docs: upload run_config for archive_4experts moe_standard
Browse files
archive_4experts/moe_standard/none/scratch/seed_42/results/run_config.json
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"allow_pretrained": true,
|
| 3 |
+
"architecture": "resnet18",
|
| 4 |
+
"augment": true,
|
| 5 |
+
"batch_size": 16,
|
| 6 |
+
"cutmix_alpha": 0.0,
|
| 7 |
+
"data_limit": null,
|
| 8 |
+
"dropout": 0.2,
|
| 9 |
+
"early_stop_patience": 8,
|
| 10 |
+
"epochs": 20,
|
| 11 |
+
"expert_family": "mobilenet",
|
| 12 |
+
"fourier_mode": "none",
|
| 13 |
+
"hidden_size": 256,
|
| 14 |
+
"image_size": 224,
|
| 15 |
+
"in_channels": 3,
|
| 16 |
+
"label_smoothing": 0.0,
|
| 17 |
+
"lr_backbone": 0.0001,
|
| 18 |
+
"lr_head": 0.001,
|
| 19 |
+
"max_grad_norm": 1.0,
|
| 20 |
+
"mixup_alpha": 0.0,
|
| 21 |
+
"model_family": "moe_standard",
|
| 22 |
+
"model_size": "base",
|
| 23 |
+
"multi_gpu": true,
|
| 24 |
+
"num_attention_heads": 8,
|
| 25 |
+
"num_experts": 4,
|
| 26 |
+
"num_hidden_layers": 6,
|
| 27 |
+
"num_workers": 4,
|
| 28 |
+
"patch_size": 16,
|
| 29 |
+
"projection_dim": 128,
|
| 30 |
+
"raw_min": false,
|
| 31 |
+
"regime": "scratch",
|
| 32 |
+
"routing_strategy": "dense",
|
| 33 |
+
"scheduler_patience": 3,
|
| 34 |
+
"seed": 42,
|
| 35 |
+
"seeds": [
|
| 36 |
+
42,
|
| 37 |
+
123,
|
| 38 |
+
2024
|
| 39 |
+
],
|
| 40 |
+
"threshold_strategy": "balanced_accuracy",
|
| 41 |
+
"top_k": 2,
|
| 42 |
+
"train_backbone": true,
|
| 43 |
+
"unfreeze_last_n": 2,
|
| 44 |
+
"use_class_weights": false,
|
| 45 |
+
"use_weighted_sampler": true,
|
| 46 |
+
"variant": "small",
|
| 47 |
+
"weight_decay": 0.0001
|
| 48 |
+
}
|