Add files using upload-large-folder tool

results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/caption_grounding/metrics.json

@@ -0,0 +1,47 @@
+{
+  "status": "pass",
+  "task": "caption_grounding",
+  "task_display_name": "Language Grounding",
+  "model_family": "neural_mlp_metadata_contrastive_retrieval",
+  "source": "128_episode_qwen_jsonl_metadata",
+  "limitation": "query text is derived from held-out labels, while candidate representations exclude answer_json fields; this is a public-safe retrieval proxy, not a raw caption-to-sensor grounding model",
+  "split_policy": "train contrastive query/window alignment on train split, report held-out test retrieval",
+  "num_train_windows": 25629,
+  "num_test_windows": 4032,
+  "history": [
+    {
+      "epoch": 1,
+      "loss": 3.8980400463398808
+    },
+    {
+      "epoch": 7,
+      "loss": 0.9906599925536399
+    },
+    {
+      "epoch": 14,
+      "loss": 0.7154027189824095
+    },
+    {
+      "epoch": 21,
+      "loss": 0.6159498697075041
+    },
+    {
+      "epoch": 28,
+      "loss": 0.5653652491909131
+    },
+    {
+      "epoch": 35,
+      "loss": 0.5247695826790848
+    }
+  ],
+  "device": "cuda",
+  "num_queries": 4032,
+  "mrr": 0.008236799389123917,
+  "median_rank": 1038.0,
+  "mean_rank": 1375.5374755859375,
+  "top1_accuracy": 0.001488095238095238,
+  "top5_accuracy": 0.008432539682539682,
+  "top10_accuracy": 0.015625,
+  "primary_metric": "mrr",
+  "primary_score": 0.008236799389123917
+}

results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/temporal_order/confusion_matrix.csv

@@ -0,0 +1,3 @@
+true\pred,reversed,correct
+reversed,1771,289
+correct,411,1547

results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/temporal_order/metrics.json

@@ -0,0 +1,52 @@
+{
+  "status": "pass",
+  "task": "temporal_order",
+  "task_display_name": "Temporal Order Verification",
+  "model_family": "neural_mlp_metadata_pair",
+  "source": "128_episode_qwen_jsonl_metadata",
+  "limitation": "metadata-only pair features are weaker than the single-episode raw-feature temporal-order task",
+  "split_policy": "train on train-split adjacent window pairs, report held-out test-split adjacent window pairs",
+  "num_train_samples": 25540,
+  "num_test_samples": 4018,
+  "history": [
+    {
+      "epoch": 1,
+      "loss": 0.5672330612770355,
+      "train_accuracy": 0.6666014095536413
+    },
+    {
+      "epoch": 7,
+      "loss": 0.0043020602993918,
+      "train_accuracy": 0.9985512920908379
+    },
+    {
+      "epoch": 14,
+      "loss": 0.004064371907577967,
+      "train_accuracy": 0.9989036805011746
+    },
+    {
+      "epoch": 21,
+      "loss": 0.002938282458758355,
+      "train_accuracy": 0.9990602975724354
+    },
+    {
+      "epoch": 28,
+      "loss": 0.0009582243631951243,
+      "train_accuracy": 0.9996476115896633
+    },
+    {
+      "epoch": 35,
+      "loss": 0.0010173755352853556,
+      "train_accuracy": 0.9997259201252937
+    }
+  ],
+  "device": "cuda",
+  "accuracy": 0.8257839721254355,
+  "balanced_accuracy": 0.8249003342027232,
+  "macro_f1": 0.8252408266656923,
+  "weighted_f1": 0.8254881518303968,
+  "num_eval_windows": 4018,
+  "num_classes": 2,
+  "primary_metric": "f1",
+  "primary_score": 0.8252408266656923
+}

results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/temporal_order/per_class_metrics.csv

@@ -0,0 +1,3 @@
+class_id,class_name,support,predicted,precision,recall,f1
+0,reversed,2060,2182,0.8116406966086159,0.8597087378640776,0.834983498349835
+1,correct,1958,1836,0.8425925925925926,0.7900919305413687,0.8154981549815498