| { | |
| "model_name": "swin_small_patch4_window7_224", | |
| "num_classes": 7, | |
| "pretrained": true, | |
| "drop_path_rate": 0.2, | |
| "data_root": "/kaggle/input/tea-leaf701515/tea_leaf_processed_dataset", | |
| "input_size": 224, | |
| "batch_size": 64, | |
| "num_workers": 0, | |
| "img_mean": [ | |
| 0.485, | |
| 0.456, | |
| 0.406 | |
| ], | |
| "img_std": [ | |
| 0.229, | |
| 0.224, | |
| 0.225 | |
| ], | |
| "hsv_use_sincos": false, | |
| "gate_warmup_epochs": 5, | |
| "epochs": 100, | |
| "warmup_epochs": 5, | |
| "lr": 0.0005, | |
| "warmup_lr_init": 1e-06, | |
| "weight_decay": 0.05, | |
| "min_lr": 1e-06, | |
| "label_smoothing": 0.1, | |
| "grad_clip_norm": 1.0, | |
| "gradient_accumulation_steps": 1, | |
| "color_jitter": 0.2, | |
| "hue_jitter": 0.1, | |
| "use_gaussian_blur": false, | |
| "gaussian_blur_prob": 0.1, | |
| "rrc_scale_min": 0.85, | |
| "use_amp": true, | |
| "use_ema": true, | |
| "ema_decay": 0.9998, | |
| "use_class_weights": false, | |
| "use_hsv_branch": true, | |
| "hsv_embed_dim": 128, | |
| "hsv_dropout": 0.1, | |
| "fuse_dropout": 0.2, | |
| "gate_hidden": 256, | |
| "gate_vector": false, | |
| "compute_val_auc": false, | |
| "early_stopping_patience": 25, | |
| "run_dir": "/kaggle/working/experiments_swin", | |
| "experiment_name": "swin_small_hsv_raw", | |
| "log_interval": 50, | |
| "save_cm_png": true, | |
| "ckpt_temp_dir": "/kaggle/working/ckpt_cache", | |
| "save_epoch_checkpoints": false, | |
| "save_epoch_every": 5, | |
| "keep_last_n_checkpoints": 3, | |
| "device": "cuda", | |
| "seed": 42, | |
| "deterministic": true, | |
| "use_data_parallel": true | |
| } |