File size: 1,461 Bytes
8b64ae3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
{
  "model_name": "SmaAtUNet",
  "model_type": "smaatunet",
  "architectures": ["SmaAtUNet", "CBAM", "DepthwiseSeparableConv"],
  "framework": "PyTorch",
  "domain": "atmosphere",
  "task": "precipitation-nowcasting",
  "implementation": {
    "entry_point": "model/smaatunet.py",
    "scope": "small attention U-Net with depthwise-separable convolutions for multi-step precipitation regression",
    "train_script": "scripts/train.py",
    "inference_script": "scripts/inference.py",
    "evaluation_script": "scripts/result.py",
    "synthetic_data_script": "scripts/fake_data.py"
  },
  "architecture": {
    "family": "five-level attention U-Net",
    "in_channels": 12,
    "out_channels": 6,
    "base_channels": 8,
    "kernels_per_layer": 2,
    "reduction_ratio": 4,
    "bilinear": true,
    "attention": "channel and spatial CBAM",
    "convolution": "depthwise separable"
  },
  "data": {
    "datasets": ["KNMI precipitation radar maps"],
    "protocol": "smaat_unet_synthetic_precip_v1",
    "format": "NPZ",
    "train_file": "data/train.npz",
    "test_file": "data/test.npz",
    "input_shape": ["N", 12, 288, 288],
    "target_shape": ["N", 6, 288, 288],
    "interval_minutes": 5,
    "forecast_horizon_minutes": 30,
    "required_metadata": ["format_version", "data_source"]
  },
  "configuration_sources": ["conf/config.yaml", "model/smaatunet.py", "scripts/fake_data.py", "scripts/train.py", "scripts/inference.py", "scripts/result.py"]
}