Add Assignment 4 summary

summary.json

@@ -0,0 +1,116 @@
+{
+  "baseline": {
+    "epochs": 100,
+    "training_time_seconds": 3139.0,
+    "training_time_minutes": 52.32,
+    "final_loss": 0.0974,
+    "bleu": 0.5247,
+    "bleu_percent": 52.47,
+    "source": "Extracted from en_to_hi.ipynb recorded outputs on 2026-03-13"
+  },
+  "tuning": {
+    "best_config": {
+      "learning_rate": 8.884411309500873e-05,
+      "batch_size": 16,
+      "num_heads": 4,
+      "d_ff": 1536,
+      "dropout": 0.19016462895143443,
+      "num_layers": 4,
+      "weight_decay": 1.149540832915101e-06,
+      "d_model": 512,
+      "max_len": 50,
+      "seed": 42,
+      "force_cpu": false,
+      "epochs": 20
+    },
+    "best_metrics": {
+      "epoch": 20,
+      "loss": 0.5256124382669275,
+      "bleu": 0.7698028731485314,
+      "bleu_percent": 76.98028731485313,
+      "timestamp": 1773401641,
+      "checkpoint_dir_name": null,
+      "done": true,
+      "training_iteration": 20,
+      "trial_id": "28792536",
+      "date": "2026-03-13_17-04-01",
+      "time_this_iter_s": 42.827330112457275,
+      "time_total_s": 857.8043477535248,
+      "pid": 3948130,
+      "hostname": "user-SSI100C3A-48",
+      "node_ip": "10.6.0.87",
+      "config": {
+        "learning_rate": 8.884411309500873e-05,
+        "batch_size": 16,
+        "num_heads": 4,
+        "d_ff": 1536,
+        "dropout": 0.19016462895143443,
+        "num_layers": 4,
+        "weight_decay": 1.149540832915101e-06,
+        "d_model": 512,
+        "max_len": 50,
+        "seed": 42,
+        "force_cpu": false,
+        "epochs": 20
+      },
+      "time_since_restore": 857.8043477535248,
+      "iterations_since_restore": 20,
+      "experiment_tag": "18_batch_size=16,d_ff=1536,d_model=512,dropout=0.1902,epochs=20,force_cpu=False,learning_rate=0.0001,max_len=50,num_heads=4,num_layers=4,seed=42,weight_decay=0.0000"
+    },
+    "num_samples": 20,
+    "tune_epochs": 20,
+    "search_space": {
+      "learning_rate": "loguniform(1e-5, 1e-3)",
+      "batch_size": "choice([16, 32, 64])",
+      "num_heads": "choice([4, 8])",
+      "d_ff": "choice([1024, 1536, 2048])",
+      "dropout": "uniform(0.10, 0.40)",
+      "num_layers": "choice([4, 6])",
+      "weight_decay": "loguniform(1e-6, 1e-3)"
+    },
+    "metric": "bleu",
+    "mode": "max",
+    "scheduler": "ASHAScheduler",
+    "search_algorithm": "OptunaSearch"
+  },
+  "best_model": {
+    "epochs": 20,
+    "training_time_seconds": 853.0576362609863,
+    "training_time_minutes": 14.217627271016438,
+    "final_loss": 0.5256124382669275,
+    "best_bleu": 0.8471540338276414,
+    "best_bleu_percent": 84.71540338276414,
+    "best_epoch": 15,
+    "epochs_to_target": 6,
+    "model_path": "b23cs1075_ass_4_best_model.pth",
+    "config": {
+      "learning_rate": 8.884411309500873e-05,
+      "batch_size": 16,
+      "num_heads": 4,
+      "d_ff": 1536,
+      "dropout": 0.19016462895143443,
+      "num_layers": 4,
+      "weight_decay": 1.149540832915101e-06,
+      "d_model": 512,
+      "max_len": 50,
+      "seed": 42,
+      "force_cpu": false,
+      "epochs": 20
+    }
+  },
+  "search_space": {
+    "learning_rate": "loguniform(1e-5, 1e-3)",
+    "batch_size": "choice([16, 32, 64])",
+    "num_heads": "choice([4, 8])",
+    "d_ff": "choice([1024, 1536, 2048])",
+    "dropout": "uniform(0.10, 0.40)",
+    "num_layers": "choice([4, 6])",
+    "weight_decay": "loguniform(1e-6, 1e-3)"
+  },
+  "artifacts": {
+    "baseline_model": "transformer_translation_final.pth",
+    "best_model": "b23cs1075_ass_4_best_model.pth",
+    "report_markdown": "b23cs1075_ass_4_report.md",
+    "summary_json": "artifacts/assignment4/summary.json"
+  }
+}