Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+DoubleInputRegressor_lc_3_ssl.json filter=lfs diff=lfs merge=lfs -text
+fit_DoubleInputRegressor_lc_3_ssl.png filter=lfs diff=lfs merge=lfs -text

DoubleInputRegressor_lc_3_ssl.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f2e0af5d8901a2b28aaa2998c525d0c878fbd3810096f26761cbde67b0b4e807
+size 14418305

DoubleInputRegressor_lc_3_ssl.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:03337e9e63170668951746145c5910f4ab08a2e66a095ffab1022fd02226f258
+size 48107861

DoubleInputRegressor_lc_3_ssl_complete_config.json

@@ -0,0 +1,129 @@
+{
+  "model_name": "DoubleInputRegressor",
+  "data_args": {
+    "log_dir": "/data/lightSpec/logs",
+    "dataset": "KeplerDataset",
+    "data_dir": "/data/lightPred/data",
+    "model_name": "DoubleInputRegressor",
+    "batch_size": 16,
+    "num_epochs": 1000,
+    "exp_num": "3_ssl",
+    "max_len_spectra": 4096,
+    "max_days_lc": 270,
+    "max_len_lc": 34560,
+    "lc_freq": 0.0208,
+    "create_umap": false,
+    "masked_transform": false,
+    "use_acf": true,
+    "use_fft": true,
+    "scale_flux": false,
+    "ssl_weight": 0.5,
+    "labels": [
+      "Prot"
+    ]
+  },
+  "model_args": {
+    "encoder_only": true,
+    "stacked_input": true,
+    "in_channels": 3,
+    "load_checkpoint": false,
+    "output_dim": 1,
+    "num_quantiles": 5,
+    "dropout_p": 0.3,
+    "checkpoint_path": "/data/lightSpec/logs/light_2025-03-19/DoubleInputRegressor_lc_1.pth"
+  },
+  "astroconf_args": {
+    "in_channels": 2,
+    "encoder": [
+      "mhsa_pro",
+      "conv"
+    ],
+    "timeshift": false,
+    "num_layers": 5,
+    "num_decoder_layers": 6,
+    "stride": 20,
+    "encoder_dim": 512,
+    "decoder_dim": 128,
+    "num_heads": 8,
+    "kernel_size": 3,
+    "dropout_p": 0.3,
+    "output_dim": 3,
+    "encoder_only": true,
+    "norm": "postnorm",
+    "load_checkpoint": false,
+    "spec_checkpoint_path": "/data/lightSpec/logs/exp6/astroconf_spectra_3.pth"
+  },
+  "cnn_args": {
+    "in_channels": 2,
+    "num_layers": 6,
+    "stride": 1,
+    "encoder_dims": [
+      32,
+      64,
+      128,
+      256,
+      512
+    ],
+    "kernel_size": 3,
+    "dropout_p": 0.3,
+    "output_dim": 2,
+    "beta": 1,
+    "load_checkpoint": false,
+    "checkpoint_num": 1,
+    "activation": "sine",
+    "sine_w0": 1.0,
+    "avg_output": true,
+    "checkpoint_path": "/data/lightSpec/logs/light_2024-11-27/CNNEncoder_lc_1.pth"
+  },
+  "optim_args": {
+    "max_lr": "5e-5",
+    "weight_decay": "1e-5",
+    "warmup_pct": 0.15,
+    "steps_per_epoch": 2500,
+    "quantiles": [
+      0.1,
+      0.25,
+      0.5,
+      0.75,
+      0.9
+    ]
+  },
+  "num_params": 11982725,
+  "model_structure": "DistributedDataParallel(\n  (module): MultiTaskSimSiam(\n    (activation): GELU(approximate='none')\n    (simsiam): SimSiam(\n      (backbone): DoubleInputRegressor(\n        (encoder1): CNNEncoder(\n          (activation): Sine()\n          (embedding): Sequential(\n            (0): Conv1d(2, 32, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n            (1): BatchNorm1d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n            (2): Sine()\n          )\n          (layers): ModuleList(\n            (0): ConvBlock(\n              (activation): Sine()\n              (layers): Sequential(\n                (0): Conv1d(32, 64, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                (1): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                (2): Sine()\n              )\n            )\n            (1): ConvBlock(\n              (activation): Sine()\n              (layers): Sequential(\n                (0): Conv1d(64, 128, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                (1): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                (2): Sine()\n              )\n            )\n            (2): ConvBlock(\n              (activation): Sine()\n              (layers): Sequential(\n                (0): Conv1d(128, 256, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                (2): Sine()\n              )\n            )\n            (3): ConvBlock(\n              (activation): Sine()\n              (layers): Sequential(\n                (0): Conv1d(256, 512, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                (1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                (2): Sine()\n              )\n            )\n            (4-5): 2 x ConvBlock(\n              (activation): Sine()\n              (layers): Sequential(\n                (0): Conv1d(512, 512, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                (1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                (2): Sine()\n              )\n            )\n          )\n          (pool): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n        )\n        (encoder2): Astroconformer(\n          (extractor): Sequential(\n            (0): Conv1d(2, 512, kernel_size=(20,), stride=(20,))\n            (1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n            (2): SiLU()\n          )\n          (pe): RotaryEmbedding()\n          (encoder): ConformerEncoder(\n            (blocks): ModuleList(\n              (0-4): 5 x ConformerBlock(\n                (modlist): ModuleList(\n                  (0): PostNorm(\n                    (module): MHA_rotary(\n                      (query): Linear(in_features=512, out_features=512, bias=True)\n                      (key): Linear(in_features=512, out_features=512, bias=True)\n                      (value): Linear(in_features=512, out_features=512, bias=True)\n                      (rotary_emb): RotaryEmbedding()\n                      (output): Linear(in_features=512, out_features=512, bias=True)\n                    )\n                    (norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n                  )\n                  (1): PostNorm(\n                    (module): ConvBlock(\n                      (layers): Sequential(\n                        (0): Conv1d(512, 512, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                        (1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                        (2): SiLU()\n                      )\n                    )\n                    (norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n                  )\n                )\n              )\n            )\n          )\n          (pred_layer): Identity()\n        )\n        (regressor): Identity()\n      )\n      (projector): projection_MLP(\n        (layer1): Sequential(\n          (0): Linear(in_features=1024, out_features=256, bias=True)\n          (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n          (2): ReLU(inplace=True)\n        )\n        (layer2): Sequential(\n          (0): Linear(in_features=256, out_features=256, bias=True)\n          (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n          (2): ReLU(inplace=True)\n        )\n        (layer3): Sequential(\n          (0): Linear(in_features=256, out_features=256, bias=True)\n          (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n        )\n      )\n      (encoder): Sequential(\n        (0): DoubleInputRegressor(\n          (encoder1): CNNEncoder(\n            (activation): Sine()\n            (embedding): Sequential(\n              (0): Conv1d(2, 32, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n              (1): BatchNorm1d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n              (2): Sine()\n            )\n            (layers): ModuleList(\n              (0): ConvBlock(\n                (activation): Sine()\n                (layers): Sequential(\n                  (0): Conv1d(32, 64, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                  (1): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                  (2): Sine()\n                )\n              )\n              (1): ConvBlock(\n                (activation): Sine()\n                (layers): Sequential(\n                  (0): Conv1d(64, 128, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                  (1): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                  (2): Sine()\n                )\n              )\n              (2): ConvBlock(\n                (activation): Sine()\n                (layers): Sequential(\n                  (0): Conv1d(128, 256, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                  (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                  (2): Sine()\n                )\n              )\n              (3): ConvBlock(\n                (activation): Sine()\n                (layers): Sequential(\n                  (0): Conv1d(256, 512, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                  (1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                  (2): Sine()\n                )\n              )\n              (4-5): 2 x ConvBlock(\n                (activation): Sine()\n                (layers): Sequential(\n                  (0): Conv1d(512, 512, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                  (1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                  (2): Sine()\n                )\n              )\n            )\n            (pool): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n          )\n          (encoder2): Astroconformer(\n            (extractor): Sequential(\n              (0): Conv1d(2, 512, kernel_size=(20,), stride=(20,))\n              (1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n              (2): SiLU()\n            )\n            (pe): RotaryEmbedding()\n            (encoder): ConformerEncoder(\n              (blocks): ModuleList(\n                (0-4): 5 x ConformerBlock(\n                  (modlist): ModuleList(\n                    (0): PostNorm(\n                      (module): MHA_rotary(\n                        (query): Linear(in_features=512, out_features=512, bias=True)\n                        (key): Linear(in_features=512, out_features=512, bias=True)\n                        (value): Linear(in_features=512, out_features=512, bias=True)\n                        (rotary_emb): RotaryEmbedding()\n                        (output): Linear(in_features=512, out_features=512, bias=True)\n                      )\n                      (norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n                    )\n                    (1): PostNorm(\n                      (module): ConvBlock(\n                        (layers): Sequential(\n                          (0): Conv1d(512, 512, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                          (1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                          (2): SiLU()\n                        )\n                      )\n                      (norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n                    )\n                  )\n                )\n              )\n            )\n            (pred_layer): Identity()\n          )\n          (regressor): Identity()\n        )\n        (1): projection_MLP(\n          (layer1): Sequential(\n            (0): Linear(in_features=1024, out_features=256, bias=True)\n            (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n            (2): ReLU(inplace=True)\n          )\n          (layer2): Sequential(\n            (0): Linear(in_features=256, out_features=256, bias=True)\n            (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n            (2): ReLU(inplace=True)\n          )\n          (layer3): Sequential(\n            (0): Linear(in_features=256, out_features=256, bias=True)\n            (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n          )\n        )\n      )\n      (predictor): prediction_MLP(\n        (layer1): Sequential(\n          (0): Linear(in_features=256, out_features=256, bias=True)\n          (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n          (2): ReLU(inplace=True)\n        )\n        (layer2): Linear(in_features=256, out_features=256, bias=True)\n      )\n    )\n    (regressor): Sequential(\n      (0): Linear(in_features=512, out_features=256, bias=True)\n      (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n      (2): GELU(approximate='none')\n      (3): Dropout(p=0.3, inplace=False)\n      (4): Linear(in_features=256, out_features=5, bias=True)\n    )\n  )\n)",
+  "transforms": "Compose(\n    RandomCrop(crop_size=34560)\n    moving_avg(kernel_size=13, stride=1)\n    ACF(max_lag=None)\n    <transforms.transforms.FFT object at 0x7f8fd43145b0>\n    Normalize(scheme='['mag_median', 'std']', axis=0)\n    ToTensor\n)",
+  "trainer": {
+    "model": "DistributedDataParallel(\n  (module): MultiTaskSimSiam(\n    (activation): GELU(approximate='none')\n    (simsiam): SimSiam(\n      (backbone): DoubleInputRegressor(\n        (encoder1): CNNEncoder(\n          (activation): Sine()\n          (embedding): Sequential(\n            (0): Conv1d(2, 32, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n            (1): BatchNorm1d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n            (2): Sine()\n          )\n          (layers): ModuleList(\n            (0): ConvBlock(\n              (activation): Sine()\n              (layers): Sequential(\n                (0): Conv1d(32, 64, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                (1): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                (2): Sine()\n              )\n            )\n            (1): ConvBlock(\n              (activation): Sine()\n              (layers): Sequential(\n                (0): Conv1d(64, 128, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                (1): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                (2): Sine()\n              )\n            )\n            (2): ConvBlock(\n              (activation): Sine()\n              (layers): Sequential(\n                (0): Conv1d(128, 256, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                (2): Sine()\n              )\n            )\n            (3): ConvBlock(\n              (activation): Sine()\n              (layers): Sequential(\n                (0): Conv1d(256, 512, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                (1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                (2): Sine()\n              )\n            )\n            (4-5): 2 x ConvBlock(\n              (activation): Sine()\n              (layers): Sequential(\n                (0): Conv1d(512, 512, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                (1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                (2): Sine()\n              )\n            )\n          )\n          (pool): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n        )\n        (encoder2): Astroconformer(\n          (extractor): Sequential(\n            (0): Conv1d(2, 512, kernel_size=(20,), stride=(20,))\n            (1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n            (2): SiLU()\n          )\n          (pe): RotaryEmbedding()\n          (encoder): ConformerEncoder(\n            (blocks): ModuleList(\n              (0-4): 5 x ConformerBlock(\n                (modlist): ModuleList(\n                  (0): PostNorm(\n                    (module): MHA_rotary(\n                      (query): Linear(in_features=512, out_features=512, bias=True)\n                      (key): Linear(in_features=512, out_features=512, bias=True)\n                      (value): Linear(in_features=512, out_features=512, bias=True)\n                      (rotary_emb): RotaryEmbedding()\n                      (output): Linear(in_features=512, out_features=512, bias=True)\n                    )\n                    (norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n                  )\n                  (1): PostNorm(\n                    (module): ConvBlock(\n                      (layers): Sequential(\n                        (0): Conv1d(512, 512, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                        (1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                        (2): SiLU()\n                      )\n                    )\n                    (norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n                  )\n                )\n              )\n            )\n          )\n          (pred_layer): Identity()\n        )\n        (regressor): Identity()\n      )\n      (projector): projection_MLP(\n        (layer1): Sequential(\n          (0): Linear(in_features=1024, out_features=256, bias=True)\n          (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n          (2): ReLU(inplace=True)\n        )\n        (layer2): Sequential(\n          (0): Linear(in_features=256, out_features=256, bias=True)\n          (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n          (2): ReLU(inplace=True)\n        )\n        (layer3): Sequential(\n          (0): Linear(in_features=256, out_features=256, bias=True)\n          (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n        )\n      )\n      (encoder): Sequential(\n        (0): DoubleInputRegressor(\n          (encoder1): CNNEncoder(\n            (activation): Sine()\n            (embedding): Sequential(\n              (0): Conv1d(2, 32, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n              (1): BatchNorm1d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n              (2): Sine()\n            )\n            (layers): ModuleList(\n              (0): ConvBlock(\n                (activation): Sine()\n                (layers): Sequential(\n                  (0): Conv1d(32, 64, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                  (1): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                  (2): Sine()\n                )\n              )\n              (1): ConvBlock(\n                (activation): Sine()\n                (layers): Sequential(\n                  (0): Conv1d(64, 128, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                  (1): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                  (2): Sine()\n                )\n              )\n              (2): ConvBlock(\n                (activation): Sine()\n                (layers): Sequential(\n                  (0): Conv1d(128, 256, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                  (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                  (2): Sine()\n                )\n              )\n              (3): ConvBlock(\n                (activation): Sine()\n                (layers): Sequential(\n                  (0): Conv1d(256, 512, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                  (1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                  (2): Sine()\n                )\n              )\n              (4-5): 2 x ConvBlock(\n                (activation): Sine()\n                (layers): Sequential(\n                  (0): Conv1d(512, 512, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                  (1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                  (2): Sine()\n                )\n              )\n            )\n            (pool): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n          )\n          (encoder2): Astroconformer(\n            (extractor): Sequential(\n              (0): Conv1d(2, 512, kernel_size=(20,), stride=(20,))\n              (1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n              (2): SiLU()\n            )\n            (pe): RotaryEmbedding()\n            (encoder): ConformerEncoder(\n              (blocks): ModuleList(\n                (0-4): 5 x ConformerBlock(\n                  (modlist): ModuleList(\n                    (0): PostNorm(\n                      (module): MHA_rotary(\n                        (query): Linear(in_features=512, out_features=512, bias=True)\n                        (key): Linear(in_features=512, out_features=512, bias=True)\n                        (value): Linear(in_features=512, out_features=512, bias=True)\n                        (rotary_emb): RotaryEmbedding()\n                        (output): Linear(in_features=512, out_features=512, bias=True)\n                      )\n                      (norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n                    )\n                    (1): PostNorm(\n                      (module): ConvBlock(\n                        (layers): Sequential(\n                          (0): Conv1d(512, 512, kernel_size=(3,), stride=(1,), padding=same, bias=False)\n                          (1): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n                          (2): SiLU()\n                        )\n                      )\n                      (norm): LayerNorm((512,), eps=1e-05, elementwise_affine=True)\n                    )\n                  )\n                )\n              )\n            )\n            (pred_layer): Identity()\n          )\n          (regressor): Identity()\n        )\n        (1): projection_MLP(\n          (layer1): Sequential(\n            (0): Linear(in_features=1024, out_features=256, bias=True)\n            (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n            (2): ReLU(inplace=True)\n          )\n          (layer2): Sequential(\n            (0): Linear(in_features=256, out_features=256, bias=True)\n            (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n            (2): ReLU(inplace=True)\n          )\n          (layer3): Sequential(\n            (0): Linear(in_features=256, out_features=256, bias=True)\n            (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n          )\n        )\n      )\n      (predictor): prediction_MLP(\n        (layer1): Sequential(\n          (0): Linear(in_features=256, out_features=256, bias=True)\n          (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n          (2): ReLU(inplace=True)\n        )\n        (layer2): Linear(in_features=256, out_features=256, bias=True)\n      )\n    )\n    (regressor): Sequential(\n      (0): Linear(in_features=512, out_features=256, bias=True)\n      (1): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n      (2): GELU(approximate='none')\n      (3): Dropout(p=0.3, inplace=False)\n      (4): Linear(in_features=256, out_features=5, bias=True)\n    )\n  )\n)",
+    "optimizer": "AdamW (\nParameter Group 0\n    amsgrad: False\n    betas: (0.9, 0.999)\n    capturable: False\n    differentiable: False\n    eps: 1e-08\n    foreach: None\n    fused: None\n    initial_lr: 5e-05\n    lr: 5e-05\n    maximize: False\n    weight_decay: 1e-05\n)",
+    "criterion": "CQR()",
+    "scaler": "<torch.cuda.amp.grad_scaler.GradScaler object at 0x7f8f6924f790>",
+    "grad_clip": false,
+    "cos_inc": false,
+    "output_dim": 1,
+    "scheduler": "<torch.optim.lr_scheduler.CosineAnnealingLR object at 0x7f8f6924f850>",
+    "train_dl": "<torch.utils.data.dataloader.DataLoader object at 0x7f8f6f27cb20>",
+    "val_dl": "<torch.utils.data.dataloader.DataLoader object at 0x7f8f6f220ac0>",
+    "train_sampler": "<torch.utils.data.distributed.DistributedSampler object at 0x7f8f6f27c910>",
+    "val_sampler": "<torch.utils.data.distributed.DistributedSampler object at 0x7f8f6f27cdc0>",
+    "max_iter": Infinity,
+    "device": 0,
+    "world_size": 1,
+    "exp_num": "light_2025-03-19",
+    "exp_name": "DoubleInputRegressor_lc_3_ssl",
+    "log_path": "/data/lightSpec/logs",
+    "best_state_dict": null,
+    "plot_every": null,
+    "logger": null,
+    "range_update": null,
+    "accumulation_step": 1,
+    "wandb": false,
+    "num_quantiles": 5,
+    "update_func": "<function Trainer.<lambda> at 0x7f8fd7d94820>",
+    "stack_pairs": false,
+    "temperature": 1,
+    "use_w": false,
+    "use_pred_coeff": false,
+    "pred_coeff_val": null,
+    "ssl_weight": 0.5,
+    "weight_decay": true
+  }
+}