinates = torch.stack(torch.meshgrid(u, v, indexing="ij"), -1)
|
| 76 |
+
self.coordinate_shape = shape
|
| 77 |
+
return self.coordinates
|
| 78 |
+
|
| 79 |
+
def forward(self, x):
|
| 80 |
+
height, width = x.shape[-2:]
|
| 81 |
+
half_width = width // 2 + 1
|
| 82 |
+
coordinates = self.mode_coordinates(height, half_width, x)
|
| 83 |
+
real = self.real_kernel(coordinates).reshape(height, half_width, self.width, self.width)
|
| 84 |
+
imag = self.imag_kernel(coordinates).reshape_as(real)
|
| 85 |
+
kernel = torch.complex(real, imag)
|
| 86 |
+
spectrum = torch.fft.rfft2(x, norm="backward")
|
| 87 |
+
mixed = torch.einsum("bihw,hwio->bohw", spectrum, kernel)
|
| 88 |
+
return torch.fft.irfft2(mixed.contiguous(), s=(height, width), norm="backward")
|
| 89 |
+
|
| 90 |
+
|
| 91 |
+
class ResidualOperatorBlock(nn.Module):
|
| 92 |
+
def __init__(self, width, expansion, activate, **kernel_args):
|
| 93 |
+
super().__init__()
|
| 94 |
+
self.spectral = FullSpectrumOperator(width, **kernel_args)
|
| 95 |
+
self.channel_mlp = PointwiseMLP(width, width * expansion, width)
|
| 96 |
+
self.activate = activate
|
| 97 |
+
|
| 98 |
+
def forward(self, x):
|
| 99 |
+
# Enhanced architecture eq. (14), not the vanilla-FNO ablation.
|
| 100 |
+
result = x + self.channel_mlp(self.spectral(x))
|
| 101 |
+
return F.gelu(result) if self.activate else result
|
| 102 |
+
|
| 103 |
+
|
| 104 |
+
class SirenFNO(nn.Module):
|
| 105 |
+
def __init__(self, in_channels=1, out_channels=1, width=32, layers=4,
|
| 106 |
+
lifting_channels=64, projection_channels=64,
|
| 107 |
+
residual_expansion=4, siren_hidden=32, siren_layers=1,
|
| 108 |
+
rff_features=16, rff_sigma=512.0, omega=30.0,
|
| 109 |
+
learnable_omega=True, add_grid=True, truncation=False,
|
| 110 |
+
factorization="dense", siren_activation="sin"):
|
| 111 |
+
super().__init__()
|
| 112 |
+
if truncation or factorization != "dense" or siren_activation != "sin":
|
| 113 |
+
raise ValueError("This reproduction requires dense full-spectrum sine SirenFNO")
|
| 114 |
+
if rff_features % 2 or siren_layers < 1:
|
| 115 |
+
raise ValueError("RFF features must be even and SIREN depth positive")
|
| 116 |
+
self.in_channels = in_channels
|
| 117 |
+
self.add_grid = add_grid
|
| 118 |
+
self.lifting = PointwiseMLP(in_channels + 2 * int(add_grid), lifting_channels, width)
|
| 119 |
+
kernel_args = dict(hidden=siren_hidden, depth=siren_layers, features=rff_features,
|
| 120 |
+
sigma=rff_sigma, omega=omega, learnable_omega=learnable_omega)
|
| 121 |
+
self.blocks = nn.ModuleList([
|
| 122 |
+
ResidualOperatorBlock(width, residual_expansion, i < layers - 1, **kernel_args)
|
| 123 |
+
for i in range(layers)
|
| 124 |
+
])
|
| 125 |
+
self.projection = PointwiseMLP(width, projection_channels, out_channels)
|
| 126 |
+
|
| 127 |
+
def forward(self, x):
|
| 128 |
+
if x.ndim != 4 or x.shape[1] != self.in_channels:
|
| 129 |
+
raise ValueError(f"Expected NCHW permeability with {self.in_channels} channels: {x.shape}")
|
| 130 |
+
if self.add_grid:
|
| 131 |
+
batch, _, height, width = x.shape
|
| 132 |
+
row = torch.linspace(0, 1, height, device=x.device, dtype=x.dtype)
|
| 133 |
+
col = torch.linspace(0, 1, width, device=x.device, dtype=x.dtype)
|
| 134 |
+
grid = torch.stack(torch.meshgrid(row, col, indexing="ij"), 0)
|
| 135 |
+
x = torch.cat((x, grid.unsqueeze(0).expand(batch, -1, -1, -1)), 1)
|
| 136 |
+
x = self.lifting(x)
|
| 137 |
+
for block in self.blocks:
|
| 138 |
+
x = block(x)
|
| 139 |
+
return self.projection(x)
|
model/__init__.py
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from .SirenFNO import SirenFNO
|
| 2 |
+
|
| 3 |
+
__all__ = ["SirenFNO"]
|
results/budget_verification.json
ADDED
|
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"status": "passed",
|
| 3 |
+
"actual_completed_epochs": 500,
|
| 4 |
+
"optimizer_steps_per_parameter": [
|
| 5 |
+
16000
|
| 6 |
+
],
|
| 7 |
+
"batches_per_epoch": 32,
|
| 8 |
+
"batch_size": 32,
|
| 9 |
+
"final_batch_size": 8,
|
| 10 |
+
"scheduler_last_epoch": 500,
|
| 11 |
+
"scheduler_final_lr": [
|
| 12 |
+
0.0
|
| 13 |
+
],
|
| 14 |
+
"history_rows": 500,
|
| 15 |
+
"selected_epoch": 500,
|
| 16 |
+
"selection_rule": "minimum_full_training_original_scale_squared_relative_l2",
|
| 17 |
+
"finite_weights": true,
|
| 18 |
+
"checkpoint_loading_device": "CPU, metadata inspection only; no model forward"
|
| 19 |
+
}
|
results/darcy_fields.pdf
ADDED
|
Binary file (65.2 kB). View file
|
|
|
results/darcy_fields.png
ADDED
|
Git LFS Details
|
results/data_manifest.json
ADDED
|
@@ -0,0 +1,90 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"selection": "first_n",
|
| 3 |
+
"no_download": true,
|
| 4 |
+
"files": {
|
| 5 |
+
"train": {
|
| 6 |
+
"path": "/public/home/liuyushuang/zhaomingsi/onescience-cfd-data/SirenFNO/darcy_train_32.pt",
|
| 7 |
+
"sha256": "f44a1151802eaf2eac65ddaded1ecf105bac019add0e3d277a3e9939cca1cd8b",
|
| 8 |
+
"selected_indices": [
|
| 9 |
+
0,
|
| 10 |
+
999
|
| 11 |
+
],
|
| 12 |
+
"x": {
|
| 13 |
+
"raw_shape": [
|
| 14 |
+
5000,
|
| 15 |
+
32,
|
| 16 |
+
32
|
| 17 |
+
],
|
| 18 |
+
"raw_dtype": "torch.float32",
|
| 19 |
+
"selected_shape": [
|
| 20 |
+
1000,
|
| 21 |
+
1,
|
| 22 |
+
32,
|
| 23 |
+
32
|
| 24 |
+
],
|
| 25 |
+
"min": 0.0,
|
| 26 |
+
"max": 1.0,
|
| 27 |
+
"meaning": "permeability"
|
| 28 |
+
},
|
| 29 |
+
"y": {
|
| 30 |
+
"raw_shape": [
|
| 31 |
+
5000,
|
| 32 |
+
32,
|
| 33 |
+
32
|
| 34 |
+
],
|
| 35 |
+
"raw_dtype": "torch.float32",
|
| 36 |
+
"selected_shape": [
|
| 37 |
+
1000,
|
| 38 |
+
1,
|
| 39 |
+
32,
|
| 40 |
+
32
|
| 41 |
+
],
|
| 42 |
+
"min": -0.42788004875183105,
|
| 43 |
+
"max": 2.0652403831481934,
|
| 44 |
+
"meaning": "pressure"
|
| 45 |
+
}
|
| 46 |
+
},
|
| 47 |
+
"test": {
|
| 48 |
+
"path": "/public/home/liuyushuang/zhaomingsi/onescience-cfd-data/SirenFNO/darcy_test_32.pt",
|
| 49 |
+
"sha256": "f9fbe30065908327f132425d760d573ed7ce29cb2a54cfdb3fcefd3d80d7ac83",
|
| 50 |
+
"selected_indices": [
|
| 51 |
+
0,
|
| 52 |
+
199
|
| 53 |
+
],
|
| 54 |
+
"x": {
|
| 55 |
+
"raw_shape": [
|
| 56 |
+
1000,
|
| 57 |
+
32,
|
| 58 |
+
32
|
| 59 |
+
],
|
| 60 |
+
"raw_dtype": "torch.float32",
|
| 61 |
+
"selected_shape": [
|
| 62 |
+
200,
|
| 63 |
+
1,
|
| 64 |
+
32,
|
| 65 |
+
32
|
| 66 |
+
],
|
| 67 |
+
"min": 0.0,
|
| 68 |
+
"max": 1.0,
|
| 69 |
+
"meaning": "permeability"
|
| 70 |
+
},
|
| 71 |
+
"y": {
|
| 72 |
+
"raw_shape": [
|
| 73 |
+
1000,
|
| 74 |
+
32,
|
| 75 |
+
32
|
| 76 |
+
],
|
| 77 |
+
"raw_dtype": "torch.float32",
|
| 78 |
+
"selected_shape": [
|
| 79 |
+
200,
|
| 80 |
+
1,
|
| 81 |
+
32,
|
| 82 |
+
32
|
| 83 |
+
],
|
| 84 |
+
"min": -0.11912135034799576,
|
| 85 |
+
"max": 2.0931766033172607,
|
| 86 |
+
"meaning": "pressure"
|
| 87 |
+
}
|
| 88 |
+
}
|
| 89 |
+
}
|
| 90 |
+
}
|
results/error_distribution.png
ADDED
|
results/execution_provenance.json
ADDED
|
@@ -0,0 +1,69 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"status": "completed",
|
| 3 |
+
"verified_at": "2026-11-12T08:13:49.919431+00:00",
|
| 4 |
+
"execution_channel": "local_slurm",
|
| 5 |
+
"config_source": "onescience.json",
|
| 6 |
+
"partition": "hx1hdnormal01",
|
| 7 |
+
"jobs": [
|
| 8 |
+
{
|
| 9 |
+
"stage": "smoke",
|
| 10 |
+
"job_id": "1100629",
|
| 11 |
+
"node": "f11r3n12",
|
| 12 |
+
"state": "COMPLETED",
|
| 13 |
+
"exit_code": "0:0",
|
| 14 |
+
"slurm_elapsed": "00:00:55",
|
| 15 |
+
"stdout": "logs/1100629.out",
|
| 16 |
+
"stderr": "logs/1100629.err",
|
| 17 |
+
"started_at": "2026-11-12T15:39:26+08:00",
|
| 18 |
+
"completed_at": "2026-11-12T15:40:19+08:00"
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"stage": "train",
|
| 22 |
+
"job_id": "1100639",
|
| 23 |
+
"node": "f11r4n01",
|
| 24 |
+
"state": "COMPLETED",
|
| 25 |
+
"exit_code": "0:0",
|
| 26 |
+
"slurm_elapsed": "00:22:28",
|
| 27 |
+
"stdout": "logs/1100639.out",
|
| 28 |
+
"stderr": "logs/1100639.err",
|
| 29 |
+
"started_at": "2026-11-12T15:43:36+08:00",
|
| 30 |
+
"completed_at": "2026-11-12T16:06:02+08:00"
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"stage": "inference",
|
| 34 |
+
"job_id": "1100802",
|
| 35 |
+
"node": "f11r3n13",
|
| 36 |
+
"state": "COMPLETED",
|
| 37 |
+
"exit_code": "0:0",
|
| 38 |
+
"slurm_elapsed": "00:00:29",
|
| 39 |
+
"stdout": "logs/1100802.out",
|
| 40 |
+
"stderr": "logs/1100802.err",
|
| 41 |
+
"started_at": "2026-11-12T16:08:38+08:00",
|
| 42 |
+
"completed_at": "2026-11-12T16:09:05+08:00"
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"stage": "visualize",
|
| 46 |
+
"job_id": "1100831",
|
| 47 |
+
"node": "f12r2n12",
|
| 48 |
+
"state": "COMPLETED",
|
| 49 |
+
"exit_code": "0:0",
|
| 50 |
+
"slurm_elapsed": "00:00:30",
|
| 51 |
+
"stdout": "logs/1100831.out",
|
| 52 |
+
"stderr": "logs/1100831.err",
|
| 53 |
+
"started_at": "2026-11-12T16:10:42+08:00",
|
| 54 |
+
"completed_at": "2026-11-12T16:11:10+08:00"
|
| 55 |
+
}
|
| 56 |
+
],
|
| 57 |
+
"source_hashes_unchanged_during_formal_run": true,
|
| 58 |
+
"source_checksums": ".onescience/source_checksums.json",
|
| 59 |
+
"previous_session_had_checkpoint": false,
|
| 60 |
+
"previous_session_had_formal_training": false,
|
| 61 |
+
"reused_existing_implementation": true,
|
| 62 |
+
"formal_training_runs": 1,
|
| 63 |
+
"formal_epochs": 500,
|
| 64 |
+
"optimizer_steps": 16000,
|
| 65 |
+
"data_redownloaded": false,
|
| 66 |
+
"git_submitted": false,
|
| 67 |
+
"platform_upload_performed": false,
|
| 68 |
+
"numerical_match_to_paper_claimed": false
|
| 69 |
+
}
|
results/independent_metric_verification.json
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"status": "passed",
|
| 3 |
+
"method": "Independent NumPy float64 square/sum/sqrt from saved original-scale prediction and target fields",
|
| 4 |
+
"test_samples": 200,
|
| 5 |
+
"prediction_shape": [
|
| 6 |
+
200,
|
| 7 |
+
1,
|
| 8 |
+
32,
|
| 9 |
+
32
|
| 10 |
+
],
|
| 11 |
+
"relative_l2_mean": 0.04524468414060868,
|
| 12 |
+
"mean_squared_relative_l2": 0.002598048037006346,
|
| 13 |
+
"difference_from_report": -2.0816681711721685e-17,
|
| 14 |
+
"per_sample_csv_verified": true,
|
| 15 |
+
"sample_ids_contiguous": true,
|
| 16 |
+
"model_forward_executed": false
|
| 17 |
+
}
|
results/metrics.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"relative_l2_mean": 0.0452446841406087,
|
| 3 |
+
"squared_relative_l2_mean": 0.0025980480370063476,
|
| 4 |
+
"relative_l2_std": 0.023472677606589745,
|
| 5 |
+
"n_test": 200,
|
| 6 |
+
"resolution": [
|
| 7 |
+
32,
|
| 8 |
+
32
|
| 9 |
+
],
|
| 10 |
+
"scale": "original pressure",
|
| 11 |
+
"aggregation": "arithmetic mean of per-sample Euclidean norm ratios",
|
| 12 |
+
"completed_epochs": 500,
|
| 13 |
+
"selected_epoch": 500,
|
| 14 |
+
"checkpoint": "/public/home/liuyushuang/zhaomingsi/onescience-cfd-repro-packages/SirenFNO/weight/best_model.pth",
|
| 15 |
+
"checkpoint_sha256": "ceb2392d5ba68e386f08e0f117ea1705df1f3a0185d3cef0971f07951113ce1e",
|
| 16 |
+
"checkpoint_selection": "minimum_full_training_original_scale_squared_relative_l2",
|
| 17 |
+
"paper_table2": 0.0351,
|
| 18 |
+
"signed_gap": 0.0101446841406087,
|
| 19 |
+
"absolute_gap": 0.0101446841406087,
|
| 20 |
+
"relative_gap_percent": 28.90223401882821,
|
| 21 |
+
"last_epoch_test_relative_l2": 0.045244683949276804,
|
| 22 |
+
"test_used_for_selection": false,
|
| 23 |
+
"loss_metric_caveat": "Paper Eq15 is squared; this report's primary metric is unsquared as requested"
|
| 24 |
+
}
|
results/normalization.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"mean": 0.38948068022727966,
|
| 3 |
+
"scale": 0.3364421533180237,
|
| 4 |
+
"fit_on": "train[0:1000] only",
|
| 5 |
+
"input_normalized": false
|
| 6 |
+
}
|
results/per_sample_metrics.csv
ADDED
|
@@ -0,0 +1,201 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
sample_id,relative_l2_original_scale,squared_relative_l2_original_scale
|
| 2 |
+
0,0.03991393346308568,0.0015931220844956307
|
| 3 |
+
1,0.021679728299641918,0.00047001061914629465
|
| 4 |
+
2,0.03593162659794281,0.001291081789973991
|
| 5 |
+
3,0.02738661021542678,0.0007500264190917185
|
| 6 |
+
4,0.04595653210512599,0.002112002843129476
|
| 7 |
+
5,0.033997285314606364,0.0011558154087627495
|
| 8 |
+
6,0.04481150023577957,0.002008070553381272
|
| 9 |
+
7,0.042064766395879354,0.0017694445719399008
|
| 10 |
+
8,0.06083198607654119,0.003700530530016501
|
| 11 |
+
9,0.0694689185491364,0.004825930644386547
|
| 12 |
+
10,0.03556714502822694,0.0012650218054589286
|
| 13 |
+
11,0.08702125530485003,0.0075726988748318885
|
| 14 |
+
12,0.02955568130823827,0.000873538297594145
|
| 15 |
+
13,0.041354514747949905,0.0017101958900384061
|
| 16 |
+
14,0.08681179304979905,0.007536287412521139
|
| 17 |
+
15,0.03735277433791372,0.001395229750739106
|
| 18 |
+
16,0.029353850647195718,0.0008616485478178725
|
| 19 |
+
17,0.04678098851939099,0.0021884608868513915
|
| 20 |
+
18,0.04169761903886579,0.0017386914335103829
|
| 21 |
+
19,0.049971361130574925,0.0024971369332423346
|
| 22 |
+
20,0.0454504229100997,0.002065740942706916
|
| 23 |
+
21,0.04761020453746115,0.002266731576098886
|
| 24 |
+
22,0.03595149277673192,0.0012925098328754076
|
| 25 |
+
23,0.040455087971743825,0.0016366141428015319
|
| 26 |
+
24,0.054192290106569126,0.00293680430699455
|
| 27 |
+
25,0.06980511242524942,0.0048727537207017115
|
| 28 |
+
26,0.07488396203907079,0.0056076077706689945
|
| 29 |
+
27,0.028260061389214445,0.0007986310697221691
|
| 30 |
+
28,0.047668972821444874,0.0022723309698516503
|
| 31 |
+
29,0.03980984350243983,0.0015848236396887507
|
| 32 |
+
30,0.021163468131333275,0.00044789238334595917
|
| 33 |
+
31,0.028214660724304615,0.0007960670797876174
|
| 34 |
+
32,0.0378043413886598,0.0014291682278303362
|
| 35 |
+
33,0.026744316091654105,0.0007152584432103087
|
| 36 |
+
34,0.20639324448800686,0.042598171370286174
|
| 37 |
+
35,0.02820882164473986,0.0007957376185847441
|
| 38 |
+
36,0.043843626651466794,0.0019222635979532093
|
| 39 |
+
37,0.031707287594991584,0.0010053520866315072
|
| 40 |
+
38,0.040338752146263875,0.0016272149247177083
|
| 41 |
+
39,0.07715900144972218,0.00595351150471823
|
| 42 |
+
40,0.05407883967625949,0.0029245209007305777
|
| 43 |
+
41,0.041919412246074185,0.0017572371230563143
|
| 44 |
+
42,0.028424104787395498,0.0008079297329648396
|
| 45 |
+
43,0.035081579545689384,0.0012307172234205317
|
| 46 |
+
44,0.028690294775966026,0.0008231330143318235
|
| 47 |
+
45,0.04041641419049958,0.0016334865360180157
|
| 48 |
+
46,0.028294568593811124,0.0008005826119098828
|
| 49 |
+
47,0.03367295303320186,0.0011338677659762186
|
| 50 |
+
48,0.029862792553058848,0.000891786379067027
|
| 51 |
+
49,0.031722657450574045,0.001006326995726461
|
| 52 |
+
50,0.06022175159545226,0.003626659365224357
|
| 53 |
+
51,0.030957311773171854,0.0009583551522213647
|
| 54 |
+
52,0.0456862317083458,0.002087231767708661
|
| 55 |
+
53,0.03407718710980627,0.0011612546813167465
|
| 56 |
+
54,0.13946495686609062,0.019450474193660517
|
| 57 |
+
55,0.029599371926144764,0.0008761228184222469
|
| 58 |
+
56,0.028858977397177202,0.0008328405764107846
|
| 59 |
+
57,0.024149403057569182,0.0005831936680369317
|
| 60 |
+
58,0.036540093409536595,0.0013351784263776598
|
| 61 |
+
59,0.05121476039953415,0.002622951682781691
|
| 62 |
+
60,0.03243728415318242,0.0010521774032342995
|
| 63 |
+
61,0.06900504509404648,0.0047616962484313885
|
| 64 |
+
62,0.06499128344973694,0.0042238669244440506
|
| 65 |
+
63,0.04442883690047314,0.0019739215483288437
|
| 66 |
+
64,0.05630035905462064,0.0031697304296792045
|
| 67 |
+
65,0.04718119823503853,0.002226065466894003
|
| 68 |
+
66,0.1720637098438123,0.029605920245215633
|
| 69 |
+
67,0.04662661714593362,0.0021740414264734714
|
| 70 |
+
68,0.02897618675997014,0.0008396193991486689
|
| 71 |
+
69,0.026975387495168646,0.000727671530514501
|
| 72 |
+
70,0.03825776261148524,0.0014636564000367579
|
| 73 |
+
71,0.037729203043091984,0.0014234927622668613
|
| 74 |
+
72,0.03220297705184893,0.0010370317310019086
|
| 75 |
+
73,0.024820173326413914,0.0006160410039532287
|
| 76 |
+
74,0.05856674830066697,0.0034300640065136776
|
| 77 |
+
75,0.06786122417109645,0.004605145745999806
|
| 78 |
+
76,0.031842155276965954,0.0010139228526824108
|
| 79 |
+
77,0.0568779329293407,0.0032350992543145793
|
| 80 |
+
78,0.03760672171010369,0.0014142655177811839
|
| 81 |
+
79,0.045592281151864165,0.0020786561006306283
|
| 82 |
+
80,0.07499112763222096,0.005623669223552054
|
| 83 |
+
81,0.034180206718301624,0.0011682865313058315
|
| 84 |
+
82,0.03570924757899693,0.0012751503626580983
|
| 85 |
+
83,0.05512731892713979,0.003039021292094585
|
| 86 |
+
84,0.05130804013439043,0.0026325149824322188
|
| 87 |
+
85,0.0505747257560241,0.002557802885297047
|
| 88 |
+
86,0.03635122805893393,0.0013214117813926257
|
| 89 |
+
87,0.031322491555979236,0.0009810984772743905
|
| 90 |
+
88,0.026458901815530324,0.0007000734852838739
|
| 91 |
+
89,0.028282195464697752,0.00079988258030337
|
| 92 |
+
90,0.03185447870002263,0.0010147078132501954
|
| 93 |
+
91,0.04021917364324085,0.0016175819285451597
|
| 94 |
+
92,0.027233080757607022,0.0007416406875503459
|
| 95 |
+
93,0.03345636464459775,0.0011193283352322904
|
| 96 |
+
94,0.03578377448735742,0.001280478516562052
|
| 97 |
+
95,0.03335664344530581,0.0011126656619372631
|
| 98 |
+
96,0.02620486403556473,0.0006866948991224338
|
| 99 |
+
97,0.020525335200463072,0.00042128938509136845
|
| 100 |
+
98,0.03938109378748043,0.0015508705478983296
|
| 101 |
+
99,0.0353132344771743,0.0012470245292398915
|
| 102 |
+
100,0.045668404572701125,0.002085603176215909
|
| 103 |
+
101,0.023792262539264353,0.0005660717567372819
|
| 104 |
+
102,0.05393053053062307,0.002908502123314467
|
| 105 |
+
103,0.046056745336179726,0.002121223790961713
|
| 106 |
+
104,0.03477769913907992,0.0012094883574083604
|
| 107 |
+
105,0.02720341143728534,0.0007400255938262269
|
| 108 |
+
106,0.044851680242453285,0.002011673220571274
|
| 109 |
+
107,0.034413974679544664,0.0011843216532443413
|
| 110 |
+
108,0.021521497681973492,0.0004631748624751904
|
| 111 |
+
109,0.04420088758641451,0.001953718463426852
|
| 112 |
+
110,0.05949700516243849,0.0035398936232992327
|
| 113 |
+
111,0.029762208265228508,0.0008857890408228361
|
| 114 |
+
112,0.1352922282690137,0.018303987029994907
|
| 115 |
+
113,0.059922659391084404,0.003590725108499916
|
| 116 |
+
114,0.022767097512928672,0.000518340729163203
|
| 117 |
+
115,0.04344540401604078,0.0018875031301170122
|
| 118 |
+
116,0.029484648613721774,0.0008693445038746453
|
| 119 |
+
117,0.030418655820115905,0.0009252946219026712
|
| 120 |
+
118,0.050512677724570854,0.0025515306109063565
|
| 121 |
+
119,0.0535301966511846,0.0028654819535144952
|
| 122 |
+
120,0.04948199939034132,0.0024484682636657386
|
| 123 |
+
121,0.04095257770724269,0.0016771136208677511
|
| 124 |
+
122,0.03302203753102791,0.001090454962700616
|
| 125 |
+
123,0.03136060231583389,0.0009834873776118859
|
| 126 |
+
124,0.03963164904192357,0.001570667605782201
|
| 127 |
+
125,0.0751286660799176,0.005644316466947761
|
| 128 |
+
126,0.03172685097302068,0.0010065930726642631
|
| 129 |
+
127,0.0468804721115417,0.002197778665401039
|
| 130 |
+
128,0.039325703592900565,0.0015465109630766725
|
| 131 |
+
129,0.02464430593510875,0.0006073418150232364
|
| 132 |
+
130,0.02778770922630199,0.0007721567840455086
|
| 133 |
+
131,0.028549447120316377,0.0008150709308757411
|
| 134 |
+
132,0.05606670803524978,0.0031434757499099425
|
| 135 |
+
133,0.03412389616746267,0.0011644402896477735
|
| 136 |
+
134,0.035877006004297725,0.001287159559832415
|
| 137 |
+
135,0.043102882958641515,0.0018578585193463491
|
| 138 |
+
136,0.020240465330536642,0.0004096764367966558
|
| 139 |
+
137,0.07336290696203467,0.005382116117920155
|
| 140 |
+
138,0.055852893230061706,0.0031195456821686727
|
| 141 |
+
139,0.08951646275271638,0.00801319710375846
|
| 142 |
+
140,0.06917694049463642,0.004785449096198468
|
| 143 |
+
141,0.046187990756406026,0.0021333304901138485
|
| 144 |
+
142,0.04054571071292625,0.0016439546572163026
|
| 145 |
+
143,0.03112737826708608,0.0009689136777822628
|
| 146 |
+
144,0.035427435942357526,0.001255103217449846
|
| 147 |
+
145,0.05333297218241139,0.002844405921809867
|
| 148 |
+
146,0.027859026464513593,0.0007761253555504687
|
| 149 |
+
147,0.03535631923036233,0.0012500693095192892
|
| 150 |
+
148,0.03517370791973029,0.0012371897288224974
|
| 151 |
+
149,0.06669649107326957,0.004448421921486727
|
| 152 |
+
150,0.028184106319618075,0.0007943438490355355
|
| 153 |
+
151,0.037721738102602806,0.0014229295254813564
|
| 154 |
+
152,0.04348080368606685,0.0018905802891862843
|
| 155 |
+
153,0.03201439067522981,0.0010249212103062416
|
| 156 |
+
154,0.058860555183258015,0.003464564956481362
|
| 157 |
+
155,0.0348876600860388,0.001217148826278985
|
| 158 |
+
156,0.04344810180066125,0.0018877375500806232
|
| 159 |
+
157,0.074161392337401,0.00549991211342192
|
| 160 |
+
158,0.030911167060406634,0.0009555002490363681
|
| 161 |
+
159,0.07459009910270302,0.005563682884151057
|
| 162 |
+
160,0.03393567218958478,0.0011516298469589578
|
| 163 |
+
161,0.017839876754936014,0.00031826120263130634
|
| 164 |
+
162,0.046926178365305125,0.002202066215972431
|
| 165 |
+
163,0.030432033141839263,0.0009261086411460033
|
| 166 |
+
164,0.048521050177554315,0.0023542923103327435
|
| 167 |
+
165,0.0407406446298148,0.0016598001248528573
|
| 168 |
+
166,0.08872422389208104,0.007871987905252124
|
| 169 |
+
167,0.04044605904585056,0.0016358836923404299
|
| 170 |
+
168,0.044042097763706385,0.0019397063754278709
|
| 171 |
+
169,0.033109917554197135,0.0010962666404457316
|
| 172 |
+
170,0.039311947615215545,0.0015454292253014512
|
| 173 |
+
171,0.028095536003793508,0.0007893591433404572
|
| 174 |
+
172,0.03749832177601354,0.0014061241360174511
|
| 175 |
+
173,0.08915933327246019,0.007949386709589626
|
| 176 |
+
174,0.03261206559558627,0.0010635468224108216
|
| 177 |
+
175,0.036261938104917416,0.0013149281551248617
|
| 178 |
+
176,0.028306313451454262,0.0008012473812119805
|
| 179 |
+
177,0.02610009210128118,0.0006812148076953602
|
| 180 |
+
178,0.03643076740410866,0.001327200813652266
|
| 181 |
+
179,0.103343831162516,0.010679947439346613
|
| 182 |
+
180,0.056840195708563565,0.003230807848187808
|
| 183 |
+
181,0.06653221225332549,0.004426535267321555
|
| 184 |
+
182,0.03096807317776706,0.0009590215563435356
|
| 185 |
+
183,0.02900650491777696,0.000841377327545019
|
| 186 |
+
184,0.05859937002614709,0.003433886167461306
|
| 187 |
+
185,0.04550637436239198,0.002070830107610166
|
| 188 |
+
186,0.06294931749790952,0.003962616573452618
|
| 189 |
+
187,0.020884910103994285,0.0004361794700519226
|
| 190 |
+
188,0.06549126853552026,0.004289106254391626
|
| 191 |
+
189,0.058037373266787824,0.003368336695708458
|
| 192 |
+
190,0.04977924179520765,0.0024779729137057484
|
| 193 |
+
191,0.06814559565663965,0.004643822207398224
|
| 194 |
+
192,0.07202909632677316,0.005188190717651567
|
| 195 |
+
193,0.02930723595550002,0.0008589140793513532
|
| 196 |
+
194,0.033552314257589075,0.0011257577920400152
|
| 197 |
+
195,0.05067543875910043,0.0025680000934273383
|
| 198 |
+
196,0.0577903614592822,0.0033397258775944897
|
| 199 |
+
197,0.026609906953305125,0.0007080871480635565
|
| 200 |
+
198,0.03669344958595133,0.001346409242516752
|
| 201 |
+
199,0.04678646470539086,0.0021889732796287844
|
results/predictions.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9878a810d3386cbdbff4f423b164a0f16c245d7194cd821c16e39b3ea9bd956a
|
| 3 |
+
size 2463216
|
results/report.md
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# SirenFNO Table 2 / Darcy 32×32 复现结果
|
| 2 |
+
|
| 3 |
+
正式训练已完成 **500 epoch**,batch size 32,训练 1000 个、测试 200 个样本。
|
| 4 |
+
加载 `weight/best_model.pth` 后,原尺度逐样本相对 L2 算术均值为 **0.0452446841**。
|
| 5 |
+
对比论文 Table 2 的 **0.0351**:有符号差距 **+0.0101446841**,绝对差距 **0.0101446841**,相对差距 **+28.902%**。
|
| 6 |
+
|
| 7 |
+
## 指标与权重
|
| 8 |
+
|
| 9 |
+
- 主指标:对每个样本计算 `||prediction-y||_2 / ||y||_2`,对 200 个样本取算术均值;由保存的原尺度 float32 场使用 float64 重新计算。
|
| 10 |
+
- 论文式 (15) 字面上的平方相对 L2 均值:**0.0025980480**。它与主指标不是同一统计量。
|
| 11 |
+
- 最优权重选择:每个 epoch 后在完整训练集上以 eval 模式计算原尺度平方相对 L2,取最小者;选中 **epoch 500**。
|
| 12 |
+
- 测试集每轮用于监测,**不用于选权重、调参或早停**;无独立验证集。论文没有说明权重选择规则,本规则属于显式工程假设。
|
| 13 |
+
- 末轮 epoch 500 的原尺度测试相对 L2:**0.0452446839**;末轮权重保存在 `weight/last_model.pth`。
|
| 14 |
+
- checkpoint SHA256:`ceb2392d5ba68e386f08e0f117ea1705df1f3a0185d3cef0971f07951113ce1e`。
|
| 15 |
+
- 实际本轮训练墙钟时间:1321.41 秒;参数量:304801。
|
| 16 |
+
|
| 17 |
+
## 对齐情况与限制
|
| 18 |
+
|
| 19 |
+
1. 本次是非分解完整 SirenFNO,四层 32 通道,式 (14) 的增强残差结构,全部 32×17 个独立 rFFT 频点生成稠密 32×32 复系数;没有任何频率截断。rFFT 半谱是实数输入的厄米冗余压缩,不是高频截断。
|
| 20 |
+
2. 论文 §5.2 的 AdamW、lr=1e-3、weight decay=1e-4、CosineAnnealingLR(T_max=500)、500 epoch、batch=32 均执行。无提前停止、无降采样、无额外训练数据。预先冒烟训练的模型已丢弃,正式训练从 seed 42 重新初始化。
|
| 21 |
+
3. 论文式 (7) 明确写 sin,本次使用纯 sin 及可学习偏置;官方仓库使用交错 sin/cos 且无偏置,本次没有用该冲突实现覆盖论文。
|
| 22 |
+
4. 论文 §5.2 指定 lifting/projection 隐藏宽度 64,本次使用 3→64→32 和 32→64→1。仓库使用单层 3→32 lifting 和 32→128→1 projection;因此本次参数量不要求等于 Table 3 的约 308.9k。
|
| 23 |
+
5. 论文式 (15) 在 PDF 上是范数平方之比,但文字称相对 L2、仓库 LpLoss 使用未平方比值。本次训练在原尺度使用平方比值以遵循字面公式,报告同时给出用户指定的未平方指标。**无法断言 Table 2 的 0.0351 究竟对应公式还是仓库口径,因此数值差距只是指定口径下的直接比较,不代表已精确匹配作者训练实现。**
|
| 24 |
+
6. 论文未给出的 RFF 特征数 16、sigma=512、SIREN 隐藏宽度 32/一层、omega 初值、midpoint 频率索引、残差 MLP 扩张 4、坐标拼接、无 padding、seed 42、无 shuffle,以及训练集压力 channel-wise 标准化,来自官方仓库只读参考。输入渗透率保持原值;压力标准化统计量仅来自 1000 个训练样本,模型输出先反归一化再计算训练 loss 和评测指标。
|
| 25 |
+
7. 运行设备为 **BW**,PyTorch **2.4.1**,HIP **6.1.25065**;论文设备是 RTX 5090。后端和随机数流不同,只有一次固定 seed 运行,没有多种子统计。
|
| 26 |
+
8. 物理单位及精确边界网格元数据未在 pt 字段中提供,不虚构单位;可视化坐标采用官方输入网格 [0,1] 的约定,不宣称额外 PDE 残差或守恒量验证。
|
| 27 |
+
|
| 28 |
+
## 证据与文件
|
| 29 |
+
|
| 30 |
+
- `results/data_manifest.json`:原始 shape/dtype、选样索引、字段语义与数据 SHA256。
|
| 31 |
+
- `results/training_history.csv`:500 轮实际记录,含 train/test 两种误差及学习率。
|
| 32 |
+
- `results/per_sample_metrics.csv`:200 个测试样本原尺度误差。
|
| 33 |
+
- `results/predictions.pt`:原尺度输入、真值、预测、样本索引。
|
| 34 |
+
- `results/darcy_fields.png`、`darcy_fields.pdf`:固定第 0 个、误差中位样本、最差样本,真值与预测使用同一色标。
|
| 35 |
+
- `results/training_curves.png`、`error_distribution.png`:训练曲线及误差分布。
|
| 36 |
+
- `results/smoke/checks.json`:六项计算节点冒烟检查。
|
| 37 |
+
- `reproduction_spec.md`:完整来源、缺口与冲突决策;`logs/`:Slurm 实时 print 日志。
|
| 38 |
+
|
| 39 |
+
论文:[IJCAI PDF](https://ijcai.org/proceedings/2026/0535.pdf)。补充细节:[官方仓库只读参考](https://github.com/pengqingshi/SirenFNO)。未下载、克隆或复制官方仓库代码,所有实现均独立编写;未提交 Git、未上传平台。
|
| 40 |
+
|
| 41 |
+
## 恢复执行与补充核验
|
| 42 |
+
|
| 43 |
+
- 恢复时确认已有实现、配置和规格,但没有 checkpoint、训练记录或活跃作业;本次复用旧代码执行首次正式训练,没有重启已有训练。
|
| 44 |
+
- 实际优化器对每个参数完成 **16000 次更新**(500×32 个 batch,末 batch 为8个样本),调度器完成500步、最终lr=0;见 `budget_verification.json`。
|
| 45 |
+
- 独立 NumPy float64 公式从保存的预测/真值重算,均值与报告相差约2.1e-17,并逐行核对200个样本CSV;见 `independent_metric_verification.json`。
|
| 46 |
+
- 六项冒烟、正式训练、权重重载推理、可视化作业分别为1100629、1100639、1100802、1100831,均以0退出;节点与起止时间见 `execution_provenance.json`。
|
| 47 |
+
- **归一化补充限制**:论文未规定归一化细节。保存的官方只读参考记录显示,仓库训练在标准化目标上计算损失、评测时反归一化;本实现依据论文公式,在训练和评测前均反归一化到原始文件尺度计算损失。该差异已在 `reproduction_spec.md` 第7节记录,可能影响与Table2的数值对齐。原始文件未提供物理单位,本报告的“原尺度”指提供的pt文件中y的数值尺度。
|
| 48 |
+
- 目视检查了场图及训练曲线;所有模型、配置、业务脚本与正式运行前的SHA256快照一致。
|
results/run_metadata.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"parameters": 304801,
|
| 3 |
+
"torch": "2.4.1",
|
| 4 |
+
"hip": "6.1.25065",
|
| 5 |
+
"device": "BW",
|
| 6 |
+
"selection": "minimum_full_training_original_scale_squared_relative_l2",
|
| 7 |
+
"seed": 42,
|
| 8 |
+
"data_sha256": {
|
| 9 |
+
"train": "f44a1151802eaf2eac65ddaded1ecf105bac019add0e3d277a3e9939cca1cd8b",
|
| 10 |
+
"test": "f9fbe30065908327f132425d760d573ed7ce29cb2a54cfdb3fcefd3d80d7ac83"
|
| 11 |
+
}
|
| 12 |
+
}
|
results/smoke/checks.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"status": "passed",
|
| 3 |
+
"attempts": 1,
|
| 4 |
+
"checks": {
|
| 5 |
+
"forward_pass": "passed",
|
| 6 |
+
"backward_pass": "passed",
|
| 7 |
+
"train_loop_dry_run": "passed",
|
| 8 |
+
"validation_loop_dry_run": "passed",
|
| 9 |
+
"domain_metric_dry_run": "passed",
|
| 10 |
+
"config_consistency": "passed"
|
| 11 |
+
},
|
| 12 |
+
"used_original_data": true,
|
| 13 |
+
"formal_epochs_consumed": 0,
|
| 14 |
+
"loss_first8_mean": 0.39898883551359177,
|
| 15 |
+
"loss_last8_mean": 0.12611230090260506,
|
| 16 |
+
"validation_metric": {
|
| 17 |
+
"relative_l2_mean": 0.34784180857241154,
|
| 18 |
+
"squared_relative_l2_mean": 0.12593160066240605,
|
| 19 |
+
"relative_l2_std": 0.07026860516247534
|
| 20 |
+
},
|
| 21 |
+
"peak_memory_bytes": 299171328
|
| 22 |
+
}
|
results/training_curves.png
ADDED
|
Git LFS Details
|
results/training_history.csv
ADDED
|
@@ -0,0 +1,501 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
epoch,lr,online_train_loss,train_relative_l2,train_squared_relative_l2,test_relative_l2,test_squared_relative_l2,seconds,selected
|
| 2 |
+
1,0.001,0.23011872267723082,0.3313449652940035,0.11436280270746278,0.33834751531481744,0.11906420606449726,10.637119774241,True
|
| 3 |
+
2,0.0009999901304280686,0.09272673469781875,0.2653466761857271,0.07515520037045052,0.27276008918881417,0.07946665244583548,2.5584473996423185,True
|
| 4 |
+
3,0.0009999605221019082,0.07086053627729416,0.23112500841915606,0.05809211669267561,0.23803621970117092,0.06144539054462512,2.567289157304913,True
|
| 5 |
+
4,0.0009999111761904045,0.056780061021447184,0.21250988959521055,0.04947933003023007,0.21801984068006278,0.052038867872964474,2.568350072018802,True
|
| 6 |
+
5,0.00099984209464165,0.04404554235935211,0.18368097335845232,0.03736924471434293,0.18691234663128853,0.03842726008010049,2.5661276909522712,True
|
| 7 |
+
6,0.0009997532801828658,0.038149921357631685,0.18850703723728657,0.03926974653594289,0.19115304198116065,0.04001370033359918,2.5660480903461576,False
|
| 8 |
+
7,0.0009996447363202945,0.03753815668821335,0.17501267348229885,0.03418916916990967,0.17891598138958215,0.0353117180053839,2.5788639280945063,True
|
| 9 |
+
8,0.0009995164673390624,0.03635310712456703,0.18100457260012626,0.03645112963011417,0.184570836648345,0.0377027804445321,2.5655471668578684,False
|
| 10 |
+
9,0.0009993684783030088,0.03304100893437863,0.1717013074234128,0.03297880418153496,0.17486063357442616,0.03392605424427333,2.565937630366534,True
|
| 11 |
+
10,0.0009992007750544876,0.030527080081403254,0.16052984242886306,0.028992700680689255,0.16335388600826264,0.029659292919647598,2.565453337971121,True
|
| 12 |
+
11,0.0009990133642141358,0.028634587690234185,0.1537148814573884,0.02668526708814456,0.1562392330542207,0.027143336138605676,2.5643864809535444,True
|
| 13 |
+
12,0.0009988062531806126,0.02696247221529484,0.1474009471386671,0.024619586173992552,0.14953956130892038,0.024903248900163114,2.5656868680380285,True
|
| 14 |
+
13,0.000998579450130307,0.024853453531861306,0.1424821178689599,0.022989330559030616,0.1440153967961669,0.023087470824171525,2.564125149976462,True
|
| 15 |
+
14,0.0009983329640170149,0.022475376412272454,0.13573458659276366,0.02089229182049701,0.13705203250050546,0.02097217856970102,2.5624765218235552,True
|
| 16 |
+
15,0.0009980668045715864,0.020002353884279726,0.12299209954217076,0.017431149523581775,0.12475247064605355,0.01762039412598429,2.5555983437225223,True
|
| 17 |
+
16,0.0009977809823015402,0.018185377418994904,0.11549638846516609,0.01556001062274098,0.11772419154644012,0.01583498310091769,2.554186108056456,True
|
| 18 |
+
17,0.0009974755084906502,0.01747658094018698,0.11385663529857994,0.014921407561005675,0.11661886485293507,0.015358245512920008,2.5591375837102532,True
|
| 19 |
+
18,0.0009971503951984996,0.015716877847909927,0.1079341348670423,0.013594763377021148,0.11106691051274538,0.014094681253279985,2.5568320811726153,True
|
| 20 |
+
19,0.0009968056552600044,0.015207389019429684,0.10665695288032294,0.013062979042657632,0.11055480513721705,0.013772982775235818,2.5601223949342966,True
|
| 21 |
+
20,0.000996441302284907,0.013726603664457798,0.10119331005960702,0.01187589117991735,0.10524132367223502,0.01259752191497802,2.5625786278396845,True
|
| 22 |
+
21,0.0009960573506572392,0.012878906369209289,0.09864396734163165,0.011271918145996197,0.10310432158410548,0.012098707345527442,2.5578143829479814,True
|
| 23 |
+
22,0.0009956538155347534,0.012028315879404545,0.09578740574792027,0.010613928927049503,0.10047760955989361,0.011499635919577362,2.5696868849918246,True
|
| 24 |
+
23,0.0009952307128483257,0.011130176071077585,0.09344548606872559,0.010098032019911081,0.0982172548212111,0.011022261234840638,2.5650640511885285,True
|
| 25 |
+
24,0.0009947880593013255,0.010159365974366666,0.09172933765500783,0.009682830416054618,0.0965478610806167,0.010645707911082021,2.5664974190294743,True
|
| 26 |
+
25,0.000994325872368957,0.009050240620970726,0.08651496526226402,0.008484483575536607,0.09241844274103642,0.009615927353324767,2.5640232460573316,True
|
| 27 |
+
26,0.0009938441702975688,0.008296732291579246,0.08214046122133732,0.007585931270728024,0.08869717972353101,0.0088355714402354,2.5804659631103277,True
|
| 28 |
+
27,0.000993342972103934,0.007823935553431511,0.08008148867636919,0.0071156572021768135,0.08733845133334399,0.008501054132644535,2.5641930461861193,True
|
| 29 |
+
28,0.000992822297574499,0.007407028742134571,0.07843859727680683,0.006766572616093959,0.08608081478625536,0.00824259368473393,2.5649724109098315,True
|
| 30 |
+
29,0.0009922821672646027,0.007098493181169033,0.07709385548904538,0.006498218564253547,0.08529313107952476,0.008105650271393691,2.563590803183615,True
|
| 31 |
+
30,0.0009917226024976648,0.006792068462818861,0.07640401224792004,0.0063566736215337144,0.08502634096890688,0.008065226625626252,2.584067951887846,True
|
| 32 |
+
31,0.0009911436253643444,0.0065291008725762364,0.07638454765826463,0.006344345648532562,0.08541791243478655,0.008167257390649253,2.566741110291332,True
|
| 33 |
+
32,0.000990545258721667,0.006372505594044924,0.0775351991765201,0.006560042032394595,0.08697620380669832,0.008517486757903559,2.5662344470620155,False
|
| 34 |
+
33,0.0009899275261921233,0.0065824190191924575,0.08909605153277517,0.008926885497696223,0.09880736442282796,0.011262517980862823,2.5640262458473444,False
|
| 35 |
+
34,0.000989290452162736,0.00708465974777937,0.0812064827606082,0.007286104941220376,0.09130397884175181,0.009452440044371365,2.563000010792166,False
|
| 36 |
+
35,0.0009886340617840967,0.007088577054440976,0.07718623568490147,0.006427109248746076,0.08853087436407804,0.008765663845867844,2.5649240012280643,False
|
| 37 |
+
36,0.0009879583809693736,0.0058493631891906265,0.06834484524652362,0.0051123807241455505,0.07853390166535973,0.006964983021466384,2.5573770911432803,True
|
| 38 |
+
37,0.0009872634363932884,0.005644763328135014,0.06862898639962077,0.005136646521128287,0.07937402920797468,0.0071130763536549615,2.556695126928389,False
|
| 39 |
+
38,0.0009865492554910632,0.005600021257996559,0.07189181484654546,0.005697156392669314,0.08119258103892207,0.0074690145024681135,2.55634384509176,False
|
| 40 |
+
39,0.0009858158664573368,0.006798306778073311,0.09290809835493564,0.00978106917827869,0.09687225708737969,0.010986638234890154,2.5670784218236804,False
|
| 41 |
+
40,0.0009850632982450526,0.009975432112812996,0.08433684770017863,0.008012980923287798,0.08924210701137782,0.00919040992072401,2.565325072966516,False
|
| 42 |
+
41,0.0009842915805643151,0.009511774022132158,0.08689182179421187,0.008277108317584783,0.09331560857594014,0.009606622255980347,2.560328036081046,False
|
| 43 |
+
42,0.000983500743881217,0.007918468572199345,0.09561714918911457,0.010042174099817643,0.10066664267331361,0.011278070868634695,2.566343978047371,False
|
| 44 |
+
43,0.0009826908194166364,0.008631027348339558,0.12638170412927865,0.017258614525192208,0.12851507680490612,0.01830723455791229,2.567069191019982,False
|
| 45 |
+
44,0.0009818618391450045,0.009168853655457497,0.08237001019716263,0.007635269041176034,0.08768096894025802,0.008969075281159162,2.5661686160601676,False
|
| 46 |
+
45,0.0009810138357930424,0.006197604700922966,0.07121317290514707,0.00551478846733437,0.07910907637327909,0.007138805648724589,2.577276024967432,False
|
| 47 |
+
46,0.000980146842838471,0.004631121296435594,0.07024689945206046,0.005325265214593349,0.07819172654300928,0.00694365778252882,2.5626429971307516,False
|
| 48 |
+
47,0.0009792608945086874,0.004242521220818162,0.06893047893047333,0.005106812424659913,0.07714236287400127,0.006740106084100235,2.5692511522211134,True
|
| 49 |
+
48,0.0009783560257794147,0.0039662871323525905,0.06772581718862057,0.004926594009919021,0.07613348290324211,0.006571850415333957,2.5668320190161467,True
|
| 50 |
+
49,0.000977432272373321,0.003778002068400383,0.06571932772547007,0.004639221295763319,0.07444882474839687,0.006304002215480147,2.5788921830244362,True
|
| 51 |
+
50,0.0009764896707586089,0.0036163107044994832,0.06398156348988414,0.004396789047618952,0.0730096984282136,0.006082700679720985,2.5665050856769085,True
|
| 52 |
+
51,0.0009755282581475763,0.0034658964537084102,0.06259288713149726,0.004206312079450859,0.07184979507699608,0.005910127386539858,2.5654852921143174,True
|
| 53 |
+
52,0.0009745480724951467,0.003326105350628495,0.061422554943710565,0.004047918770761926,0.07087761718779802,0.005768117134584389,2.567937864921987,True
|
| 54 |
+
53,0.0009735491524973714,0.003193105174228549,0.06032400604337454,0.0039019108824363017,0.06997125558555126,0.005636618435528611,2.570183396805078,True
|
| 55 |
+
54,0.0009725315375899016,0.0030714216362684966,0.05926784256845713,0.0037635617112741137,0.06911215477623045,0.005513040679747296,2.564647426828742,True
|
| 56 |
+
55,0.0009714952679464315,0.00295872110966593,0.0582781183514744,0.0036367954644797302,0.06831070385873318,0.005399581395247671,2.5671348506584764,True
|
| 57 |
+
56,0.000970440384477112,0.002857062103226781,0.05734957391954958,0.0035205242403543017,0.06755908012390137,0.005295757122814651,2.5641720262356102,True
|
| 58 |
+
57,0.0009693669288269363,0.002765212352387607,0.0565040982067585,0.003417056207375184,0.06687436548061669,0.005203745199025191,2.564331886358559,True
|
| 59 |
+
58,0.0009682749433740955,0.0026846351865679025,0.05579880888201296,0.0033328831865564183,0.06628739179112017,0.005128157788721371,2.564931548666209,True
|
| 60 |
+
59,0.0009671644712283052,0.002615703874267638,0.05528688405454159,0.0032734406473947015,0.065841680848971,0.005074429778913127,2.5639375438913703,True
|
| 61 |
+
60,0.0009660355562291046,0.0025605580881237983,0.05496132144890726,0.0032372715535116804,0.06553709450177848,0.005042666081973659,2.5646593081764877,True
|
| 62 |
+
61,0.0009648882429441249,0.0025179629307240247,0.05517714847438037,0.0032667349080528524,0.06559456217102706,0.005065417576887594,2.558146263938397,False
|
| 63 |
+
62,0.0009637225766673299,0.002493561667390168,0.05554283691197634,0.0033152134017309627,0.06578504736535251,0.005108264372228067,2.55665340507403,False
|
| 64 |
+
63,0.0009625386034172282,0.0024948015697300436,0.056775613414123655,0.003473362943957692,0.06661032926291227,0.005252409117027897,2.5567756961099803,False
|
| 65 |
+
64,0.0009613363699350566,0.002542872468009591,0.06040049103833735,0.003952095487754304,0.06921572622843086,0.005695192613618667,2.592729837168008,False
|
| 66 |
+
65,0.0009601159236829344,0.0027937439568340777,0.07229453689604998,0.005735596862681839,0.07869159342721105,0.007420177648968174,2.572648261208087,False
|
| 67 |
+
66,0.0009588773128419897,0.004461058864369988,0.09648860345780849,0.010916196082539303,0.09929459610953927,0.012201429271516717,2.5659818947315216,False
|
| 68 |
+
67,0.0009576205863104579,0.006316237011924386,0.07198153082281351,0.00549384424372652,0.07746704783290624,0.006646920841304725,2.5773275452665985,False
|
| 69 |
+
68,0.0009563457937017506,0.006290275786072016,0.06581961356475949,0.00456651603570181,0.07524818882346153,0.006428195265018689,2.566531514748931,False
|
| 70 |
+
69,0.000955052985342497,0.00492642025463283,0.06043841591104865,0.003816172209196845,0.07265109756961464,0.006014819360397514,2.565406625159085,False
|
| 71 |
+
70,0.0009537422122705576,0.004248074799776077,0.07024021013081073,0.005277131567339625,0.07865529436618089,0.007180262318183496,2.566366230137646,False
|
| 72 |
+
71,0.0009524135262330089,0.003395844019949436,0.09036575805768371,0.008560805294228029,0.09507007714360953,0.010152810782464298,2.569888069294393,False
|
| 73 |
+
72,0.0009510669796841005,0.004332255113869906,0.10401228494942188,0.011248528663240148,0.1070145303569734,0.012635497252560295,2.571131265256554,False
|
| 74 |
+
73,0.0009497026257831847,0.007285923287272453,0.07350741055607796,0.005900196245600507,0.08067413456737996,0.007623894851657226,2.567106891889125,False
|
| 75 |
+
74,0.0009483205183926171,0.008228020027279855,0.06229212472960353,0.004254720906436664,0.07714805144816637,0.006955963150764625,2.5681875301524997,False
|
| 76 |
+
75,0.0009469207120756311,0.004342329774051905,0.05671875550039113,0.003389920542466947,0.06855658011510968,0.005378306408144293,2.572283540852368,False
|
| 77 |
+
76,0.000945503262094183,0.0028249176982790233,0.05660701757110655,0.003348441078470282,0.0677539779432118,0.005280363049231077,2.5649589826352894,False
|
| 78 |
+
77,0.0009440682244067714,0.0029501428902149202,0.05304657864570618,0.0029458419770979597,0.06518665011972188,0.004996913668993484,2.55882972991094,True
|
| 79 |
+
78,0.0009426156556662268,0.003218168534338474,0.0505450551584363,0.0027580860184071646,0.06463644220493733,0.005148668449491891,2.5568398595787585,True
|
| 80 |
+
79,0.0009411456132174758,0.00362340422347188,0.054512755244970325,0.0033141520171982323,0.06936905906535686,0.006002311347152205,2.5677423467859626,False
|
| 81 |
+
80,0.0009396581550952772,0.003981819273903966,0.06017931488156319,0.0037966259372307406,0.07195701733231545,0.00587021080218122,2.5655872458592057,False
|
| 82 |
+
81,0.0009381533400219309,0.0032400401588529347,0.06971435874700546,0.005055979984084403,0.07937380183488131,0.006930851191600564,2.567250625230372,False
|
| 83 |
+
82,0.0009366312274049592,0.003012609675526619,0.07106486460193992,0.0052612572310802306,0.08049739185720682,0.0071804770151383665,2.5671055829152465,False
|
| 84 |
+
83,0.0009350918773347619,0.0032811715472489595,0.059506516922265294,0.003686018395697716,0.07136778578162194,0.0057669525610124315,2.5660532587207854,False
|
| 85 |
+
84,0.0009335353505822441,0.00359749561175704,0.05765139153786004,0.0035854034208251773,0.07177492247894407,0.00613602020920287,2.5636062948033214,False
|
| 86 |
+
85,0.0009319617085964167,0.0042092655822634694,0.06910787498578429,0.004942844759854384,0.0804097437299788,0.007069614872448464,2.567337095271796,False
|
| 87 |
+
86,0.000930371013501971,0.0037197784893214705,0.07666838195547461,0.006064186015377886,0.08689219016581774,0.008125043198699161,2.5585057181306183,False
|
| 88 |
+
87,0.0009287633280968253,0.0034869415182620287,0.07240271708741784,0.005421708576297873,0.08332995932549238,0.007535561394992958,2.5963355489075184,False
|
| 89 |
+
88,0.0009271387158496467,0.0035653607323765757,0.06809174473956227,0.0048237982641851015,0.07922071859240531,0.0068902589643955415,2.5720575801096857,False
|
| 90 |
+
89,0.0009254972408973451,0.003764865198172629,0.06294671488925814,0.0041993625913226236,0.07415258904919028,0.0061543551661394245,2.5581151959486306,False
|
| 91 |
+
90,0.0009238389680425407,0.0038044645097106695,0.06054220035299659,0.003969701710952109,0.07124870156869292,0.00585412880268496,2.572791294194758,False
|
| 92 |
+
91,0.0009221639627510066,0.00391887860186398,0.05276982384175062,0.0030271103626080074,0.06465587674640119,0.004948504324154538,2.565351014956832,False
|
| 93 |
+
92,0.0009204722911490836,0.004697892338037491,0.062175781587138775,0.004128919479655477,0.07298047534190118,0.006087471290378588,2.56685250159353,False
|
| 94 |
+
93,0.0009187640200210698,0.006488717440515756,0.09276506761461496,0.009265344681706214,0.10170121101662516,0.01153275932744783,2.5666527478024364,False
|
| 95 |
+
94,0.0009170392168065845,0.0068957748413085935,0.06487252815812826,0.004472793230738061,0.07562839103862644,0.0065593371523309015,2.572520449757576,False
|
| 96 |
+
95,0.0009152979495979053,0.0046047277152538296,0.0849772990271449,0.007878000344514295,0.09792011609300971,0.011052956098313374,2.563144620973617,False
|
| 97 |
+
96,0.0009135402871372799,0.00599562562070787,0.10200192229077219,0.011263287867106796,0.11458894463256002,0.014879526764548089,2.557912182994187,False
|
| 98 |
+
97,0.0009117662988142127,0.007366633247584104,0.05845916881412268,0.003702500239911418,0.07256522361189127,0.006093240898818542,2.558225834276527,False
|
| 99 |
+
98,0.0009099760546627253,0.005252546552568674,0.05764118300564587,0.003576790003448981,0.07035216618329286,0.005729343871506531,2.561742043122649,False
|
| 100 |
+
99,0.000908169625358591,0.0031647010073065757,0.04945677692256868,0.0026448826946244364,0.06361635521985591,0.0048323838673077985,2.5688918423838913,True
|
| 101 |
+
100,0.0009063470822165459,0.002321729678660631,0.04494553457200527,0.0021290605895557587,0.06042539575137198,0.004297627102938149,2.564945090096444,True
|
| 102 |
+
101,0.0009045084971874727,0.0019785691322758794,0.04458463542535901,0.0020751590992924463,0.06038989895954728,0.004326523240840677,2.565557074267417,True
|
| 103 |
+
102,0.0009026539428555598,0.0019382316265255212,0.048224121415987614,0.002439203221607464,0.0635453556291759,0.00484348019519756,2.5640631350688636,False
|
| 104 |
+
103,0.0009007834924354372,0.002006389127112925,0.051647963317111134,0.0028108338157242707,0.06679703208617865,0.0053501306229561596,2.565298622008413,False
|
| 105 |
+
104,0.0008988972197692843,0.00208973485045135,0.052934815041720866,0.002954547856886289,0.06827873327769339,0.00557654297920148,2.563653592951596,False
|
| 106 |
+
105,0.0008969951993239164,0.0021761700138449667,0.05147438714466989,0.002789936317678487,0.06737231488339603,0.005420451213398165,2.581804800312966,False
|
| 107 |
+
106,0.0008950775061878439,0.002260584242641926,0.04790959915146232,0.0024125506153156637,0.06459802624769509,0.004973917896597238,2.5641098362393677,False
|
| 108 |
+
107,0.0008931442160683082,0.002352035300806165,0.04455481308139861,0.00209668433649712,0.06169311793521046,0.004512286474203973,2.556839626748115,False
|
| 109 |
+
108,0.0008911954052882928,0.0024967999421060087,0.04629314012080431,0.0023106446729462826,0.062109901877120136,0.004549214463119682,2.5560998329892755,False
|
| 110 |
+
109,0.0008892311507835107,0.0027833748906850816,0.05598600714094937,0.0034112122789541513,0.06888698033988476,0.005548154825658836,2.5554502196609974,False
|
| 111 |
+
110,0.0008872515300993659,0.003325183743610978,0.06680453371070326,0.0048110562373574976,0.0781294091232121,0.006998681463494793,2.5568561367690563,False
|
| 112 |
+
111,0.0008852566213878935,0.0037577109895646573,0.056450067210942506,0.0034186340490421296,0.06949354998767376,0.005631050552974733,2.565683573950082,False
|
| 113 |
+
112,0.0008832465034046738,0.003418859455734491,0.046306167675182225,0.002313030824033781,0.060862105805426836,0.004520897484898333,2.5637438031844795,False
|
| 114 |
+
113,0.0008812212555057229,0.002875426998361945,0.0592359892167151,0.0038569010027558795,0.07255039244890213,0.0065326626380601705,2.573260175064206,False
|
| 115 |
+
114,0.0008791809576443599,0.002897406838834286,0.06469571809098125,0.004484687778775596,0.07835261758416891,0.007302857140832524,2.56515732081607,False
|
| 116 |
+
115,0.0008771256903680509,0.0026555989496409894,0.056001394456252455,0.0033109602627416136,0.07087390111759305,0.005941839368901835,2.5641540959477425,False
|
| 117 |
+
116,0.0008750555348152287,0.0024452983383089306,0.04479455678910017,0.002100610324783961,0.0615457974281162,0.004496992437285598,2.56564231403172,False
|
| 118 |
+
117,0.00087297057271209,0.002305742573924363,0.04385874695517123,0.0020359754546431024,0.06038989811204374,0.0042941245112627716,2.5747730638831854,True
|
| 119 |
+
118,0.0008708708863693686,0.002304076423868537,0.04654554572328925,0.002315042787110609,0.061935571152716876,0.004508505423898416,2.5635759262368083,False
|
| 120 |
+
119,0.0008687565586790859,0.0023067719973623754,0.04612204563058913,0.0022493374500234258,0.06118261215277016,0.004398557106583197,2.5594692332670093,False
|
| 121 |
+
120,0.000866627673111279,0.002169262832030654,0.042407334480434655,0.0018816768231413534,0.05793659233488142,0.004016448837074448,2.5583887579850852,True
|
| 122 |
+
121,0.0008644843137107047,0.001998709068633616,0.041619008278474215,0.0018467326942782254,0.05749307120218873,0.004105324248581526,2.56917180493474,True
|
| 123 |
+
122,0.0008623265650935222,0.0019737559892237188,0.04753740469738841,0.0024730879583185887,0.06251963838003576,0.004972565604673215,2.5642844601534307,False
|
| 124 |
+
123,0.0008601545124439522,0.002051879582926631,0.0590143783595413,0.0038037218633963066,0.07285517798736692,0.006601590664012555,2.563020111992955,False
|
| 125 |
+
124,0.0008579682415109144,0.002230637283064425,0.07132831915467977,0.005443920644765413,0.084647865742445,0.00852487874675281,2.5629475922323763,False
|
| 126 |
+
125,0.0008557678386046416,0.0025459462180733683,0.07033196802996099,0.005243348067578054,0.0844033139012754,0.008365746309530547,2.559733154717833,False
|
| 127 |
+
126,0.0008535533905932725,0.002910746292211115,0.05160014964267611,0.002820814476006632,0.06792762408033014,0.005530283231342954,2.5564482579939067,False
|
| 128 |
+
127,0.0008513249848994235,0.0033181062368676066,0.04863003455661237,0.0025394699252987134,0.06386574110947549,0.004787606195850335,2.5572719518095255,False
|
| 129 |
+
128,0.0008490827094967353,0.003529355271719396,0.05345786305889487,0.0030837449548511285,0.06758213767781854,0.005347123233467555,2.569212664850056,False
|
| 130 |
+
129,0.0008468266529064013,0.003457077255472541,0.05083655494637787,0.002705674395344734,0.06488923255354166,0.004966362544679093,2.5610758722759783,False
|
| 131 |
+
130,0.0008445569041936731,0.002674631109461188,0.06012633798830211,0.003952960342235855,0.07322019881568849,0.006557275421890556,2.5604353290982544,False
|
| 132 |
+
131,0.0008422735529643432,0.002421605519019067,0.07150508976727724,0.00541773022684191,0.08423949820920824,0.008323779696153303,2.5661676288582385,False
|
| 133 |
+
132,0.0008399766893612084,0.0024015842759981753,0.05533353518694639,0.0032375014896766246,0.07053908867761492,0.005971751296592784,2.564624140970409,False
|
| 134 |
+
133,0.0008376664040605111,0.0023487880118191243,0.04454201533831656,0.0020816233188842913,0.061218437617644665,0.004495356383458814,2.5682048401795328,False
|
| 135 |
+
134,0.000835342788268359,0.002401053560897708,0.048620478712022304,0.0024787438823578685,0.06373412641696632,0.004739508619971925,2.566875282675028,False
|
| 136 |
+
135,0.0008330059337171247,0.0023384863398969173,0.058906867761164904,0.003698345506313343,0.07165288239717484,0.005957873104005907,2.5672204750590026,False
|
| 137 |
+
136,0.0008306559326618247,0.0021716753821820023,0.07139826677367091,0.005424277509321977,0.08330079357139766,0.008068086559712638,2.5764447231777012,False
|
| 138 |
+
137,0.000828292877876477,0.002320586549118161,0.07292797410488129,0.0056397750903267145,0.08562430528923869,0.008529039800173508,2.5654991259798408,False
|
| 139 |
+
138,0.0008259168626504383,0.0025702017433941364,0.0653750497289002,0.004562014140932005,0.07913765789940953,0.007300067979396013,2.565485275350511,False
|
| 140 |
+
139,0.0008235279807847211,0.002685874971561134,0.06291632586158812,0.004256562069207298,0.07647752169519663,0.00673464016078007,2.565181344281882,False
|
| 141 |
+
140,0.0008211263265882911,0.0027582657895982264,0.06351043319702149,0.004338942153026106,0.07664929415099322,0.006742891751819146,2.578418724704534,False
|
| 142 |
+
141,0.0008187119948743437,0.002901010524481535,0.0750162251777947,0.006090360371631153,0.0858969434350729,0.0086118754755438,2.570732394233346,False
|
| 143 |
+
142,0.0008162850809565612,0.0030432240422815083,0.07721033032611012,0.006540166619918811,0.08749595824629068,0.008916595454368459,2.5670142038725317,False
|
| 144 |
+
143,0.0008138456806453491,0.0031535631120204926,0.07833938736468554,0.0067376856677349465,0.08880273655056953,0.009100286602261327,2.568365128710866,False
|
| 145 |
+
144,0.0008113938902440552,0.0033422199012711645,0.07407825278863311,0.006091056082909712,0.08423359654843807,0.008351590390706334,2.583665191195905,False
|
| 146 |
+
145,0.0008089298065451661,0.004505084127187729,0.06335224427841604,0.004488860815462415,0.07395936897955835,0.006259644342419929,2.5673707742244005,False
|
| 147 |
+
146,0.0008064535268264872,0.005564936868846416,0.0719777609705925,0.005516104437252477,0.08106405237689614,0.00718742126504786,2.5634122230112553,False
|
| 148 |
+
147,0.0008039651488473016,0.004950250627472997,0.056541126012802125,0.003437991742914221,0.0675853172224015,0.005254394823290528,2.5673419553786516,False
|
| 149 |
+
148,0.0008014647708445113,0.004529491044580936,0.06963970717787743,0.005082283642990748,0.08127224791795015,0.00741577204818813,2.565632314886898,False
|
| 150 |
+
149,0.0007989524915287584,0.005275851972401142,0.06936435543000698,0.004937248411874615,0.08021623406559229,0.006910119467698683,2.565884836949408,False
|
| 151 |
+
150,0.0007964284100805285,0.004542543645948171,0.06524617022275925,0.004399939599005399,0.07765598552301527,0.006468464707781995,2.563049850985408,False
|
| 152 |
+
151,0.0007938926261462355,0.0035108226193115116,0.052417761605232954,0.0028212797733868753,0.06687991369515657,0.004928196154304326,2.563915536273271,False
|
| 153 |
+
152,0.0007913452398342869,0.0026610688734799623,0.041595497308298944,0.001815408435992122,0.057506346348673104,0.0039607220028409305,2.571034293156117,True
|
| 154 |
+
153,0.0007887863517111326,0.0022670710273087025,0.04123623658716678,0.0017953470992900012,0.05680535746738315,0.003937248799776139,2.5677480855956674,True
|
| 155 |
+
154,0.0007862160627972942,0.0027529502706602218,0.047083343924954536,0.0023524671178908924,0.06161165999248624,0.004537401369368094,2.564458919223398,False
|
| 156 |
+
155,0.0007836344745633772,0.0036485069077461956,0.06152480845153332,0.00404625448849949,0.07318298475816847,0.0062167167554535085,2.562280527781695,False
|
| 157 |
+
156,0.0007810416889260641,0.00434305989742279,0.0723422144278884,0.005577715828505083,0.08239123642444611,0.007793893978895211,2.5645690700039268,False
|
| 158 |
+
157,0.0007784378082440928,0.004569766715168953,0.08492628109827638,0.00778153068001674,0.0928861551731825,0.009897506215812625,2.5666019809432328,False
|
| 159 |
+
158,0.000775822935314214,0.005842843625694513,0.08802531189098954,0.008315405471193107,0.0956279425136745,0.010248480252996302,2.564291798043996,False
|
| 160 |
+
159,0.0007731971733671334,0.006197527766227722,0.06705365975759923,0.004820280277166362,0.07720101546496153,0.006555657128124128,2.5664284992963076,False
|
| 161 |
+
160,0.0007705606260634367,0.006028309971094131,0.0583452924862504,0.0036838437651165955,0.07192955257371068,0.005927244088706819,2.565605904906988,False
|
| 162 |
+
161,0.000767913397489497,0.004039040863513947,0.06373949460312724,0.00431895852853562,0.07569510873407126,0.006741226495998291,2.564181548077613,False
|
| 163 |
+
162,0.0007652555921533658,0.0027771076615899803,0.04699879749119282,0.0023875007514438246,0.062387252459302545,0.00478433054384667,2.5609922111034393,False
|
| 164 |
+
163,0.0007625873149806466,0.0020246254112571476,0.03891520436666906,0.0015633825842016173,0.057203522427007554,0.003860296738096637,2.5742316809482872,True
|
| 165 |
+
164,0.0007599086713103534,0.0018230886310338975,0.0423587240986526,0.001863322916709173,0.05926284778863192,0.004123491043202661,2.556215805001557,False
|
| 166 |
+
165,0.0007572197668907521,0.0018959095515310763,0.04343555142730474,0.001954391004851718,0.059148823032155635,0.004059657269366301,2.5646121399477124,False
|
| 167 |
+
166,0.0007545207078751845,0.0018097144924104213,0.04100180636905134,0.0017374492692332137,0.057539321184158325,0.0038174935593853045,2.5642644576728344,False
|
| 168 |
+
167,0.0007518116008178792,0.0015863050715997815,0.03848011053353548,0.001529809185730683,0.05629147357307374,0.0037292645996419077,2.5711121428757906,True
|
| 169 |
+
168,0.0007490925526697443,0.0014099249951541424,0.0380998870972544,0.001505120148948887,0.056287946850061416,0.0038132571154819174,2.5665016788989305,True
|
| 170 |
+
169,0.0007463636707741447,0.0012875484726391734,0.03818449300155043,0.0015133992287407324,0.05634614314883948,0.0038699371874074103,2.5638968446291983,False
|
| 171 |
+
170,0.000743625062862665,0.0012089688386768102,0.03780013648234308,0.0014824887117285091,0.05598388854414225,0.0038404489105422802,2.5714907348155975,True
|
| 172 |
+
171,0.0007408768370508564,0.001166842669248581,0.036876399341970685,0.0014104407224632175,0.05527525831013918,0.0037471204598522865,2.5639885747805238,True
|
| 173 |
+
172,0.0007381191018339683,0.0011499176137149333,0.03566514201834798,0.0013188070463397055,0.05444085182622075,0.0036303310614157164,2.5648754513822496,True
|
| 174 |
+
173,0.0007353519660826651,0.0011495589222759008,0.034547761227935556,0.0012357537669198393,0.05373872086405754,0.0035262438324042333,2.5636824341490865,True
|
| 175 |
+
174,0.000732575539038728,0.0011621408527716995,0.033833251601085065,0.0011818996603378297,0.05332844975404441,0.0034549915408435233,2.556303574703634,True
|
| 176 |
+
175,0.0007297899303107428,0.0011890802592970432,0.0338207793366164,0.001177474705864586,0.053359112357720735,0.00343575995270918,2.557042147964239,True
|
| 177 |
+
176,0.0007269952498697721,0.001232922739814967,0.034715239476412536,0.0012389581067734875,0.05392167722806335,0.0034799802766338466,2.557080059312284,False
|
| 178 |
+
177,0.0007241916080450151,0.0012995348814874887,0.03655183805152774,0.0013754544715358277,0.055029450133442875,0.0035932047034189594,2.5552678988315165,False
|
| 179 |
+
178,0.0007213791155194495,0.00139394820202142,0.03905606634169817,0.0015745927813135036,0.05649909373372793,0.0037555151648777396,2.55843588616699,False
|
| 180 |
+
179,0.0007185578833254652,0.0015167739018797873,0.04188086572475731,0.001813850909075158,0.0581766969896853,0.003945645507940065,2.5628143190406263,False
|
| 181 |
+
180,0.0007157280228404783,0.0016557110799476505,0.04464122118987143,0.0020611794183910473,0.05995890486985445,0.004147875939996516,2.5607638489454985,False
|
| 182 |
+
181,0.0007128896457825352,0.001781119652092457,0.046358875265344976,0.0022243812341441204,0.06119547558017075,0.004286785781888787,2.5589045989327133,False
|
| 183 |
+
182,0.000710042864205902,0.0018715443275868892,0.04593886970356107,0.0021923204818521847,0.0609455035906285,0.004247727270397255,2.5724099799990654,False
|
| 184 |
+
183,0.000707187790496641,0.0019349660677835346,0.044146000392735,0.002036023115705846,0.05971268271096051,0.004096163869930317,2.5584391462616622,False
|
| 185 |
+
184,0.0007043245373681734,0.0019895951729267834,0.04353047945536673,0.001980217017813068,0.05945595348253846,0.004068101809929533,2.556890548206866,False
|
| 186 |
+
185,0.0007014532178568302,0.0020821481058374046,0.046468441916629674,0.0022559140071873006,0.062137915706261995,0.004421740661572985,2.5585868270136416,False
|
| 187 |
+
186,0.0006985739453173892,0.0023500415245071055,0.0540913973134011,0.003118367440495299,0.06882547804154455,0.00544355566703149,2.5821426128968596,False
|
| 188 |
+
187,0.0006956868334186001,0.002981441413983703,0.0689264428075403,0.0051200711466051225,0.08125413938425481,0.0076753324872332685,2.5582127142697573,False
|
| 189 |
+
188,0.000692791996138697,0.0036926783360540865,0.07349684016406537,0.005873554421205387,0.08462404182180762,0.008519925378836835,2.5569930281490088,False
|
| 190 |
+
189,0.0006898895477608994,0.0035836106445640325,0.06922404375113547,0.005457205232100821,0.08250449612736702,0.008676464592780065,2.5580810350365937,False
|
| 191 |
+
190,0.000686979602868899,0.0028815331766381858,0.06348451809957624,0.004308205786954352,0.07580537620931864,0.006771582509377738,2.5795458182692528,False
|
| 192 |
+
191,0.0006840622763423378,0.002502119220793247,0.05552709472179413,0.003240227024487278,0.06959647281095385,0.00559444483047818,2.5639403671957552,False
|
| 193 |
+
192,0.0006811376833522717,0.0030431882124394177,0.04243019798770547,0.0019050049166067144,0.06032578905113042,0.004295862465825301,2.560951924882829,False
|
| 194 |
+
193,0.0006782059393566241,0.0035487843975424765,0.05295029836334288,0.0029656734195567347,0.06744294316507876,0.005297469702667927,2.566629745066166,False
|
| 195 |
+
194,0.0006752671600956284,0.002913558840751648,0.053661274489015344,0.003023302372064311,0.06742949430830776,0.005202984992424779,2.5665755746886134,False
|
| 196 |
+
195,0.0006723214615872575,0.002614184046164155,0.0450010378677398,0.002116227761780131,0.06178952928632498,0.00434447583413906,2.5664272750727832,False
|
| 197 |
+
196,0.0006693689601226446,0.002185130486264825,0.04786678860709071,0.002450208385704656,0.06432269608601927,0.00483574052413007,2.5634320778772235,False
|
| 198 |
+
197,0.0006664097722614922,0.0018938513938337564,0.06032139373384416,0.0038498144295978786,0.0739873921405524,0.006536114365557986,2.566688686143607,False
|
| 199 |
+
198,0.0006634440148274703,0.0018977203220129012,0.04725954243354499,0.002351129737825992,0.06362457357347012,0.0049322605254287,2.565787401050329,False
|
| 200 |
+
199,0.0006604718049036037,0.0018860683487728238,0.03508323007449508,0.0012737560363575437,0.0552078515291214,0.0036409751955937685,2.563431649003178,False
|
| 201 |
+
200,0.0006574932598276514,0.0018591785449534656,0.04117560114711523,0.0017592335961278485,0.05900303614325821,0.004044019488295096,2.568869097623974,False
|
| 202 |
+
201,0.0006545084971874728,0.0018069933149963617,0.041262733604758975,0.0017684574812060932,0.05938800456933677,0.004025115995073856,2.564959306269884,False
|
| 203 |
+
202,0.0006515176348163861,0.0016340033793821932,0.04492695078998804,0.0021366439889944655,0.06252396067604422,0.004559191580260087,2.5628934958949685,False
|
| 204 |
+
203,0.0006485207907885165,0.0014624451720155775,0.05268226587772369,0.0029076504560378164,0.06806401945650578,0.005514330732939684,2.5675286008045077,False
|
| 205 |
+
204,0.000645518083414135,0.0014314320739358664,0.04850710329413414,0.002456968943016047,0.06465204635635019,0.005046540136110427,2.566754206083715,False
|
| 206 |
+
205,0.000642509631234987,0.001416117798537016,0.03756713351607323,0.0014726659046223362,0.05678184727206826,0.003891617166458425,2.570998759008944,False
|
| 207 |
+
206,0.0006394955530196136,0.0014402546724304557,0.035160122837871316,0.0012811495875413768,0.05519585817120969,0.0036006927850191018,2.5695765810087323,False
|
| 208 |
+
207,0.0006364759677586616,0.001470587879884988,0.03847533715330064,0.0015380823012626158,0.05737438938580453,0.0038164917023143594,2.567013908177614,False
|
| 209 |
+
208,0.0006334509946601868,0.001479851054958999,0.040202391255646945,0.0016932254880594063,0.059105751365423204,0.004017610977553899,2.56655440479517,False
|
| 210 |
+
209,0.0006304207531449474,0.0014255631221458316,0.04399356294982135,0.0020521076975749154,0.062359060691669584,0.004541914741230569,2.585212233942002,False
|
| 211 |
+
210,0.0006273853628416901,0.0013409438678063453,0.04906546632200479,0.002529859768680545,0.06576950983144343,0.005140203037565496,2.566684365272522,False
|
| 212 |
+
211,0.0006243449435824264,0.0012849541842006147,0.04981829264014959,0.002596323350498582,0.06593803430907429,0.0052093677467077,2.5652159680612385,False
|
| 213 |
+
212,0.0006212996153977028,0.0012418448701500893,0.046217165416106584,0.0022388172819964173,0.06291196308098733,0.004763673796027762,2.5656589888967574,False
|
| 214 |
+
213,0.0006182494985118615,0.001235579897183925,0.0401429371740669,0.0016923073249673472,0.05850938805378973,0.004114671521781258,2.5824229889549315,False
|
| 215 |
+
214,0.0006151947133382945,0.0012540119308978319,0.03636847909167409,0.0013831690489713245,0.056125461598858235,0.0037457782221881736,2.565607429947704,False
|
| 216 |
+
215,0.0006121353804746897,0.0012853789799846709,0.037012102918699386,0.0014367475895004957,0.056683310633525254,0.003775522910160695,2.5636268300004303,False
|
| 217 |
+
216,0.0006090716206982704,0.0013237437820062042,0.03912565107643604,0.0016081966261190027,0.05826238194480538,0.003941722455420239,2.5648530698381364,False
|
| 218 |
+
217,0.0006060035549610265,0.001368418830446899,0.04082941557653248,0.0017571362596333026,0.059969081394374374,0.004150571173484522,2.5651682429015636,False
|
| 219 |
+
218,0.0006029313043849398,0.001384262289851904,0.04295468635112047,0.001964621341245561,0.0621221726667136,0.004498211230588442,2.567461724858731,False
|
| 220 |
+
219,0.0005998549902572027,0.0013508212151937186,0.045596721736714245,0.0022173706689381364,0.06390778465196491,0.0048563592307467466,2.5634267241694033,False
|
| 221 |
+
220,0.0005967747340254292,0.0012868172330781818,0.04841937085241079,0.0024902186596576785,0.06544675837270915,0.0051429011629038545,2.5609912811778486,False
|
| 222 |
+
221,0.0005936906572928614,0.0012256855904124676,0.050810214644297955,0.0027405589597634884,0.06694102665409446,0.005366465299932617,2.557075909804553,False
|
| 223 |
+
222,0.0005906028818135678,0.0011940667857415974,0.05148964210413396,0.0028161734631965444,0.06733812807127833,0.005402011859302043,2.553943432867527,False
|
| 224 |
+
223,0.0005875115294876372,0.001210704818367958,0.04992774617299438,0.002649799884458061,0.06606564511545003,0.005184850036345234,2.5502706640399992,False
|
| 225 |
+
224,0.0005844167223563661,0.001269049544353038,0.04629736001417041,0.0022762915742798543,0.06331247623078524,0.0047506314010292115,2.5589584899134934,False
|
| 226 |
+
225,0.000581318582597441,0.0013509643971920014,0.04219148040376604,0.0018892763994793056,0.06039362667128444,0.004313282140018761,2.5613026027567685,False
|
| 227 |
+
226,0.0005782172325201146,0.0014345245650038124,0.040779002353549004,0.001781431839797007,0.05951527329161763,0.00417171725187109,2.559615673031658,False
|
| 228 |
+
227,0.0005751127945603777,0.0015080327875912189,0.04279816571436822,0.0019634155412973156,0.06095219175331294,0.004323802380997122,2.5584907783195376,False
|
| 229 |
+
228,0.0005720053912761253,0.0015995995183475315,0.04452589438483119,0.0020935455460934956,0.06207357396371663,0.004421165291430142,2.56315107177943,False
|
| 230 |
+
229,0.0005688951453423183,0.0017590206782333554,0.04399758687801659,0.0020262912622139037,0.062055723816156386,0.004397946264237393,2.5653735040687025,False
|
| 231 |
+
230,0.0005657821795461405,0.0018873773012310266,0.044642633095383646,0.002157566093201036,0.06392001886852086,0.004770800767488818,2.559520953334868,False
|
| 232 |
+
231,0.0005626666167821513,0.0018127088551409543,0.048606153219938275,0.002637407597886317,0.06681485760957002,0.005385375026063732,2.5574103416875005,False
|
| 233 |
+
232,0.000559548580047434,0.0017160756569355726,0.06048610414192081,0.003970596299485788,0.07563261041417718,0.006736601881521781,2.557283982168883,False
|
| 234 |
+
233,0.0005564281924367399,0.0017737092655152082,0.06277219347283244,0.00423042993843436,0.0773552742600441,0.006922417006471517,2.555886634159833,False
|
| 235 |
+
234,0.0005533055771376291,0.0019243917716667057,0.05306313656270504,0.003036783098200331,0.06861771298572421,0.005546237876104714,2.5549215390346944,False
|
| 236 |
+
235,0.0005501808574256065,0.001930237107910216,0.046436549121513966,0.0023273134499477645,0.06309491277672351,0.004684264418190384,2.5599620258435607,False
|
| 237 |
+
236,0.0005470541566592563,0.001802462193183601,0.04499832124449313,0.0021928435435938135,0.06203688810579479,0.004485139746760685,2.562228078022599,False
|
| 238 |
+
237,0.0005439255982753708,0.0018247383646667003,0.043451037514954806,0.0020395836421010917,0.060739025976508855,0.004328335405911285,2.5598861454054713,False
|
| 239 |
+
238,0.0005407953057840779,0.0018647385444492102,0.04216424688696861,0.001914941141186839,0.05965570726431906,0.004225954315156344,2.5588747397996485,False
|
| 240 |
+
239,0.0005376634027639655,0.0019304181793704629,0.04322846089489758,0.002021914755446886,0.05965254610404372,0.004253086728088432,2.559587084222585,False
|
| 241 |
+
240,0.0005345300128572021,0.002082858959212899,0.04948445299267769,0.002677278094089351,0.06377683505415917,0.004841902718996077,2.560993320774287,False
|
| 242 |
+
241,0.0005313952597646558,0.002287665957584977,0.05818486618250608,0.0036573951874589367,0.07067368359304965,0.005811993989354336,2.5598655487410724,False
|
| 243 |
+
242,0.0005282592672410113,0.0025698748491704464,0.061762630537152294,0.004074540822607559,0.07396539290435612,0.006235109796509628,2.5612020269036293,False
|
| 244 |
+
243,0.0005251221590898839,0.0028553489055484532,0.053323650170117615,0.003066412535803058,0.06670179893262684,0.005145419761085293,2.55887112300843,False
|
| 245 |
+
244,0.0005219840591589315,0.002743379954248667,0.044719157859683036,0.0021066247833835673,0.05925954002887011,0.0042021290914933795,2.5519774868153036,False
|
| 246 |
+
245,0.0005188450913349664,0.002217074335552752,0.046598987236619,0.0023531643829871303,0.06148745664395392,0.004640618580132701,2.5529828420840204,False
|
| 247 |
+
246,0.0005157053795390632,0.0017068424909375609,0.04161094822734594,0.001818047364449304,0.058443317171186206,0.004028705600467305,2.5526103409938514,False
|
| 248 |
+
247,0.0005125650477216679,0.0015207535778172315,0.03651306180097163,0.0013918652313412415,0.055484638419002294,0.003634921351099862,2.566340954042971,False
|
| 249 |
+
248,0.0005094242198577033,0.0015157859334722162,0.03905330972000957,0.0015849972272274678,0.05751927076838911,0.0038632780937158874,2.5546049810945988,False
|
| 250 |
+
249,0.0005062830199416755,0.0014399823443964125,0.042150946244597434,0.0018246185931763436,0.059439749494194984,0.004071375092179751,2.5618287990801036,False
|
| 251 |
+
250,0.0005031415719827788,0.001416943754069507,0.03979826589860022,0.0016324916896466673,0.05694564376026392,0.0038328067473409775,2.5623950827866793,False
|
| 252 |
+
251,0.0004999999999999993,0.0013658661302179099,0.03488359469734132,0.0012642974610494865,0.052972391210496426,0.003420683055742073,2.574483957141638,False
|
| 253 |
+
252,0.0004968584280172197,0.001193027726840228,0.032568625936284665,0.0011033510872510883,0.051302376622334125,0.0032306888768737106,2.56016882089898,True
|
| 254 |
+
253,0.0004937169800583229,0.0010501801455393433,0.03151967662759125,0.0010281993142447225,0.05071947325952351,0.00314833095466213,2.5620633619837463,True
|
| 255 |
+
254,0.000490575780142295,0.0009693631129339337,0.030337691314518453,0.000948054550403434,0.05019049900583923,0.003075407578060417,2.561861040070653,True
|
| 256 |
+
255,0.00048743495227833043,0.0009134742622263729,0.02945964914560318,0.0008917436014009924,0.04983590823598206,0.0030261984379839468,2.5682856841012836,True
|
| 257 |
+
256,0.00048429462046093504,0.0008674470032565295,0.028951844083145262,0.0008602860986850192,0.04966420678421855,0.003002163065788704,2.561756579205394,True
|
| 258 |
+
257,0.00048115490866503186,0.0008277865438722074,0.028635552573949098,0.0008413141969841724,0.049575919760391114,0.0029927236898178505,2.5525910300202668,True
|
| 259 |
+
258,0.0004780159408410668,0.0007975142989307642,0.02838964765518904,0.0008270044313741878,0.04950109810568392,0.002988433341252703,2.5531060942448676,True
|
| 260 |
+
259,0.00047487784091011445,0.0007760454586241395,0.02817086249217391,0.0008143006189966262,0.04942169286310673,0.002984425452209319,2.552499991375953,True
|
| 261 |
+
260,0.00047174073275898696,0.0007608208362944424,0.02799766689259559,0.0008042161518297594,0.04934790764003992,0.0029816349993344675,2.553718297276646,True
|
| 262 |
+
261,0.00046860474023534266,0.0007487049687188119,0.027849656836129725,0.0007956820620572955,0.04927411461248994,0.0029794466204430387,2.5543651599437,True
|
| 263 |
+
262,0.0004654699871427964,0.0007377188373357057,0.02771357572171837,0.0007879743268473571,0.049188792910426854,0.0029760962981845,2.551662715151906,True
|
| 264 |
+
263,0.00046233659723603286,0.0007270605657249689,0.02759366043098271,0.0007813566703731842,0.04908756671473384,0.002970403306021042,2.5529407830908895,True
|
| 265 |
+
264,0.0004592046942159205,0.0007174868697766214,0.027481800314038993,0.0007752763099311775,0.04899075477384031,0.002964149514617353,2.5618711397983134,True
|
| 266 |
+
265,0.0004560744017246277,0.0007097317839507014,0.02736033314000815,0.0007684866527461346,0.04889730737544596,0.0029575036175812975,2.560376562178135,True
|
| 267 |
+
266,0.00045294584334074214,0.00070377724384889,0.027235091883689166,0.0007612795510061709,0.04882710313424468,0.0029523168405583117,2.5595785109326243,True
|
| 268 |
+
267,0.00044981914257439187,0.0006992709918413312,0.02717584485374391,0.000757872715790621,0.04883420281112194,0.0029560745520356225,2.56078798789531,True
|
| 269 |
+
268,0.0004466944228623693,0.000696656265296042,0.027251982444897294,0.0007624204714853491,0.04895054012537003,0.0029719416807554733,2.554567124694586,False
|
| 270 |
+
269,0.0004435718075632584,0.0006961344231385737,0.027453404758125542,0.0007743961623559017,0.049171971511095765,0.00299889244528866,2.5529980161227286,False
|
| 271 |
+
270,0.0004404514199525645,0.0006973720076493919,0.027635824006050826,0.0007852875302447215,0.049376677935943004,0.003021383883992875,2.573216195218265,False
|
| 272 |
+
271,0.0004373333832178472,0.0007006414425559342,0.027657202240079642,0.000786611245614347,0.04945057505741715,0.003026221286812327,2.573233363684267,False
|
| 273 |
+
272,0.00043421782045385804,0.0007065740306861698,0.027516384826041757,0.0007794640677689727,0.049322552094236013,0.003008307669534608,2.5622453750111163,False
|
| 274 |
+
273,0.00043110485465768026,0.0007204340489115566,0.027875474160537123,0.0008087195183917292,0.049249782608821985,0.003007532224718785,2.570579740218818,False
|
| 275 |
+
274,0.00042799460872387313,0.0007593826653901488,0.030661045247688888,0.0010099564266760328,0.050356295304372904,0.0031699002693100814,2.552285011857748,False
|
| 276 |
+
275,0.00042488720543962065,0.0008766583069227636,0.03727723906561732,0.0015274906178385237,0.054115444319322706,0.003679767899722816,2.5603101360611618,False
|
| 277 |
+
276,0.0004217827674798838,0.0012096786247566343,0.03558666317537427,0.0013826869018007599,0.053739895699545744,0.0036324586528243084,2.560702627990395,False
|
| 278 |
+
277,0.0004186814174025574,0.0016099670007824899,0.04533224201947451,0.002254207813846969,0.06109303569421172,0.004604203217207072,2.55740237981081,False
|
| 279 |
+
278,0.0004155832776436323,0.002032036936841905,0.038774415384978056,0.0016069420534133387,0.05737414733506739,0.0039673631813473996,2.5634013218805194,False
|
| 280 |
+
279,0.00041248847051236125,0.0016092960117384792,0.03410588321276009,0.0012627677931485457,0.05228916081599891,0.00336020029058958,2.5573469987139106,False
|
| 281 |
+
280,0.00040939711818643057,0.0011399275669828058,0.033427454994991425,0.0011634279312827697,0.053461573347449304,0.003422956081365307,2.556630176026374,False
|
| 282 |
+
281,0.0004063093427071369,0.0008827969962731004,0.029806076137349008,0.0009156270958696117,0.05075657548382878,0.0030927315009085963,2.5595875908620656,False
|
| 283 |
+
282,0.0004032252659745692,0.0008476274942513555,0.02972348503023386,0.0009212295568289757,0.05014185663312674,0.00309190445435513,2.5630535599775612,False
|
| 284 |
+
283,0.0004001450097427958,0.0008149217809550464,0.028739890126511455,0.0008515239675980709,0.05005400522612035,0.003077791349266156,2.553545258939266,False
|
| 285 |
+
284,0.0003970686956150585,0.000777039871783927,0.02881426153331995,0.0008522028870839561,0.05001244106329977,0.0030472591970196845,2.552367902826518,False
|
| 286 |
+
285,0.00039399644503897196,0.000753547758795321,0.02939186924137175,0.0008912805706568084,0.05000566124916077,0.0030546574509168834,2.55161538021639,False
|
| 287 |
+
286,0.000390928379301728,0.0007514341077767313,0.02993232280574739,0.0009330249993221915,0.0503247722517699,0.0031036283701342342,2.5521032218821347,False
|
| 288 |
+
287,0.00038786461952530863,0.0007595436740666628,0.030172429874539376,0.0009443506900433118,0.050514242900535467,0.0031078809761144093,2.5500045409426093,False
|
| 289 |
+
288,0.00038480528666170403,0.0007586922128684819,0.031831665435805914,0.0010499895568478902,0.05122689646668732,0.003174606638040248,2.5556003507226706,False
|
| 290 |
+
289,0.0003817505014881369,0.0007694362085312605,0.034691910879686476,0.0012534808585759523,0.05255131235346198,0.0033352108806214433,2.557861901819706,False
|
| 291 |
+
290,0.00037870038460229553,0.0007935684253461659,0.03613597342930734,0.001361311221168258,0.05330301785841584,0.0034268055285590045,2.5607204693369567,False
|
| 292 |
+
291,0.00037565505641757197,0.0008335259081795811,0.03583983075246215,0.0013362884945861393,0.05313288615085185,0.0033921695480730474,2.5582207650877535,False
|
| 293 |
+
292,0.00037261463715830827,0.0008712851000018417,0.03689513729512692,0.0014184014953342872,0.053662767885252836,0.003442999258350559,2.557804183103144,False
|
| 294 |
+
293,0.0003695792468550508,0.0009134735108818859,0.039635832734406,0.0016555165603993945,0.055050960322842,0.0036353735431052326,2.5571271190419793,False
|
| 295 |
+
294,0.0003665490053398116,0.0009725478624459356,0.04032937925495207,0.0017402747922627703,0.05527141264639795,0.0037139989318749006,2.561520243063569,False
|
| 296 |
+
295,0.00036352403224133655,0.0010673980505671351,0.03696356179565191,0.0014796270887819742,0.053143053548410535,0.0034749657602439455,2.560049084946513,False
|
| 297 |
+
296,0.00036050444698038455,0.0011663512839004398,0.03595615794509649,0.0013878915852802525,0.05284577057696879,0.003407614607776774,2.569106883369386,False
|
| 298 |
+
297,0.00035749036876501117,0.001294415612705052,0.039913550913333894,0.0016851385475738644,0.05606663263402879,0.0037387412117004235,2.5642164670862257,False
|
| 299 |
+
298,0.00035448191658586333,0.0014803429911844433,0.041880969958379866,0.0018360805557945308,0.05800409560091793,0.003933903432848623,2.5608079093508422,False
|
| 300 |
+
299,0.00035147920921148177,0.0017007860145531594,0.03924773856624961,0.0016114589046175305,0.05654476024210453,0.0038295826959058726,2.559432261157781,False
|
| 301 |
+
300,0.00034848236518361236,0.0018718322096392512,0.04262162704207003,0.0019525692396347649,0.05988541940227151,0.004468116569235834,2.5581964952871203,False
|
| 302 |
+
301,0.0003454915028125256,0.0020184507984668015,0.045740937991067765,0.0023108861447096295,0.062188071962445975,0.004831916274340731,2.565791266039014,False
|
| 303 |
+
302,0.0003425067401723469,0.002350353091955185,0.04074839976243675,0.0017662618492988514,0.05826944279484451,0.003989300065854352,2.5559409027919173,False
|
| 304 |
+
303,0.00033952819509639457,0.0027627468332648277,0.0597560972943902,0.003809415038062644,0.07334551084786653,0.005967651037435879,2.552851356100291,False
|
| 305 |
+
304,0.00033655598517252813,0.002654587909579277,0.07064307678118348,0.005169057549975312,0.081300127543509,0.00718366680733419,2.551532309036702,False
|
| 306 |
+
305,0.00033359022773850604,0.002474536132067442,0.03993875016085804,0.001669177545244397,0.056248827250674364,0.0036344859732913,2.552574255038053,False
|
| 307 |
+
306,0.00033063103987735366,0.0017116621304303409,0.036720929272472856,0.0014005206691244298,0.05627379612997174,0.003704834930623241,2.5506199840456247,False
|
| 308 |
+
307,0.0003276785384127408,0.00124484571069479,0.04746867265366018,0.0023802130304242503,0.06332586980424822,0.004557410283554006,2.550831186119467,False
|
| 309 |
+
308,0.0003247328399043698,0.0011781896487809717,0.05094584594480693,0.002713385195547488,0.06479960149154067,0.004718561281892304,2.5494684083387256,False
|
| 310 |
+
309,0.00032179406064337403,0.0013477850356139244,0.044079219583421946,0.002050538015480663,0.060566525775939226,0.004159707477946373,2.552022193092853,False
|
| 311 |
+
310,0.00031886231664772654,0.0013709445362910628,0.04935615801624954,0.0026011948859785143,0.06523746833205223,0.004827369366760209,2.5496465894393623,False
|
| 312 |
+
311,0.0003159377236576604,0.001503075978718698,0.05455285195447505,0.0031694706638471116,0.06858936073258519,0.005320309470453356,2.5524663338437676,False
|
| 313 |
+
312,0.0003130203971310992,0.001780443698167801,0.05248792878352106,0.0029153419389670336,0.06674100890755653,0.0050105354392140245,2.552708604838699,False
|
| 314 |
+
313,0.00031011045223909887,0.001931993680074811,0.054593439891934395,0.0031195355545271396,0.06873815635219216,0.005233172267315772,2.551316157914698,False
|
| 315 |
+
314,0.00030720800386130115,0.001883190620690584,0.0533433037288487,0.0030008859451401524,0.06793914256617427,0.005139337825681715,2.550435303710401,False
|
| 316 |
+
315,0.00030431316658139826,0.0017461250675842165,0.051022334983572365,0.002758122086826971,0.06586006816476583,0.004893545831631095,2.55251638405025,False
|
| 317 |
+
316,0.000301426054682609,0.0016705568060278893,0.052347576193511484,0.002883629778686658,0.06633955711498857,0.004931970928781607,2.5523136942647398,False
|
| 318 |
+
317,0.00029854678214316806,0.0017607580721378327,0.052588436910882594,0.002888098326840705,0.06651759181171656,0.004959860837665424,2.5576808536425233,False
|
| 319 |
+
318,0.00029567546263182496,0.001908898388966918,0.05129448153451085,0.002759675788604169,0.06575724733993411,0.004927974255270994,2.5524686463177204,False
|
| 320 |
+
319,0.00029281220950335736,0.0020546583840623496,0.04689861568808556,0.0023288329504512644,0.06250919181853533,0.004565318791588729,2.561109361704439,False
|
| 321 |
+
320,0.0002899571357940963,0.002289232515729964,0.04329885865561664,0.0019841404542978534,0.059678752021864054,0.004248224151411845,2.566213928628713,False
|
| 322 |
+
321,0.0002871103542174631,0.0026754467226564886,0.04674643863365054,0.002382466538653639,0.06223516455851495,0.004684789960224267,2.56040743784979,False
|
| 323 |
+
322,0.00028427197715952013,0.0027181632714346053,0.051705582512542606,0.002881625128884565,0.06582135179080069,0.005153327429752852,2.5610734219662845,False
|
| 324 |
+
323,0.0002814421166745331,0.0026447963267564774,0.06011750865355134,0.00384827349071754,0.072246113140136,0.006141129244080191,2.563632076140493,False
|
| 325 |
+
324,0.0002786208844805488,0.0026306796111166477,0.059229263570159674,0.0036963302857379032,0.07226588907651603,0.006007862545683912,2.576428664382547,False
|
| 326 |
+
325,0.00027580839195498344,0.0022773090796545147,0.04482415304891765,0.002105675477232035,0.06206703456118703,0.004509147785220246,2.5594929731450975,False
|
| 327 |
+
326,0.0002730047501302261,0.0016007495606318116,0.03488785531930626,0.0012666113406343354,0.055168075747787954,0.003634495453551303,2.5573259009979665,False
|
| 328 |
+
327,0.00027021006968925556,0.0011461711851879954,0.03155994496587664,0.0010295533037495358,0.053091494087129834,0.003378007507043986,2.5587637592107058,False
|
| 329 |
+
328,0.0002674244609612703,0.0009255822752602398,0.029268977103754878,0.000880554797481471,0.05146296763792634,0.0032032652783813877,2.56453788978979,False
|
| 330 |
+
329,0.0002646480339173334,0.0008110615778714419,0.028198652755469084,0.000815390699643382,0.050670777540653945,0.003127529565854369,2.5579577437601984,False
|
| 331 |
+
330,0.00026188089816602987,0.0007578553203493357,0.027620686132460832,0.0007814393236501866,0.050246624844148755,0.003091806428049527,2.5583339668810368,False
|
| 332 |
+
331,0.000259123162949142,0.0007297988815698772,0.027254332246258856,0.0007604548297614509,0.04998537027277052,0.0030699939452859867,2.556665188167244,False
|
| 333 |
+
332,0.00025637493713733347,0.0007106432474683971,0.02698546092398465,0.0007453797696708252,0.04979869605973363,0.003053079660045511,2.5581778259947896,True
|
| 334 |
+
333,0.0002536363292258536,0.0006958063230849802,0.026732718168757857,0.0007314221887756399,0.04962336424738169,0.003035615770972998,2.552730165887624,True
|
| 335 |
+
334,0.00025090744733025414,0.0006829153597354889,0.02648637720197439,0.0007179730929774558,0.04944621472619474,0.003016811089607111,2.5546657368540764,True
|
| 336 |
+
335,0.00024818839918211915,0.0006711663643363863,0.02625290019903332,0.0007053442531487386,0.04927410467527807,0.002998054774690889,2.553604062180966,True
|
| 337 |
+
336,0.00024547929212481387,0.0006604636625852435,0.026031647481024264,0.000693477437558197,0.04910683734342456,0.002979802670723056,2.550413393881172,True
|
| 338 |
+
337,0.00024278023310924626,0.0006506215627305209,0.025827343416400254,0.0006826011136631796,0.04894760104827583,0.0029622625310119955,2.558538377750665,True
|
| 339 |
+
338,0.00024009132868964492,0.0006417301967740059,0.025636373709887268,0.0006725069947852122,0.048797979475930335,0.0029460582219910975,2.5563707766123116,True
|
| 340 |
+
339,0.00023741268501935162,0.00063356545381248,0.02546448839828372,0.0006634832787413907,0.04866078197956085,0.002931246571929116,2.558843599166721,True
|
| 341 |
+
340,0.00023474440784663263,0.0006261040992103517,0.025307408588938415,0.0006552904799016759,0.0485326237604022,0.0029174490524848705,2.555152656044811,True
|
| 342 |
+
341,0.00023208660251050134,0.0006193270285148174,0.02516277564689517,0.0006477905321562975,0.04841325845569372,0.0029048052060554243,2.558808504138142,True
|
| 343 |
+
342,0.00022943937393656168,0.0006131476387381553,0.02503137975744903,0.0006410112793532413,0.04830283105373383,0.002893208288085782,2.56443571485579,True
|
| 344 |
+
343,0.00022680282663286507,0.0006075377203524113,0.024913627802394332,0.0006349692577176733,0.04820228223688901,0.0028827173232670723,2.5608118348754942,True
|
| 345 |
+
344,0.0002241770646857845,0.0006023075680714101,0.024806528197601437,0.0006294976666420798,0.04811026010662317,0.0028731214223736918,2.558662293944508,True
|
| 346 |
+
345,0.00022156219175590558,0.0005975106679834425,0.024708743660710752,0.0006245268670223904,0.04802308300510049,0.0028641787114505805,2.5605102539993823,True
|
| 347 |
+
346,0.00021895831107393427,0.0005930794565938413,0.02461747553385794,0.0006199066880872053,0.047942389296367766,0.0028560188480509204,2.5632971487939358,True
|
| 348 |
+
347,0.00021636552543662148,0.0005889412933029235,0.024536217020824553,0.0006158090792214277,0.04787149727344513,0.002848820494680842,2.557747988961637,True
|
| 349 |
+
348,0.0002137839372027041,0.0005850474403705448,0.024459974475204944,0.0006119786336066878,0.04780242026783526,0.0028418787849164512,2.5483231083489954,True
|
| 350 |
+
349,0.00021121364828886592,0.0005813627853058278,0.024389767341315746,0.000608464673615893,0.047737689837813374,0.002835338093450577,2.5492663141340017,True
|
| 351 |
+
350,0.00020865476016571167,0.0005779226354788989,0.024322666959837078,0.0006051186435341729,0.047679344285279514,0.0028297176780035602,2.5483583291061223,True
|
| 352 |
+
351,0.00020610737385376313,0.000574650589376688,0.02425798952486366,0.0006019022337784368,0.0476184101216495,0.002823666708881164,2.5502626090310514,True
|
| 353 |
+
352,0.00020357158991947004,0.0005714897993020714,0.02420084789954126,0.0005990725354805288,0.047566500948742034,0.0028186507257415166,2.5498258867301047,True
|
| 354 |
+
353,0.00020104750847124041,0.000568498236592859,0.024146015727892517,0.0005963657336875005,0.047517435094341635,0.002814033818258768,2.5489676729775965,True
|
| 355 |
+
354,0.00019853522915548727,0.0005656492095440626,0.024091046242974697,0.0005936564061017902,0.047467759549617766,0.0028092131695676087,2.5488313222303987,True
|
| 356 |
+
355,0.00019603485115269692,0.0005628550003748387,0.02404207983240485,0.0005912575764063868,0.04742236765101552,0.002805002135398311,2.557072455994785,True
|
| 357 |
+
356,0.00019354647317351156,0.0005601750114001333,0.023996083125472068,0.0005890087951158114,0.04738169131800532,0.0028012732368972604,2.5570287848822773,True
|
| 358 |
+
357,0.00019107019345483242,0.000557614334160462,0.023946103353053333,0.0005865656092474996,0.04733908397145569,0.002797366999169643,2.558145351242274,True
|
| 359 |
+
358,0.00018860610975594353,0.0005551157561130822,0.0239014544878155,0.0005843940472672886,0.047299389131367206,0.0027937959582965393,2.55797916976735,True
|
| 360 |
+
359,0.00018615431935464953,0.0005526853592600674,0.023857536485418676,0.000582260808131333,0.047261957908049226,0.002790415811955012,2.5578087689355016,True
|
| 361 |
+
360,0.0001837149190434375,0.0005503184592816979,0.023818560829386114,0.0005803785546979927,0.04722921268083155,0.0027876779390418644,2.557081745006144,True
|
| 362 |
+
361,0.00018128800512565488,0.0005480337045155466,0.023776035100221634,0.0005783196926844465,0.047195617388933894,0.0027846461801200027,2.550187658984214,True
|
| 363 |
+
362,0.0001788736734117077,0.0005458248329814524,0.02373342288658023,0.0005762605795505439,0.047159921191632745,0.0027814791775228253,2.5791130531579256,True
|
| 364 |
+
363,0.00017647201921527754,0.0005436418410390615,0.023692142624408007,0.0005742704872667699,0.04712871444411576,0.002778786377606697,2.57003102498129,True
|
| 365 |
+
364,0.0001740831373495603,0.000541491872863844,0.02365353077929467,0.0005724186849296832,0.04709738811478019,0.0027761891629385265,2.5592921022325754,True
|
| 366 |
+
365,0.0001717071221235216,0.0005394080830737948,0.023617106656543912,0.0005706730366272752,0.04707164563238621,0.002774246493617184,2.566253498662263,True
|
| 367 |
+
366,0.00016934406733817376,0.000537394730374217,0.023581573519855736,0.0005689742147385187,0.04704544235020876,0.0027720411234613516,2.558302897028625,True
|
| 368 |
+
367,0.000166994066282874,0.0005354291938710957,0.023542556655593218,0.0005671039997227353,0.04701765939593315,0.002769759741601703,2.5584620968438685,True
|
| 369 |
+
368,0.00016465721173163978,0.0005335077790077776,0.023501722941175102,0.000565146750175394,0.04698819076642394,0.0027671902742306864,2.558165195863694,True
|
| 370 |
+
369,0.00016233359593948756,0.0005315894160885364,0.023463342742063104,0.0005633140244876451,0.04696244144812226,0.0027651139018789205,2.573972119949758,True
|
| 371 |
+
370,0.00016002331063879019,0.0005297250603325665,0.023430246254429223,0.0005617411904079165,0.04694070316851139,0.002763472869071279,2.556048203725368,True
|
| 372 |
+
371,0.0001577264470356556,0.0005278749326243997,0.02339607672113925,0.0005601172937586356,0.04691956090740859,0.002761991092247467,2.5590598708949983,True
|
| 373 |
+
372,0.0001554430958063255,0.0005260999805759639,0.023361226974055172,0.0005584638580019794,0.046896358747035266,0.00276022718710249,2.556942578870803,True
|
| 374 |
+
373,0.0001531733470935974,0.0005243628425523638,0.02332237218506634,0.0005566126369939687,0.046873540449887514,0.00275844414299063,2.558728478848934,True
|
| 375 |
+
374,0.00015091729050326357,0.000522654959699139,0.02328375716973096,0.0005547766399930994,0.04684820307418704,0.002756376643967807,2.5565004269592464,True
|
| 376 |
+
375,0.0001486750151005751,0.0005209464312065392,0.023246994917280973,0.0005530323511691518,0.04682599468156695,0.002754649590006318,2.5579461539164186,True
|
| 377 |
+
376,0.0001464466094067261,0.0005192834688350558,0.02321129791159183,0.0005513430754133153,0.046807056963443755,0.0027533100901185837,2.557307511102408,True
|
| 378 |
+
377,0.00014423216139535715,0.0005176487376447768,0.023175641272217034,0.0005496578639043262,0.0467856634221971,0.002751679329982348,2.555996594019234,True
|
| 379 |
+
378,0.0001420317584890842,0.0005160525743849576,0.02314307399839163,0.0005481239341766372,0.046769250854849814,0.0027506244359888038,2.550234912894666,True
|
| 380 |
+
379,0.00013984548755604636,0.0005144720119424164,0.0231111930180341,0.000546624486472592,0.04675241951830685,0.0027494744380681686,2.553165089339018,True
|
| 381 |
+
380,0.0001376734349064766,0.0005129304442089051,0.023073215268552303,0.0005448293606821178,0.04673168397508562,0.002747957897135096,2.551706811413169,True
|
| 382 |
+
381,0.000135515686289294,0.0005114258490502835,0.023032310146838427,0.0005428943433964077,0.04670939291827381,0.0027461266817013596,2.5526444567367435,True
|
| 383 |
+
382,0.00013337232688871993,0.0005099399290047586,0.022995097655802967,0.0005411445059224948,0.04668905825354159,0.002744547942902276,2.552232373971492,True
|
| 384 |
+
383,0.00013124344132091294,0.0005084557135123759,0.022960624386556445,0.0005395266340679787,0.04667328239418566,0.002743507258229813,2.5524604860693216,True
|
| 385 |
+
384,0.0001291291136306303,0.0005069943678099662,0.022927433638833462,0.00053797116221237,0.04665876343846321,0.0027425464013363203,2.550222094170749,True
|
| 386 |
+
385,0.0001270294272879088,0.000505578497890383,0.022894643197767437,0.0005364379196616493,0.04664397367276251,0.002741673234156975,2.5563006256707013,True
|
| 387 |
+
386,0.00012494446518477006,0.0005041846411768347,0.022858849602751433,0.0005347575774988492,0.04662607937119901,0.0027402637014493393,2.5567652080208063,True
|
| 388 |
+
387,0.00012287430963194793,0.0005028106011450291,0.022822608657181262,0.0005330616045899317,0.046608450310304764,0.002738949779605908,2.5574217019602656,True
|
| 389 |
+
388,0.00012081904235563893,0.0005014689951203763,0.022783350600861015,0.0005312235482888856,0.04658874989487231,0.0027373357692130197,2.5533108902163804,True
|
| 390 |
+
389,0.00011877874449427604,0.0005001393037382514,0.022747796564362942,0.0005295646834348802,0.04657165438868106,0.00273609606306991,2.5501565989106894,True
|
| 391 |
+
390,0.00011675349659532489,0.0004988243045518175,0.02271438989508897,0.0005280060216636455,0.04655788704752922,0.002735084595189367,2.5499072689563036,True
|
| 392 |
+
391,0.0001147433786121053,0.0004975255114259198,0.02268253767117858,0.0005265257365948489,0.046545987548306585,0.0027344018525851683,2.5501025803387165,True
|
| 393 |
+
392,0.00011274846990063301,0.0004962687460938469,0.022648372217081488,0.0005249368578486296,0.04653287892229855,0.002733475506588959,2.5736436750739813,True
|
| 394 |
+
393,0.0001107688492164882,0.0004950408751610666,0.022612041695043446,0.0005232469397629693,0.04651609000749886,0.002732268816603014,2.5580432121641934,True
|
| 395 |
+
394,0.00010880459471170581,0.0004938162717735394,0.022575393083505334,0.0005215434817625361,0.046498841047286986,0.002730874919200343,2.5582392127253115,True
|
| 396 |
+
395,0.00010685578393169056,0.0004926000788109377,0.022538731750100852,0.0005198393241540189,0.04648192317225039,0.002729451714607302,2.5605161651037633,True
|
| 397 |
+
396,0.00010492249381215466,0.0004914129953831434,0.022506920034065844,0.0005183672824781936,0.046468576742336155,0.002728481144466115,2.5661468259058893,True
|
| 398 |
+
397,0.00010300480067608218,0.0004902405693428591,0.022477358484640716,0.0005169999779950443,0.04645857335999608,0.002727895163351487,2.561635422054678,True
|
| 399 |
+
398,0.00010110278023071442,0.0004890937311574817,0.022447781692259014,0.000515631581975309,0.04644833031110465,0.0027272284478837717,2.558919607196003,True
|
| 400 |
+
399,9.921650756456149e-05,0.0004879798081237823,0.022416385035030545,0.0005141787970199994,0.04643446861766279,0.0027261617918694163,2.5502938111312687,True
|
| 401 |
+
400,9.734605714443892e-05,0.00048687914200127127,0.022381278523243963,0.0005125544464684904,0.04641702162101865,0.002724651935261203,2.5497852680273354,True
|
| 402 |
+
401,9.549150281252619e-05,0.00048579383408650753,0.022350273499265313,0.0005111213927455387,0.046403841907158495,0.002723591038775635,2.5537856691516936,True
|
| 403 |
+
402,9.365291778345289e-05,0.0004847141453064978,0.02232101165596396,0.0005097728349026269,0.0463913244381547,0.0027226073642645154,2.556791335809976,True
|
| 404 |
+
403,9.18303746414079e-05,0.00048365699802525343,0.022298005477525293,0.0005087139489597515,0.046385278711095455,0.002722361525164365,2.5511759351938963,True
|
| 405 |
+
404,9.00239453372738e-05,0.0004826180066447705,0.022275782840326427,0.0005076906028946179,0.046378203686326745,0.002721990303947155,2.5491317552514374,True
|
| 406 |
+
405,8.823370118578614e-05,0.0004816218678606674,0.02225052284169942,0.0005065271954236755,0.046368592027574776,0.0027213794995438827,2.5524923927150667,True
|
| 407 |
+
406,8.645971286271902e-05,0.00048065773071721195,0.02222289121337235,0.0005052559868657301,0.04635438612662256,0.0027201741150439805,2.5491392049007118,True
|
| 408 |
+
407,8.47020504020936e-05,0.0004797069950727746,0.022200182909145952,0.0005042123010495951,0.04634461358189583,0.002719476352761973,2.557718290016055,True
|
| 409 |
+
408,8.29607831934142e-05,0.00047877824306488036,0.022181659374386074,0.0005033618660552616,0.04633718314580619,0.0027190493762499703,2.5658109332434833,True
|
| 410 |
+
409,8.123597997892905e-05,0.00047787021193653346,0.022165343347936867,0.0005026149280332169,0.046332723507657644,0.002718942980277492,2.5574021390639246,True
|
| 411 |
+
410,7.952770885091536e-05,0.0004770029443316162,0.022149863578379156,0.0005019086347612726,0.046324883736670014,0.002718466895213913,2.5673703220672905,True
|
| 412 |
+
411,7.783603724899233e-05,0.0004761663753306493,0.022136152771301567,0.0005012873791441047,0.04631973963230848,0.0027182730965682267,2.560468304902315,True
|
| 413 |
+
412,7.616103195745825e-05,0.000475365730933845,0.022121042003855108,0.0005006044282726818,0.046311051258817314,0.0027177359133855766,2.5585253750905395,True
|
| 414 |
+
413,7.450275910265401e-05,0.00047462253062985837,0.02211000914964825,0.0005001117948576732,0.04630586489103734,0.0027175358693010488,2.5599721390753984,True
|
| 415 |
+
414,7.286128415035225e-05,0.00047390123386867344,0.022102831202559174,0.0004997998674505295,0.04630252136848867,0.0027176308260489944,2.5639168224297464,True
|
| 416 |
+
415,7.123667190317384e-05,0.00047322896833065895,0.022095466047525405,0.0004994874754535756,0.046296483911573885,0.0027174436507068855,2.567202349193394,True
|
| 417 |
+
416,6.962898649802811e-05,0.00047262297803536056,0.02208773245103657,0.0004991681182511297,0.046287943702191116,0.0027170479167430107,2.5580792501568794,True
|
| 418 |
+
417,6.803829140358213e-05,0.00047208446054719386,0.02208212523534894,0.0004989552946992132,0.0462802307959646,0.0027166022591645104,2.560586513020098,True
|
| 419 |
+
418,6.646464941775488e-05,0.0004715998799074441,0.02208021808974445,0.0004989156013860759,0.04627390990965068,0.0027163528150908577,2.5600894610397518,True
|
| 420 |
+
419,6.490812266523705e-05,0.0004712049481458962,0.022080429006367922,0.0004989882625216935,0.04626721396110952,0.002716193621625851,2.5618662713095546,False
|
| 421 |
+
420,6.336877259503983e-05,0.0004709033065009862,0.02207937979698181,0.000499014935335421,0.04625549604184925,0.0027155310469608278,2.5576217980124056,False
|
| 422 |
+
421,6.184665997806812e-05,0.00047071942361071706,0.022081333394162357,0.0004992009575199348,0.046242428757250306,0.00271461845498573,2.558843925129622,False
|
| 423 |
+
422,6.034184490472185e-05,0.00047065864491742105,0.02208229679055512,0.0004993542496639819,0.046225028038024904,0.0027133937872481197,2.5580309410579503,False
|
| 424 |
+
423,5.885438678252322e-05,0.0004707373288692907,0.02208996888063848,0.0004998363398984505,0.04620820201002061,0.0027121359787110853,2.55795359890908,False
|
| 425 |
+
424,5.738434433377235e-05,0.0004710084138205275,0.022096975744701923,0.0005003032788453587,0.04618270116858184,0.002710187279595837,2.5584262418560684,False
|
| 426 |
+
425,5.593177559322768e-05,0.00047146322461776437,0.0221096080429852,0.0005010615677610727,0.046152431555092334,0.0027076293320395877,2.5516691971570253,False
|
| 427 |
+
426,5.449673790581603e-05,0.0004721941010793671,0.022127020405605435,0.0005020626850366004,0.04611456003040075,0.002704486036587893,2.549164842814207,False
|
| 428 |
+
427,5.307928792436804e-05,0.0004732287754304707,0.022156240332871675,0.000503646845405385,0.046069303974509236,0.0027004268770009486,2.549139402806759,False
|
| 429 |
+
428,5.167948160738199e-05,0.00047462932160124185,0.022206180411390962,0.000506242510321543,0.046020084358751774,0.002696191403445074,2.550845212303102,False
|
| 430 |
+
429,5.029737421681439e-05,0.00047654907894320787,0.022286603510379793,0.0005102836704810883,0.045966380378231406,0.0026915228879157626,2.548736450728029,False
|
| 431 |
+
430,4.893302031589851e-05,0.00047904293285682797,0.022424397131428124,0.0005170752154671055,0.045918553248047826,0.0026877482931359797,2.5903751123696566,False
|
| 432 |
+
431,4.7586473766990253e-05,0.00048247107327915726,0.02266131047345698,0.0005286401559959296,0.04588856006041169,0.0026858436003467947,2.569987942930311,False
|
| 433 |
+
432,4.625778772944144e-05,0.00048700022301636635,0.023056863860227168,0.0005479421091954593,0.0459067053347826,0.0026891579970884662,2.5494088139384985,False
|
| 434 |
+
433,4.494701465750211e-05,0.0004933318025432527,0.02367870109528303,0.0005786070781048848,0.046013890914618966,0.0027016445175491993,2.5554319070652127,False
|
| 435 |
+
434,4.365420629824856e-05,0.0005021538564469665,0.024601243833079933,0.0006251044928828091,0.04626811685040593,0.0027290369175535945,2.558879094198346,False
|
| 436 |
+
435,4.237941368954117e-05,0.0005148883662186563,0.025705565333366393,0.0006822947162010356,0.046672441232949495,0.0027708656749770274,2.557094665709883,False
|
| 437 |
+
436,4.1122687158009364e-05,0.0005336831044405699,0.026305059204809366,0.0007132084729834573,0.046968515319749715,0.0027993303188531375,2.557498678099364,False
|
| 438 |
+
437,3.988407631706486e-05,0.0005597523583564907,0.02514540765620768,0.0006496149538320818,0.046640425994992254,0.0027605437390607803,2.5604886231012642,False
|
| 439 |
+
438,3.866363006494249e-05,0.0005807653781957924,0.02261679178290069,0.0005236482411806201,0.04600949652493,0.002693462507575835,2.558639505878091,False
|
| 440 |
+
439,3.746139658277086e-05,0.0005661299994681031,0.022061121891252696,0.0004975678817808585,0.04627365916967392,0.002720270409359107,2.5529734822921455,True
|
| 441 |
+
440,3.6277423332669306e-05,0.0005226567834615707,0.022084401784464715,0.0004982833303714322,0.04628556280396879,0.0027131468436752287,2.5537331849336624,False
|
| 442 |
+
441,3.511175705587421e-05,0.00048785282438620924,0.021719415848143398,0.00048223482740670365,0.04579528667032719,0.002657048851892389,2.5540500869974494,True
|
| 443 |
+
442,3.396444377089447e-05,0.0004709926983341575,0.021727400704286994,0.00048283232955399156,0.0455453862156719,0.002631581831239805,2.5521390670910478,False
|
| 444 |
+
443,3.283552877169394e-05,0.0004667012074496597,0.021747253851033748,0.0004836794260972107,0.04547667884267867,0.0026256131354954553,2.5495972731150687,False
|
| 445 |
+
444,3.172505662590381e-05,0.0004661804112838581,0.021708474936895074,0.00048189242932004126,0.04545596801675856,0.0026238244339686174,2.5573651753365993,True
|
| 446 |
+
445,3.0633071173062916e-05,0.00046509965066798034,0.0216876532305032,0.0004809425891697395,0.045439944947138426,0.002622122540616699,2.559261784888804,True
|
| 447 |
+
446,2.955961552288728e-05,0.0004636811900418252,0.02168313167989254,0.0004807350951410804,0.04542609940283,0.002620878343343355,2.5599946291185915,True
|
| 448 |
+
447,2.850473205356769e-05,0.00046257704775780437,0.0216772244758904,0.00048046531577380333,0.04542016756720841,0.0026206816127776304,2.5579633079469204,True
|
| 449 |
+
448,2.7468462410097604e-05,0.000461755700642243,0.021664727358147503,0.0004799043656705404,0.04541913672350347,0.002621101882349747,2.5697674807161093,True
|
| 450 |
+
449,2.645084750262784e-05,0.0004610910504125059,0.02165422693826258,0.00047943356061809153,0.045420471150428056,0.0026217529291300222,2.5598212480545044,True
|
| 451 |
+
450,2.5451927504852606e-05,0.0004605731591582298,0.02164530115947127,0.0004790323537862006,0.045422800527885554,0.0026224715218768372,2.559096423909068,True
|
| 452 |
+
451,2.4471741852423198e-05,0.00046020531048998235,0.02163353347685188,0.0004785034608748586,0.04542434622533619,0.0026230394361395886,2.559359715785831,True
|
| 453 |
+
452,2.351032924139059e-05,0.0004599948422983289,0.02162219799589366,0.0004779895231740925,0.045424304958432915,0.002623287611733557,2.5532791931182146,True
|
| 454 |
+
453,2.256772762667849e-05,0.000459916241816245,0.02161004978045821,0.00047743337719980403,0.04542257323861122,0.0026232735287726842,2.568313993047923,True
|
| 455 |
+
454,2.1643974220584693e-05,0.0004599569101119414,0.021596775357611477,0.00047681925032981137,0.0454205846786499,0.0026230507113069413,2.573499039746821,True
|
| 456 |
+
455,2.073910549131205e-05,0.00046007182239554825,0.021584174382500352,0.00047622420897563543,0.04541764222085476,0.0026225318073237304,2.55859615188092,True
|
| 457 |
+
456,1.9853157161528435e-05,0.00046022202307358383,0.0215750071760267,0.0004757681743984799,0.04541678065434098,0.00262204771116208,2.563054455909878,True
|
| 458 |
+
457,1.898616420695706e-05,0.0004603750503156334,0.021572420010343193,0.0004755884642342441,0.045418479973450304,0.002621615256048309,2.5588563829660416,True
|
| 459 |
+
458,1.8138160854995114e-05,0.0004605050889076665,0.021578967416658996,0.0004757995173196394,0.04542421870864928,0.0026214620691626138,2.5591087541542947,False
|
| 460 |
+
459,1.730918058336303e-05,0.0004605729035101831,0.02159572578035295,0.00047645250523434176,0.045434767846018076,0.0026216565772201926,2.5584597000852227,False
|
| 461 |
+
460,1.6499256118782475e-05,0.00046058671176433565,0.02162106681242585,0.0004774843655453065,0.04544887201860547,0.00262218411370089,2.5679997918196023,False
|
| 462 |
+
461,1.570841943568444e-05,0.0004605640168301761,0.021656931241974233,0.00047898448897840714,0.04546543255448341,0.0026229437926459983,2.558754383120686,False
|
| 463 |
+
462,1.4936701754947081e-05,0.00046052242000587285,0.021695879293605686,0.0004806344378132879,0.045481054950505494,0.002623731268521895,2.558632332831621,False
|
| 464 |
+
463,1.4184133542662995e-05,0.0004605420823208988,0.021734727540053428,0.0004823003631347511,0.04549360102973878,0.0026242653585046537,2.549875389318913,False
|
| 465 |
+
464,1.3450744508936723e-05,0.0004606571133481339,0.02176436234638095,0.0004835817865912516,0.04549888697452843,0.0026241981597449194,2.5515958382748067,False
|
| 466 |
+
465,1.2736563606711365e-05,0.00046086573833599686,0.021778180100023747,0.0004841884514041488,0.04549473942257464,0.0026233030099942684,2.5500834798440337,False
|
| 467 |
+
466,1.2041619030626323e-05,0.0004611914452398196,0.02176887992862612,0.00048379774023822565,0.04547923317179084,0.002621368321291899,2.5498513686470687,False
|
| 468 |
+
467,1.1365938215903245e-05,0.0004616624569753185,0.021730931094847618,0.0004821654493438469,0.04545222216285765,0.0026183946722743846,2.5488247629255056,False
|
| 469 |
+
468,1.0709547837263954e-05,0.000462183176074177,0.02165511800069362,0.00047889097048792526,0.04541343571618199,0.002614257284922338,2.548900913912803,False
|
| 470 |
+
469,1.0072473807876557e-05,0.00046267157630063596,0.0215407551638782,0.00047395471935110723,0.04536648228764534,0.0026093881785703155,2.547921628691256,True
|
| 471 |
+
470,9.454741278333e-06,0.00046293522510677576,0.021394327723421157,0.00046765842450973447,0.0453158797416836,0.0026042422641959894,2.5492724156938493,True
|
| 472 |
+
471,8.856374635655629e-06,0.0004627854183781892,0.02123379262071103,0.00046079385981803527,0.045270123621448874,0.002599731206621551,2.5484633217565715,True
|
| 473 |
+
472,8.277397502335129e-06,0.0004619953500805423,0.02108486524131149,0.0004544663841824969,0.0452381575666368,0.0025967146986538513,2.5483250408433378,True
|
| 474 |
+
473,7.717832735397325e-06,0.0004604585723718628,0.02097031331807375,0.0004496273073349506,0.04522405862808228,0.002595561498425079,2.5502918320707977,True
|
| 475 |
+
474,7.177702425500968e-06,0.00045828906237147747,0.02089705898053944,0.0004465448626299425,0.04522521671839058,0.0025959492128592773,2.5479517388157547,True
|
| 476 |
+
475,6.657027896065974e-06,0.0004557772648986429,0.020856810215860604,0.00044485299030118326,0.04523419439792633,0.0025970357646980836,2.548458101693541,True
|
| 477 |
+
476,6.1558297024311634e-06,0.00045329147367738186,0.020835173181258142,0.00044394099750916206,0.045242308704182506,0.002597926362481962,2.557194897904992,True
|
| 478 |
+
477,5.674127631043018e-06,0.0004510805739555508,0.020821700416505336,0.0004433699225251514,0.045247474201023576,0.002598438970001009,2.5585255748592317,True
|
| 479 |
+
478,5.211940698674527e-06,0.0004492569043068215,0.020811486014164986,0.0004429354932944589,0.04524944490753114,0.0025986183118215504,2.557143977843225,True
|
| 480 |
+
479,4.769287151674401e-06,0.00044779743265826257,0.020802950178273022,0.00044257194025717914,0.045249870037660005,0.0025986355833568318,2.5684830583631992,True
|
| 481 |
+
480,4.346184465246701e-06,0.0004466234690044075,0.020795626875944436,0.00044225986652810916,0.04524948740378022,0.002598578704588319,2.558436114806682,True
|
| 482 |
+
481,3.942649342761114e-06,0.00044566567079164086,0.020789283252321183,0.0004419896313338578,0.04524874616414309,0.0025984985386988376,2.558506514877081,True
|
| 483 |
+
482,3.558697715093203e-06,0.00044487374764867125,0.020783826113678516,0.00044175714106145693,0.045248054806143044,0.0025984181270921865,2.55705237807706,True
|
| 484 |
+
483,3.1943447399957993e-06,0.0004442077630665153,0.020779121154919266,0.0004415568107251316,0.04524740329012275,0.002598352954880559,2.5622243750840425,True
|
| 485 |
+
484,2.8496048015005352e-06,0.0004436437232652679,0.02077506650891155,0.00044138425001368924,0.045246780971065166,0.0025982919408911835,2.5590565782040358,True
|
| 486 |
+
485,2.5244915093499103e-06,0.0004431627143640071,0.020771603632718325,0.0004412369134857664,0.04524633146822452,0.002598246609379955,2.5580129721201956,True
|
| 487 |
+
486,2.2190176984599993e-06,0.00044274926534853875,0.020768651879392565,0.0004411114156601001,0.04524593081325293,0.0025982096665124895,2.5526436041109264,True
|
| 488 |
+
487,1.9331954284137454e-06,0.0004423923718277365,0.020766150713898243,0.0004410050743120789,0.04524560496211052,0.00259817665206789,2.551325744949281,True
|
| 489 |
+
488,1.6670359829850637e-06,0.000442084030713886,0.02076404622849077,0.0004409156027426701,0.045245356243103746,0.0025981511707175576,2.5518376040272415,True
|
| 490 |
+
489,1.4205498696930313e-06,0.0004418170923599973,0.02076229054480791,0.0004408409655183956,0.0452451653778553,0.0025981308207432894,2.5532494820654392,True
|
| 491 |
+
490,1.1937468193873854e-06,0.0004415862385649234,0.020760842024348677,0.0004407793868751515,0.045245016450062396,0.002598113546758245,2.5559587171301246,True
|
| 492 |
+
491,9.866357858642194e-07,0.00044138721632771194,0.02075966370664537,0.0004407292945858269,0.04524491653777659,0.0025981009423273415,2.5554272537119687,True
|
| 493 |
+
492,7.99224945512488e-07,0.00044121647940482945,0.020758721941150724,0.0004406892570526589,0.04524484386667609,0.002598089738639806,2.555413282942027,True
|
| 494 |
+
493,6.315216969912655e-07,0.00044107107678428294,0.020757986566983164,0.0004406579936767332,0.0452447921410203,0.0025980802354585884,2.55427507776767,True
|
| 495 |
+
494,4.835326609376461e-07,0.0004409485488431528,0.020757427864708005,0.0004406342392902109,0.045244758566841486,0.002598072177127321,2.555480174254626,True
|
| 496 |
+
495,3.552636797053972e-07,0.00044084691628813743,0.020757020458579065,0.00044061691515262925,0.04524474010802806,0.0025980657670760573,2.554173336364329,True
|
| 497 |
+
496,2.4671981713419966e-07,0.0004407642579171807,0.020756738672964275,0.00044060493625409827,0.04524473208934068,0.002598060765579577,2.5517233232967556,True
|
| 498 |
+
497,1.5790535835002987e-07,0.00044069921190384773,0.02075655865482986,0.00044059727961834707,0.045244717746973036,0.002598055565536102,2.552299396134913,True
|
| 499 |
+
498,8.882380959551728e-08,0.00044065024144947527,0.020756457422859968,0.00044059297335676304,0.04524470739997923,0.0025980518757507324,2.548582026734948,True
|
| 500 |
+
499,3.947789809194408e-08,0.0004406164406100288,0.02075641248561442,0.0004405910533074976,0.04524469004943967,0.0025980488478794237,2.5502143460325897,True
|
| 501 |
+
500,9.869571931442322e-09,0.0004405951842200011,0.020756403462029994,0.00044059067160258456,0.045244683949276804,0.002598048033511887,2.557273573242128,True
|
results/training_status.json
ADDED
|
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"status": "completed",
|
| 3 |
+
"completed_epochs": 500,
|
| 4 |
+
"best_epoch": 500,
|
| 5 |
+
"best_training_loss": 0.00044059067160258456,
|
| 6 |
+
"last_epoch": {
|
| 7 |
+
"epoch": 500,
|
| 8 |
+
"lr": 9.869571931442322e-09,
|
| 9 |
+
"online_train_loss": 0.0004405951842200011,
|
| 10 |
+
"train_relative_l2": 0.020756403462029994,
|
| 11 |
+
"train_squared_relative_l2": 0.00044059067160258456,
|
| 12 |
+
"test_relative_l2": 0.045244683949276804,
|
| 13 |
+
"test_squared_relative_l2": 0.002598048033511887,
|
| 14 |
+
"seconds": 2.557273573242128,
|
| 15 |
+
"selected": true
|
| 16 |
+
},
|
| 17 |
+
"wall_seconds": 1321.4138603489846
|
| 18 |
+
}
|
results/verification.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"status": "passed",
|
| 3 |
+
"completed_epochs": 500,
|
| 4 |
+
"history_rows": 500,
|
| 5 |
+
"test_samples": 200,
|
| 6 |
+
"original_scale_metric_recomputed": true,
|
| 7 |
+
"checkpoint_selection_verified": true,
|
| 8 |
+
"figure_sample_ids": [
|
| 9 |
+
0,
|
| 10 |
+
98,
|
| 11 |
+
34
|
| 12 |
+
],
|
| 13 |
+
"figures": [
|
| 14 |
+
"darcy_fields.png",
|
| 15 |
+
"darcy_fields.pdf",
|
| 16 |
+
"training_curves.png",
|
| 17 |
+
"error_distribution.png"
|
| 18 |
+
],
|
| 19 |
+
"optimizer_steps_per_parameter": 16000,
|
| 20 |
+
"scheduler_final_epoch": 500,
|
| 21 |
+
"independent_numpy_metric_verified": true,
|
| 22 |
+
"independent_metric_difference": -2.0816681711721685e-17,
|
| 23 |
+
"per_sample_csv_verified": true,
|
| 24 |
+
"visual_inspection": {
|
| 25 |
+
"darcy_fields.png": "passed: visible original32x32 fields, shared truth/prediction scales, fixed/median/worst samples",
|
| 26 |
+
"training_curves.png": "passed:500-epoch curves, both metric conventions and paper reference shown"
|
| 27 |
+
},
|
| 28 |
+
"all_slurm_jobs_exit_zero": true,
|
| 29 |
+
"source_hashes_unchanged": true,
|
| 30 |
+
"numerical_match_to_paper_claimed": false
|
| 31 |
+
}
|
scripts/common.py
ADDED
|
@@ -0,0 +1,150 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Configuration, original-data loading and explicit pressure-scale transforms."""
|
| 2 |
+
|
| 3 |
+
import argparse
|
| 4 |
+
import hashlib
|
| 5 |
+
import json
|
| 6 |
+
import os
|
| 7 |
+
from pathlib import Path
|
| 8 |
+
import random
|
| 9 |
+
import sys
|
| 10 |
+
|
| 11 |
+
import numpy as np
|
| 12 |
+
import torch
|
| 13 |
+
import yaml
|
| 14 |
+
|
| 15 |
+
PACKAGE_ROOT = Path(__file__).resolve().parents[1]
|
| 16 |
+
sys.path.insert(0, str(PACKAGE_ROOT))
|
| 17 |
+
from model.SirenFNO import SirenFNO
|
| 18 |
+
|
| 19 |
+
|
| 20 |
+
def arguments(description):
|
| 21 |
+
parser = argparse.ArgumentParser(description=description)
|
| 22 |
+
parser.add_argument("--config", type=Path, default=PACKAGE_ROOT / "config/config.yaml")
|
| 23 |
+
return parser
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
def load_config(path):
|
| 27 |
+
with open(path) as stream:
|
| 28 |
+
config = yaml.safe_load(stream)
|
| 29 |
+
assert config["training"]["epochs"] == config["training"]["t_max"] == 500
|
| 30 |
+
assert config["training"]["batch_size"] == 32
|
| 31 |
+
assert config["training"]["early_stopping"] is False
|
| 32 |
+
assert config["model"]["width"] == 32 and config["model"]["layers"] == 4
|
| 33 |
+
assert config["model"]["lifting_channels"] == config["model"]["projection_channels"] == 64
|
| 34 |
+
assert not config["model"]["truncation"] and config["model"]["factorization"] == "dense"
|
| 35 |
+
assert config["data"]["n_train"] == 1000 and config["data"]["n_test"] == 200
|
| 36 |
+
assert config["data"]["resolution"] == [32, 32]
|
| 37 |
+
return config
|
| 38 |
+
|
| 39 |
+
|
| 40 |
+
def configure_runtime(config):
|
| 41 |
+
torch.set_num_threads(config["runtime"]["cpu_threads"])
|
| 42 |
+
device = torch.device(config["runtime"]["device"])
|
| 43 |
+
if device.type != "cuda" or not torch.cuda.is_available():
|
| 44 |
+
raise RuntimeError("Allocated accelerator is required; CPU fallback is disabled")
|
| 45 |
+
seed = config["training"]["seed"]
|
| 46 |
+
random.seed(seed)
|
| 47 |
+
np.random.seed(seed)
|
| 48 |
+
torch.manual_seed(seed)
|
| 49 |
+
torch.cuda.manual_seed_all(seed)
|
| 50 |
+
torch.backends.cudnn.benchmark = False
|
| 51 |
+
torch.backends.cudnn.deterministic = True
|
| 52 |
+
print(f"Slurm job={os.environ.get('SLURM_JOB_ID', 'none')} node={os.uname().nodename} "
|
| 53 |
+
f"torch={torch.__version__} HIP={torch.version.hip} "
|
| 54 |
+
f"accelerator={torch.cuda.get_device_name(device)} seed={seed}", flush=True)
|
| 55 |
+
return device
|
| 56 |
+
|
| 57 |
+
|
| 58 |
+
def sha256(path):
|
| 59 |
+
digest = hashlib.sha256()
|
| 60 |
+
with open(path, "rb") as stream:
|
| 61 |
+
for block in iter(lambda: stream.read(1024 * 1024), b""):
|
| 62 |
+
digest.update(block)
|
| 63 |
+
return digest.hexdigest()
|
| 64 |
+
|
| 65 |
+
|
| 66 |
+
def atomic_json(path, content):
|
| 67 |
+
path = Path(path)
|
| 68 |
+
path.parent.mkdir(parents=True, exist_ok=True)
|
| 69 |
+
temporary = path.with_name(path.name + ".tmp")
|
| 70 |
+
temporary.write_text(json.dumps(content, indent=2, ensure_ascii=False, allow_nan=False) + "\n")
|
| 71 |
+
temporary.replace(path)
|
| 72 |
+
|
| 73 |
+
|
| 74 |
+
def load_data(config, record_path=None):
|
| 75 |
+
settings = config["data"]
|
| 76 |
+
tensors, manifest = {}, {"selection": "first_n", "no_download": True, "files": {}}
|
| 77 |
+
for split, number in [("train", settings["n_train"]), ("test", settings["n_test"])]:
|
| 78 |
+
path = Path(settings["root"]) / settings[f"{split}_file"]
|
| 79 |
+
raw = torch.load(path, map_location="cpu", weights_only=False)
|
| 80 |
+
info = {"path": str(path), "sha256": sha256(path), "selected_indices": [0, number - 1]}
|
| 81 |
+
for key in ("x", "y"):
|
| 82 |
+
value = raw[key]
|
| 83 |
+
if value.ndim != 3 or list(value.shape[-2:]) != settings["resolution"]:
|
| 84 |
+
raise ValueError(f"Unexpected original shape {path}:{key} {value.shape}")
|
| 85 |
+
if value.shape[0] < number or value.dtype != torch.float32:
|
| 86 |
+
raise ValueError(f"Unexpected original sample count or dtype: {path}:{key}")
|
| 87 |
+
selected = value[:number].unsqueeze(1).contiguous()
|
| 88 |
+
if not torch.isfinite(selected).all():
|
| 89 |
+
raise ValueError(f"Nonfinite data {path}:{key}")
|
| 90 |
+
tensors[f"{split}_{key}"] = selected
|
| 91 |
+
info[key] = {"raw_shape": list(value.shape), "raw_dtype": str(value.dtype),
|
| 92 |
+
"selected_shape": list(selected.shape), "min": selected.min().item(),
|
| 93 |
+
"max": selected.max().item(), "meaning": "permeability" if key == "x" else "pressure"}
|
| 94 |
+
manifest["files"][split] = info
|
| 95 |
+
print(f"DATA {split}: {json.dumps(info)}", flush=True)
|
| 96 |
+
if record_path:
|
| 97 |
+
atomic_json(record_path, manifest)
|
| 98 |
+
return tensors, manifest
|
| 99 |
+
|
| 100 |
+
|
| 101 |
+
def normalizer_from_train(config, y):
|
| 102 |
+
if config["data"]["normalize_output"]:
|
| 103 |
+
mean = y.mean().item()
|
| 104 |
+
scale = y.std(correction=1).item() + config["data"]["normalization_epsilon"]
|
| 105 |
+
else:
|
| 106 |
+
mean, scale = 0.0, 1.0
|
| 107 |
+
return {"mean": mean, "scale": scale, "fit_on": "train[0:1000] only", "input_normalized": False}
|
| 108 |
+
|
| 109 |
+
|
| 110 |
+
def decode(prediction, normalizer):
|
| 111 |
+
return prediction * normalizer["scale"] + normalizer["mean"]
|
| 112 |
+
|
| 113 |
+
|
| 114 |
+
def relative_l2(prediction, target):
|
| 115 |
+
difference = (prediction - target).flatten(1)
|
| 116 |
+
denominator = torch.linalg.vector_norm(target.flatten(1), dim=1)
|
| 117 |
+
if torch.any(denominator <= 0):
|
| 118 |
+
raise ValueError("Relative L2 is undefined for zero-norm pressure sample")
|
| 119 |
+
return torch.linalg.vector_norm(difference, dim=1) / denominator
|
| 120 |
+
|
| 121 |
+
|
| 122 |
+
@torch.inference_mode()
|
| 123 |
+
def evaluate(model, x, y, normalizer, batch_size, device, return_predictions=False):
|
| 124 |
+
model.eval()
|
| 125 |
+
errors, predictions = [], []
|
| 126 |
+
for start in range(0, len(x), batch_size):
|
| 127 |
+
prediction = decode(model(x[start:start + batch_size].to(device)), normalizer)
|
| 128 |
+
target = y[start:start + batch_size].to(device)
|
| 129 |
+
errors.append(relative_l2(prediction, target).cpu())
|
| 130 |
+
if return_predictions:
|
| 131 |
+
predictions.append(prediction.cpu())
|
| 132 |
+
errors = torch.cat(errors)
|
| 133 |
+
if not torch.isfinite(errors).all():
|
| 134 |
+
raise FloatingPointError("Nonfinite evaluation errors")
|
| 135 |
+
result = {"relative_l2_mean": errors.double().mean().item(),
|
| 136 |
+
"squared_relative_l2_mean": errors.double().square().mean().item(),
|
| 137 |
+
"relative_l2_std": errors.double().std(correction=0).item()}
|
| 138 |
+
return result, errors, torch.cat(predictions) if return_predictions else None
|
| 139 |
+
|
| 140 |
+
|
| 141 |
+
def build_model(config, device):
|
| 142 |
+
return SirenFNO(**config["model"]).to(device)
|
| 143 |
+
|
| 144 |
+
|
| 145 |
+
def save_checkpoint(path, checkpoint):
|
| 146 |
+
path = Path(path)
|
| 147 |
+
path.parent.mkdir(parents=True, exist_ok=True)
|
| 148 |
+
temporary = path.with_name(path.name + ".tmp")
|
| 149 |
+
torch.save(checkpoint, temporary)
|
| 150 |
+
temporary.replace(path)
|
scripts/inference.py
ADDED
|
@@ -0,0 +1,72 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Reload saved weights and evaluate 200 original-scale 32x32 pressure fields."""
|
| 2 |
+
|
| 3 |
+
import csv
|
| 4 |
+
from copy import deepcopy
|
| 5 |
+
from pathlib import Path
|
| 6 |
+
|
| 7 |
+
import torch
|
| 8 |
+
|
| 9 |
+
from common import (arguments, atomic_json, build_model, configure_runtime,
|
| 10 |
+
evaluate, load_config, load_data, relative_l2, sha256)
|
| 11 |
+
|
| 12 |
+
|
| 13 |
+
def main():
|
| 14 |
+
parser = arguments(__doc__)
|
| 15 |
+
parser.add_argument("--checkpoint", type=Path)
|
| 16 |
+
args = parser.parse_args()
|
| 17 |
+
config = load_config(args.config)
|
| 18 |
+
device = configure_runtime(config)
|
| 19 |
+
output = Path(config["output"]["root"])
|
| 20 |
+
results = output / config["output"]["results"]
|
| 21 |
+
checkpoint_path = args.checkpoint or output / config["output"]["checkpoint"]
|
| 22 |
+
last = torch.load(output / config["output"]["last_checkpoint"], map_location="cpu", weights_only=False)
|
| 23 |
+
assert last["epoch"] == config["training"]["epochs"] == 500
|
| 24 |
+
checkpoint = torch.load(checkpoint_path, map_location="cpu", weights_only=False)
|
| 25 |
+
trained_config = deepcopy(checkpoint["config"])
|
| 26 |
+
trained_config["data"]["root"] = config["data"]["root"]
|
| 27 |
+
trained_config["output"]["root"] = config["output"]["root"]
|
| 28 |
+
if trained_config != config:
|
| 29 |
+
raise ValueError("Inference configuration does not match trained checkpoint")
|
| 30 |
+
data, manifest = load_data(config)
|
| 31 |
+
for split, info in manifest["files"].items():
|
| 32 |
+
assert info["sha256"] == checkpoint["data_sha256"][split], "Dataset changed after training"
|
| 33 |
+
model = build_model(config, device)
|
| 34 |
+
model.load_state_dict(checkpoint["model_state_dict"], strict=True)
|
| 35 |
+
metrics, errors, predictions = evaluate(model, data["test_x"], data["test_y"],
|
| 36 |
+
checkpoint["normalizer"], config["training"]["batch_size"], device, return_predictions=True)
|
| 37 |
+
# Recompute using saved physical fields in float64, avoiding batch-average or
|
| 38 |
+
# normalized-target metric ambiguities in the final reported value.
|
| 39 |
+
errors64 = relative_l2(predictions.double(), data["test_y"].double())
|
| 40 |
+
metrics["relative_l2_mean"] = errors64.mean().item()
|
| 41 |
+
metrics["squared_relative_l2_mean"] = errors64.square().mean().item()
|
| 42 |
+
metrics["relative_l2_std"] = errors64.std(correction=0).item()
|
| 43 |
+
reference = config["paper_reference"]
|
| 44 |
+
metrics.update({"n_test": len(errors), "resolution": [32, 32], "scale": "original pressure",
|
| 45 |
+
"aggregation": "arithmetic mean of per-sample Euclidean norm ratios",
|
| 46 |
+
"completed_epochs": last["epoch"], "selected_epoch": checkpoint["epoch"],
|
| 47 |
+
"checkpoint": str(checkpoint_path), "checkpoint_sha256": sha256(checkpoint_path),
|
| 48 |
+
"checkpoint_selection": checkpoint["selection_rule"], "paper_table2": reference,
|
| 49 |
+
"signed_gap": metrics["relative_l2_mean"] - reference,
|
| 50 |
+
"absolute_gap": abs(metrics["relative_l2_mean"] - reference),
|
| 51 |
+
"relative_gap_percent": 100 * (metrics["relative_l2_mean"] / reference - 1),
|
| 52 |
+
"last_epoch_test_relative_l2": last["test_metrics"]["relative_l2_mean"],
|
| 53 |
+
"test_used_for_selection": False,
|
| 54 |
+
"loss_metric_caveat": "Paper Eq15 is squared; this report's primary metric is unsquared as requested"})
|
| 55 |
+
torch.save({"x": data["test_x"], "y": data["test_y"], "prediction": predictions,
|
| 56 |
+
"sample_ids": torch.arange(len(errors)), "relative_l2": errors64,
|
| 57 |
+
"checkpoint_sha256": metrics["checkpoint_sha256"]}, results / "predictions.pt")
|
| 58 |
+
with (results / "per_sample_metrics.csv").open("w", newline="") as stream:
|
| 59 |
+
writer = csv.writer(stream)
|
| 60 |
+
writer.writerow(["sample_id", "relative_l2_original_scale", "squared_relative_l2_original_scale"])
|
| 61 |
+
for index, error in enumerate(errors64.tolist()):
|
| 62 |
+
writer.writerow([index, error, error * error])
|
| 63 |
+
atomic_json(results / "metrics.json", metrics)
|
| 64 |
+
print(f"RELOADED {checkpoint_path} epoch={checkpoint['epoch']}; actual training=500 epochs", flush=True)
|
| 65 |
+
print(f"INFERENCE original-scale per-sample relative L2 mean={metrics['relative_l2_mean']:.10f}; "
|
| 66 |
+
f"paper=0.0351; signed_gap={metrics['signed_gap']:+.10f}; "
|
| 67 |
+
f"relative_gap={metrics['relative_gap_percent']:+.3f}%", flush=True)
|
| 68 |
+
print(f"Squared relative L2 mean (paper Eq15)={metrics['squared_relative_l2_mean']:.10f}", flush=True)
|
| 69 |
+
|
| 70 |
+
|
| 71 |
+
if __name__ == "__main__":
|
| 72 |
+
main()
|
scripts/result.py
ADDED
|
@@ -0,0 +1,132 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Plot real saved inference fields and summarize the completed experiment."""
|
| 2 |
+
|
| 3 |
+
import csv
|
| 4 |
+
import json
|
| 5 |
+
from pathlib import Path
|
| 6 |
+
|
| 7 |
+
import matplotlib
|
| 8 |
+
matplotlib.use("Agg")
|
| 9 |
+
import matplotlib.pyplot as plt
|
| 10 |
+
import numpy as np
|
| 11 |
+
import torch
|
| 12 |
+
|
| 13 |
+
from common import arguments, atomic_json, load_config, relative_l2, sha256
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
def main():
|
| 17 |
+
args = arguments(__doc__).parse_args()
|
| 18 |
+
config = load_config(args.config)
|
| 19 |
+
root = Path(config["output"]["root"])
|
| 20 |
+
results = root / config["output"]["results"]
|
| 21 |
+
metrics = json.loads((results / "metrics.json").read_text())
|
| 22 |
+
saved = torch.load(results / "predictions.pt", map_location="cpu", weights_only=False)
|
| 23 |
+
assert saved["checkpoint_sha256"] == metrics["checkpoint_sha256"]
|
| 24 |
+
assert metrics["checkpoint_sha256"] == sha256(metrics["checkpoint"])
|
| 25 |
+
assert saved["prediction"].shape == saved["y"].shape == (200, 1, 32, 32)
|
| 26 |
+
errors = relative_l2(saved["prediction"].double(), saved["y"].double()).numpy()
|
| 27 |
+
assert np.isfinite(errors).all() and abs(errors.mean() - metrics["relative_l2_mean"]) < 1e-12
|
| 28 |
+
with (results / "training_history.csv").open() as stream:
|
| 29 |
+
history = list(csv.DictReader(stream))
|
| 30 |
+
assert [int(row["epoch"]) for row in history] == list(range(1, 501))
|
| 31 |
+
training_losses = np.array([float(row["train_squared_relative_l2"]) for row in history])
|
| 32 |
+
assert int(training_losses.argmin()) + 1 == metrics["selected_epoch"]
|
| 33 |
+
selected = [0, int(np.argsort(errors)[len(errors) // 2]), int(errors.argmax())]
|
| 34 |
+
fig, axes = plt.subplots(3, 4, figsize=(14, 9), constrained_layout=True)
|
| 35 |
+
for row, sample_id in enumerate(selected):
|
| 36 |
+
permeability = saved["x"][sample_id, 0].numpy()
|
| 37 |
+
truth = saved["y"][sample_id, 0].numpy()
|
| 38 |
+
prediction = saved["prediction"][sample_id, 0].numpy()
|
| 39 |
+
difference = np.abs(prediction - truth)
|
| 40 |
+
lower, upper = min(truth.min(), prediction.min()), max(truth.max(), prediction.max())
|
| 41 |
+
for column, (array, title, cmap, vmin, vmax) in enumerate([
|
| 42 |
+
(permeability, "Permeability", "viridis", permeability.min(), permeability.max()),
|
| 43 |
+
(truth, "True pressure", "viridis", lower, upper),
|
| 44 |
+
(prediction, "Predicted pressure", "viridis", lower, upper),
|
| 45 |
+
(difference, "Absolute error", "inferno", 0, difference.max()),
|
| 46 |
+
]):
|
| 47 |
+
ax = axes[row, column]
|
| 48 |
+
im = ax.imshow(array.T, origin="lower", extent=[0, 1, 0, 1],
|
| 49 |
+
interpolation="nearest", cmap=cmap, vmin=vmin, vmax=vmax)
|
| 50 |
+
ax.set_title(f"{title}: sample {sample_id}" if column != 3 else
|
| 51 |
+
f"{title}: rel L2={errors[sample_id]:.4f}")
|
| 52 |
+
ax.set_xlabel("Coordinate 1")
|
| 53 |
+
ax.set_ylabel("Coordinate 2")
|
| 54 |
+
fig.colorbar(im, ax=ax, shrink=.8, format="%.3g")
|
| 55 |
+
fig.suptitle("SirenFNO Darcy 32x32: original-scale fields; fixed, median-error and worst samples")
|
| 56 |
+
fig.savefig(results / "darcy_fields.png", dpi=180)
|
| 57 |
+
fig.savefig(results / "darcy_fields.pdf")
|
| 58 |
+
plt.close(fig)
|
| 59 |
+
epochs = np.arange(1, 501)
|
| 60 |
+
fig, axes = plt.subplots(1, 2, figsize=(12, 4), constrained_layout=True)
|
| 61 |
+
for name, label in [("train_relative_l2", "Train, original scale"), ("test_relative_l2", "Test, original scale")]:
|
| 62 |
+
axes[0].semilogy(epochs, [float(row[name]) for row in history], label=label)
|
| 63 |
+
axes[0].axhline(.0351, color="black", ls="--", label="Paper Table 2: 0.0351")
|
| 64 |
+
axes[0].axvline(metrics["selected_epoch"], color="grey", ls=":", label="Training-selected checkpoint")
|
| 65 |
+
axes[0].set(xlabel="Epoch", ylabel="Mean per-sample relative L2")
|
| 66 |
+
axes[0].legend(fontsize=8)
|
| 67 |
+
axes[1].semilogy(epochs, training_losses, label="Full training set, paper Eq15")
|
| 68 |
+
axes[1].semilogy(epochs, [float(row["test_squared_relative_l2"]) for row in history], label="Test, squared relative L2")
|
| 69 |
+
axes[1].set(xlabel="Epoch", ylabel="Mean squared per-sample relative L2")
|
| 70 |
+
axes[1].legend(fontsize=8)
|
| 71 |
+
fig.savefig(results / "training_curves.png", dpi=180)
|
| 72 |
+
plt.close(fig)
|
| 73 |
+
fig, ax = plt.subplots(figsize=(7, 4), constrained_layout=True)
|
| 74 |
+
ax.hist(errors, bins=25, color="#317a9b", edgecolor="white")
|
| 75 |
+
ax.axvline(errors.mean(), color="#c45b35", label=f"Run mean {errors.mean():.6f}")
|
| 76 |
+
ax.axvline(.0351, color="black", ls="--", label="Paper Table 2: 0.0351")
|
| 77 |
+
ax.set(xlabel="Per-sample relative L2, original scale", ylabel="Test samples")
|
| 78 |
+
ax.legend()
|
| 79 |
+
fig.savefig(results / "error_distribution.png", dpi=180)
|
| 80 |
+
plt.close(fig)
|
| 81 |
+
metadata = json.loads((results / "run_metadata.json").read_text())
|
| 82 |
+
status = json.loads((results / "training_status.json").read_text())
|
| 83 |
+
text = f"""# SirenFNO Table 2 / Darcy 32×32 复现结果
|
| 84 |
+
|
| 85 |
+
正式训练已完成 **{metrics['completed_epochs']} epoch**,batch size 32,训练 1000 个、测试 200 个样本。
|
| 86 |
+
加载 `weight/best_model.pth` 后,原尺度逐样本相对 L2 算术均值为 **{metrics['relative_l2_mean']:.10f}**。
|
| 87 |
+
对比论文 Table 2 的 **0.0351**:有符号差距 **{metrics['signed_gap']:+.10f}**,绝对差距 **{metrics['absolute_gap']:.10f}**,相对差距 **{metrics['relative_gap_percent']:+.3f}%**。
|
| 88 |
+
|
| 89 |
+
## 指标与权重
|
| 90 |
+
|
| 91 |
+
- 主指标:对每个样本计算 `||prediction-y||_2 / ||y||_2`,对 200 个样本取算术均值;由保存的原尺度 float32 场使用 float64 重新计算。
|
| 92 |
+
- 论文式 (15) 字面上的平方相对 L2 均值:**{metrics['squared_relative_l2_mean']:.10f}**。它与主指标不是同一统计量。
|
| 93 |
+
- 最优权重选择:每个 epoch 后在完整训练集上以 eval 模式计算原尺度平方相对 L2,取最小者;选中 **epoch {metrics['selected_epoch']}**。
|
| 94 |
+
- 测试集每轮用于监测,**不用于选权重、调参或早停**;无独立验证集。论文没有说明权重选择规则,本规则属于显式工程假设。
|
| 95 |
+
- 末轮 epoch 500 的原尺度测试相对 L2:**{metrics['last_epoch_test_relative_l2']:.10f}**;末轮权重保存在 `weight/last_model.pth`。
|
| 96 |
+
- checkpoint SHA256:`{metrics['checkpoint_sha256']}`。
|
| 97 |
+
- 实际本轮训练墙钟时间:{status['wall_seconds']:.2f} 秒;参数量:{metadata['parameters']}。
|
| 98 |
+
|
| 99 |
+
## 对齐情况与限制
|
| 100 |
+
|
| 101 |
+
1. 本次是非分解完整 SirenFNO,四层 32 通道,式 (14) 的增强残差结构,全部 32×17 个独立 rFFT 频点生成稠密 32×32 复系数;没有任何频率截断。rFFT 半谱是实数输入的厄米冗余压缩,不是高频截断。
|
| 102 |
+
2. 论文 §5.2 的 AdamW、lr=1e-3、weight decay=1e-4、CosineAnnealingLR(T_max=500)、500 epoch、batch=32 均执行。无提前停止、无降采样、无额外训练数据。预先冒烟训练的模型已丢弃,正式训练从 seed 42 重新初始化。
|
| 103 |
+
3. 论文式 (7) 明确写 sin,本次使用纯 sin 及可学习偏置;官方仓库使用交错 sin/cos 且无偏置,本次没有用该冲突实现覆盖论文。
|
| 104 |
+
4. 论文 §5.2 指定 lifting/projection 隐藏宽度 64,本次使用 3→64→32 和 32→64→1。仓库使用单层 3→32 lifting 和 32→128→1 projection;因此本次参数量不要求等于 Table 3 的约 308.9k。
|
| 105 |
+
5. 论文式 (15) 在 PDF 上是范数平方之比,但文字称相对 L2、仓库 LpLoss 使用未平方比值。本次训练在原尺度使用平方比值以遵循字面公式,报告同时给出用户指定的未平方指标。**无法断言 Table 2 的 0.0351 究竟对应公式还是仓库口径,因此数值差距只是指定口径下的直接比较,不代表已精确匹配作者训练实现。**
|
| 106 |
+
6. 论文未给出的 RFF 特征数 16、sigma=512、SIREN 隐藏宽度 32/一层、omega 初值、midpoint 频率索引、残差 MLP 扩张 4、坐标拼接、无 padding、seed 42、无 shuffle,以及训练集压力 channel-wise 标准化,来自官方仓库只读参考。输入渗透率保持原值;压力标准化统计量仅来自 1000 个训练样本,模型输出先反归一化再计算训练 loss 和评测指标。
|
| 107 |
+
7. 运行设备为 **{metadata['device']}**,PyTorch **{metadata['torch']}**,HIP **{metadata['hip']}**;论文设备是 RTX 5090。后端和随机数流不同,只有一次固定 seed 运行,没有多种子统计。
|
| 108 |
+
8. 物理单位及精确边界网格元数据未在 pt 字段中提供,不虚构单位;可视化坐标采用官方输入网格 [0,1] 的约定,不宣称额外 PDE 残差或守恒量验证。
|
| 109 |
+
|
| 110 |
+
## 证据与文件
|
| 111 |
+
|
| 112 |
+
- `results/data_manifest.json`:原始 shape/dtype、选样索引、字段语义与数据 SHA256。
|
| 113 |
+
- `results/training_history.csv`:500 轮实际记录,含 train/test 两种误差及学习率。
|
| 114 |
+
- `results/per_sample_metrics.csv`:200 个测试样本原尺度误差。
|
| 115 |
+
- `results/predictions.pt`:原尺度输入、真值、预测、样本索引。
|
| 116 |
+
- `results/darcy_fields.png`、`darcy_fields.pdf`:固定第 0 个、误差中位样本、最差样本,真值与预测使用同一色标。
|
| 117 |
+
- `results/training_curves.png`、`error_distribution.png`:训练曲线及误差分布。
|
| 118 |
+
- `results/smoke/checks.json`:六项计算节点冒烟检查。
|
| 119 |
+
- `reproduction_spec.md`:完整来源、缺口与冲突决策;`logs/`:Slurm 实时 print 日志。
|
| 120 |
+
|
| 121 |
+
论文:[IJCAI PDF](https://ijcai.org/proceedings/2026/0535.pdf)。补充细节:[官方仓库只读参考](https://github.com/pengqingshi/SirenFNO)。未下载、克隆或复制官方仓库代码,所有实现均独立编写;未提交 Git、未上传平台。
|
| 122 |
+
"""
|
| 123 |
+
(results / "report.md").write_text(text)
|
| 124 |
+
atomic_json(results / "verification.json", {"status": "passed", "completed_epochs": 500,
|
| 125 |
+
"history_rows": 500, "test_samples": 200, "original_scale_metric_recomputed": True,
|
| 126 |
+
"checkpoint_selection_verified": True, "figure_sample_ids": selected,
|
| 127 |
+
"figures": ["darcy_fields.png", "darcy_fields.pdf", "training_curves.png", "error_distribution.png"]})
|
| 128 |
+
print(f"RESULTS VERIFIED: {results / 'report.md'}", flush=True)
|
| 129 |
+
|
| 130 |
+
|
| 131 |
+
if __name__ == "__main__":
|
| 132 |
+
main()
|
scripts/smoke.py
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Six compute-node checks using only actual supplied Darcy samples."""
|
| 2 |
+
|
| 3 |
+
from pathlib import Path
|
| 4 |
+
import torch
|
| 5 |
+
from torch.utils.data import DataLoader, TensorDataset
|
| 6 |
+
from common import (arguments, atomic_json, build_model, configure_runtime, decode,
|
| 7 |
+
evaluate, load_config, load_data, normalizer_from_train, relative_l2)
|
| 8 |
+
|
| 9 |
+
|
| 10 |
+
def main():
|
| 11 |
+
args = arguments(__doc__).parse_args()
|
| 12 |
+
config = load_config(args.config)
|
| 13 |
+
device = configure_runtime(config)
|
| 14 |
+
data, _ = load_data(config)
|
| 15 |
+
normalizer = normalizer_from_train(config, data["train_y"])
|
| 16 |
+
model = build_model(config, device)
|
| 17 |
+
checks = {}
|
| 18 |
+
x, y = data["train_x"][:32].to(device), data["train_y"][:32].to(device)
|
| 19 |
+
prediction = decode(model(x), normalizer)
|
| 20 |
+
assert prediction.shape == y.shape and torch.isfinite(prediction).all()
|
| 21 |
+
checks["forward_pass"] = "passed"
|
| 22 |
+
loss = relative_l2(prediction, y).square().mean()
|
| 23 |
+
loss.backward()
|
| 24 |
+
assert all(p.grad is not None and torch.isfinite(p.grad).all() for p in model.parameters() if p.requires_grad)
|
| 25 |
+
checks["backward_pass"] = "passed"
|
| 26 |
+
optimizer = torch.optim.AdamW(model.parameters(), lr=0.001, weight_decay=0.0001)
|
| 27 |
+
losses = []
|
| 28 |
+
for bx, by in DataLoader(TensorDataset(data["train_x"], data["train_y"]), batch_size=32):
|
| 29 |
+
optimizer.zero_grad(set_to_none=True)
|
| 30 |
+
prediction = decode(model(bx.to(device)), normalizer)
|
| 31 |
+
batch_loss = relative_l2(prediction, by.to(device)).square().mean()
|
| 32 |
+
assert torch.isfinite(batch_loss)
|
| 33 |
+
batch_loss.backward()
|
| 34 |
+
optimizer.step()
|
| 35 |
+
losses.append(batch_loss.item())
|
| 36 |
+
assert sum(losses[-8:]) / 8 < sum(losses[:8]) / 8
|
| 37 |
+
checks["train_loop_dry_run"] = "passed"
|
| 38 |
+
metrics, errors, pred = evaluate(model, data["test_x"][:32], data["test_y"][:32],
|
| 39 |
+
normalizer, 32, device, return_predictions=True)
|
| 40 |
+
assert torch.isfinite(errors).all()
|
| 41 |
+
checks["validation_loop_dry_run"] = "passed"
|
| 42 |
+
# Independent manual metric computation on real fields; no synthetic data.
|
| 43 |
+
manual = ((pred.double() - data["test_y"][:32].double()).flatten(1).square().sum(1).sqrt()
|
| 44 |
+
/ data["test_y"][:32].double().flatten(1).square().sum(1).sqrt())
|
| 45 |
+
assert torch.allclose(errors.double(), manual, rtol=1e-5, atol=1e-8)
|
| 46 |
+
checks["domain_metric_dry_run"] = "passed"
|
| 47 |
+
assert len(model.blocks) == 4 and model.projection.layers[0].out_channels == 64
|
| 48 |
+
for block in model.blocks:
|
| 49 |
+
assert tuple(block.spectral.coordinates.shape) == (32, 17, 2)
|
| 50 |
+
checks["config_consistency"] = "passed"
|
| 51 |
+
path = Path(config["output"]["root"]) / config["output"]["results"] / "smoke/checks.json"
|
| 52 |
+
atomic_json(path, {"status": "passed", "attempts": 1, "checks": checks,
|
| 53 |
+
"used_original_data": True, "formal_epochs_consumed": 0,
|
| 54 |
+
"loss_first8_mean": sum(losses[:8]) / 8, "loss_last8_mean": sum(losses[-8:]) / 8,
|
| 55 |
+
"validation_metric": metrics, "peak_memory_bytes": torch.cuda.max_memory_allocated()})
|
| 56 |
+
print("SMOKE: all six checks passed; discarded smoke model; formal run starts from fresh seed", flush=True)
|
| 57 |
+
|
| 58 |
+
|
| 59 |
+
if __name__ == "__main__":
|
| 60 |
+
main()
|
scripts/train.py
ADDED
|
@@ -0,0 +1,139 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Full 500-epoch paper-priority Darcy experiment; see reproduction_spec.md."""
|
| 2 |
+
|
| 3 |
+
import csv
|
| 4 |
+
import json
|
| 5 |
+
from pathlib import Path
|
| 6 |
+
import time
|
| 7 |
+
|
| 8 |
+
import torch
|
| 9 |
+
from torch.utils.data import DataLoader, TensorDataset
|
| 10 |
+
|
| 11 |
+
from common import (arguments, atomic_json, build_model, configure_runtime, decode,
|
| 12 |
+
evaluate, load_config, load_data, normalizer_from_train,
|
| 13 |
+
relative_l2, save_checkpoint)
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
def main():
|
| 17 |
+
parser = arguments(__doc__)
|
| 18 |
+
parser.add_argument("--resume", type=Path, help="Resume model, optimizer, scheduler and RNG from last checkpoint")
|
| 19 |
+
args = parser.parse_args()
|
| 20 |
+
config = load_config(args.config)
|
| 21 |
+
device = configure_runtime(config)
|
| 22 |
+
output = Path(config["output"]["root"])
|
| 23 |
+
results = output / config["output"]["results"]
|
| 24 |
+
results.mkdir(parents=True, exist_ok=True)
|
| 25 |
+
smoke_path = results / "smoke/checks.json"
|
| 26 |
+
if not smoke_path.exists() or json.loads(smoke_path.read_text())["status"] != "passed":
|
| 27 |
+
raise RuntimeError("All six compute-node smoke checks must pass before full training")
|
| 28 |
+
tensors, manifest = load_data(config, results / "data_manifest.json")
|
| 29 |
+
normalizer = normalizer_from_train(config, tensors["train_y"])
|
| 30 |
+
atomic_json(results / "normalization.json", normalizer)
|
| 31 |
+
model = build_model(config, device)
|
| 32 |
+
train = config["training"]
|
| 33 |
+
loader = DataLoader(TensorDataset(tensors["train_x"], tensors["train_y"]),
|
| 34 |
+
batch_size=train["batch_size"], shuffle=train["shuffle"],
|
| 35 |
+
drop_last=train["drop_last"], num_workers=0)
|
| 36 |
+
optimizer = torch.optim.AdamW(model.parameters(), lr=train["lr"],
|
| 37 |
+
weight_decay=train["weight_decay"],
|
| 38 |
+
betas=tuple(train["betas"]), eps=train["optimizer_epsilon"])
|
| 39 |
+
scheduler = torch.optim.lr_scheduler.CosineAnnealingLR(
|
| 40 |
+
optimizer, T_max=train["t_max"], eta_min=train["eta_min"])
|
| 41 |
+
first_epoch, best_score, best_epoch, history = 1, float("inf"), None, []
|
| 42 |
+
if args.resume:
|
| 43 |
+
restored = torch.load(args.resume, map_location="cpu", weights_only=False)
|
| 44 |
+
if restored["config"] != config:
|
| 45 |
+
raise ValueError("Resume configuration differs from checkpoint")
|
| 46 |
+
model.load_state_dict(restored["model_state_dict"], strict=True)
|
| 47 |
+
optimizer.load_state_dict(restored["optimizer_state_dict"])
|
| 48 |
+
scheduler.load_state_dict(restored["scheduler_state_dict"])
|
| 49 |
+
first_epoch = restored["epoch"] + 1
|
| 50 |
+
best_score, best_epoch = restored["best_score"], restored["best_epoch"]
|
| 51 |
+
history = restored["history"]
|
| 52 |
+
torch.set_rng_state(restored["torch_rng_state"])
|
| 53 |
+
torch.cuda.set_rng_state_all(restored["cuda_rng_state"])
|
| 54 |
+
else:
|
| 55 |
+
for name in ("checkpoint", "last_checkpoint"):
|
| 56 |
+
if (output / config["output"][name]).exists():
|
| 57 |
+
raise FileExistsError("Existing experiment weights: use --resume or a separate output directory")
|
| 58 |
+
parameters = sum(p.numel() for p in model.parameters())
|
| 59 |
+
print(f"MODEL parameters={parameters}, dense rFFT kernel grid=32x17, full 500 epochs", flush=True)
|
| 60 |
+
print("LOSS: original-scale mean squared sample-relative L2, paper eq. (15). "
|
| 61 |
+
"Checkpoint: minimum full-training-set loss; test set never selects weights.", flush=True)
|
| 62 |
+
start_time = time.perf_counter()
|
| 63 |
+
runtime = {"parameters": parameters, "torch": torch.__version__, "hip": torch.version.hip,
|
| 64 |
+
"device": torch.cuda.get_device_name(device), "selection": train["checkpoint_selection"],
|
| 65 |
+
"seed": train["seed"], "data_sha256": {k: v["sha256"] for k, v in manifest["files"].items()}}
|
| 66 |
+
atomic_json(results / "run_metadata.json", runtime)
|
| 67 |
+
fields = ["epoch", "lr", "online_train_loss", "train_relative_l2", "train_squared_relative_l2",
|
| 68 |
+
"test_relative_l2", "test_squared_relative_l2", "seconds", "selected"]
|
| 69 |
+
csv_path = results / "training_history.csv"
|
| 70 |
+
if history:
|
| 71 |
+
with csv_path.open("w", newline="") as stream:
|
| 72 |
+
writer = csv.DictWriter(stream, fieldnames=fields)
|
| 73 |
+
writer.writeheader()
|
| 74 |
+
writer.writerows(history)
|
| 75 |
+
for epoch in range(first_epoch, train["epochs"] + 1):
|
| 76 |
+
epoch_start = time.perf_counter()
|
| 77 |
+
model.train()
|
| 78 |
+
learning_rate = optimizer.param_groups[0]["lr"]
|
| 79 |
+
total, count = 0.0, 0
|
| 80 |
+
for batch_index, (x, y) in enumerate(loader):
|
| 81 |
+
x, y = x.to(device), y.to(device)
|
| 82 |
+
optimizer.zero_grad(set_to_none=True)
|
| 83 |
+
prediction = decode(model(x), normalizer)
|
| 84 |
+
sample_errors = relative_l2(prediction, y)
|
| 85 |
+
loss = sample_errors.square().mean()
|
| 86 |
+
if not torch.isfinite(loss):
|
| 87 |
+
raise FloatingPointError(f"Nonfinite loss at epoch {epoch} batch {batch_index}")
|
| 88 |
+
loss.backward()
|
| 89 |
+
optimizer.step()
|
| 90 |
+
value = loss.item()
|
| 91 |
+
total += value * len(x)
|
| 92 |
+
count += len(x)
|
| 93 |
+
if batch_index % train["print_batch_interval"] == 0 or batch_index == len(loader) - 1:
|
| 94 |
+
print(f"Epoch {epoch:03d}/500 batch {batch_index + 1:02d}/{len(loader)} "
|
| 95 |
+
f"loss={value:.8f} sample_rel_l2={sample_errors.mean().item():.8f}", flush=True)
|
| 96 |
+
training, _, _ = evaluate(model, tensors["train_x"], tensors["train_y"],
|
| 97 |
+
normalizer, train["batch_size"], device)
|
| 98 |
+
testing, _, _ = evaluate(model, tensors["test_x"], tensors["test_y"],
|
| 99 |
+
normalizer, train["batch_size"], device)
|
| 100 |
+
score = training["squared_relative_l2_mean"]
|
| 101 |
+
improved = score < best_score
|
| 102 |
+
if improved:
|
| 103 |
+
best_score, best_epoch = score, epoch
|
| 104 |
+
scheduler.step()
|
| 105 |
+
row = {"epoch": epoch, "lr": learning_rate, "online_train_loss": total / count,
|
| 106 |
+
"train_relative_l2": training["relative_l2_mean"], "train_squared_relative_l2": score,
|
| 107 |
+
"test_relative_l2": testing["relative_l2_mean"],
|
| 108 |
+
"test_squared_relative_l2": testing["squared_relative_l2_mean"],
|
| 109 |
+
"seconds": time.perf_counter() - epoch_start, "selected": improved}
|
| 110 |
+
history.append(row)
|
| 111 |
+
with csv_path.open("a", newline="") as stream:
|
| 112 |
+
writer = csv.DictWriter(stream, fieldnames=fields)
|
| 113 |
+
if epoch == 1:
|
| 114 |
+
writer.writeheader()
|
| 115 |
+
writer.writerow(row)
|
| 116 |
+
checkpoint = {"model_state_dict": model.state_dict(), "config": config,
|
| 117 |
+
"epoch": epoch, "completed_epochs_at_save": epoch,
|
| 118 |
+
"normalizer": normalizer, "best_score": best_score, "best_epoch": best_epoch,
|
| 119 |
+
"selection_rule": train["checkpoint_selection"], "train_metrics": training,
|
| 120 |
+
"test_metrics": testing, "optimizer_state_dict": optimizer.state_dict(),
|
| 121 |
+
"scheduler_state_dict": scheduler.state_dict(), "history": history,
|
| 122 |
+
"torch_rng_state": torch.get_rng_state(), "cuda_rng_state": torch.cuda.get_rng_state_all(),
|
| 123 |
+
"data_sha256": runtime["data_sha256"]}
|
| 124 |
+
if improved:
|
| 125 |
+
save_checkpoint(output / config["output"]["checkpoint"], checkpoint)
|
| 126 |
+
save_checkpoint(output / config["output"]["last_checkpoint"], checkpoint)
|
| 127 |
+
atomic_json(results / "training_status.json", {"status": "running" if epoch < 500 else "completed",
|
| 128 |
+
"completed_epochs": epoch, "best_epoch": best_epoch, "best_training_loss": best_score,
|
| 129 |
+
"last_epoch": row, "wall_seconds": time.perf_counter() - start_time})
|
| 130 |
+
print(f"EPOCH {epoch:03d}/500 lr={learning_rate:.7g} train_loss={score:.9f} "
|
| 131 |
+
f"train_rel_l2={training['relative_l2_mean']:.8f} "
|
| 132 |
+
f"test_rel_l2={testing['relative_l2_mean']:.8f} "
|
| 133 |
+
f"test_squared_rel_l2={testing['squared_relative_l2_mean']:.9f} "
|
| 134 |
+
f"best_epoch={best_epoch} seconds={row['seconds']:.2f}", flush=True)
|
| 135 |
+
print(f"TRAINING COMPLETE: completed_epochs=500 selected_epoch={best_epoch}", flush=True)
|
| 136 |
+
|
| 137 |
+
|
| 138 |
+
if __name__ == "__main__":
|
| 139 |
+
main()
|
weight/best_model.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ceb2392d5ba68e386f08e0f117ea1705df1f3a0185d3cef0971f07951113ce1e
|
| 3 |
+
size 3807256
|
weight/last_model.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:adca832dee12c6eddfc29ea41c9ebaf85ea73e3a787f51724947069680af06eb
|
| 3 |
+
size 3807256
|