results/episode_task_suite/neural_mlp/cross_modal_retrieval/metrics.json

{
  "mrr": 0.1299971898648288,
  "median_rank": 40.0,
  "mean_rank": 66.60057471264368,
  "num_queries": 348,
  "top1_accuracy": 0.05172413793103448,
  "top5_accuracy": 0.19827586206896552,
  "top10_accuracy": 0.2413793103448276,
  "task": "cross_modal_retrieval",
  "input": "motion/IMU/camera/audio query",
  "split": "chronological",
  "num_train_windows": 813,
  "num_test_windows": 348,
  "target_dim": 5096,
  "output": "matching depth/video window",
  "model": "neural_mlp",
  "head": "z-score -> MLP projection/regression",
  "neural_epochs": 80,
  "neural_hidden_dim": 128,
  "neural_batch_size": 128,
  "neural_learning_rate": 0.001,
  "neural_weight_decay": 0.0001,
  "neural_dropout": 0.1,
  "neural_device": "cpu",
  "train_final_loss": 0.21891545446596464,
  "task_display_name": "Cross-Modal Retrieval"
}