Resume SynthData0523 main/c2 batch 25

.gitattributes

@@ -3909,3 +3909,21 @@ SynthData0523/main/c2/tvae/tvae-c2-20260504_174121/gen_20260504_174137.log filte
 SynthData0523/main/c2/tvae/tvae-c2-20260504_174121/models_300epochs/train_20260504_174121.log filter=lfs diff=lfs merge=lfs -text
 SynthData0523/main/c2/tvae/tvae-c2-20260504_174121/models_300epochs/tvae_300epochs.pt filter=lfs diff=lfs merge=lfs -text
 SynthData0523/main/c2/tvae/tvae-c2-20260504_174121/staged/public/test.csv filter=lfs diff=lfs merge=lfs -text
+SynthData0523/main/c2/tvae/tvae-c2-20260504_174121/staged/public/train.csv filter=lfs diff=lfs merge=lfs -text
+SynthData0523/main/c2/tvae/tvae-c2-20260504_174121/staged/public/val.csv filter=lfs diff=lfs merge=lfs -text
+SynthData0523/main/c2/tvae/tvae-c2-20260504_174121/tvae-c2-1382-20260504_174137.csv filter=lfs diff=lfs merge=lfs -text
+SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/gen_20260504_174210.log filter=lfs diff=lfs merge=lfs -text
+SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/models_300epochs/train_20260504_174154.log filter=lfs diff=lfs merge=lfs -text
+SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/models_300epochs/tvae_300epochs.pt filter=lfs diff=lfs merge=lfs -text
+SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/staged/public/test.csv filter=lfs diff=lfs merge=lfs -text
+SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/staged/public/train.csv filter=lfs diff=lfs merge=lfs -text
+SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/staged/public/val.csv filter=lfs diff=lfs merge=lfs -text
+SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/tvae-c2-1382-20260504_174210.csv filter=lfs diff=lfs merge=lfs -text
+SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/gen_20260504_174245.log filter=lfs diff=lfs merge=lfs -text
+SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/models_300epochs/train_20260504_174228.log filter=lfs diff=lfs merge=lfs -text
+SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/models_300epochs/tvae_300epochs.pt filter=lfs diff=lfs merge=lfs -text
+SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/staged/public/test.csv filter=lfs diff=lfs merge=lfs -text
+SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/staged/public/train.csv filter=lfs diff=lfs merge=lfs -text
+SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/staged/public/val.csv filter=lfs diff=lfs merge=lfs -text
+SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/tvae-c2-1382-20260504_174245.csv filter=lfs diff=lfs merge=lfs -text
+SynthData0523/main/c2/tvae/tvae-c2-20260504_175027/gen_20260504_175041.log filter=lfs diff=lfs merge=lfs -text

SynthData0523/main/c2/tvae/tvae-c2-20260504_174121/staged/public/train.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4aed00c2c2b3f88a55a7ebff31b2e1b5e0e32fb0a7267e0b9d2779cd23e434dd
+size 41565

SynthData0523/main/c2/tvae/tvae-c2-20260504_174121/staged/public/val.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:26e90c1170a57a14c05832ac88027722b1f3848f9662c7c09ef7c93dcba4cc01
+size 5176

SynthData0523/main/c2/tvae/tvae-c2-20260504_174121/staged/tvae/adapter_report.json

@@ -0,0 +1,7 @@
+{
+  "adapter_ready_status": "pass",
+  "adapter_fail_reason_code": null,
+  "adapter_fail_detail": null,
+  "adapter_transforms_applied": [],
+  "model_input_manifest": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174121/staged/tvae/model_input_manifest.json"
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174121/staged/tvae/adapter_transforms_applied.json

@@ -0,0 +1 @@
+[]

SynthData0523/main/c2/tvae/tvae-c2-20260504_174121/staged/tvae/model_input_manifest.json

@@ -0,0 +1,151 @@
+{
+  "dataset_id": "c2",
+  "model": "tvae",
+  "target_column": "class",
+  "task_type": "classification",
+  "column_schema": [
+    {
+      "name": "buying",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "vhigh",
+          "med",
+          "high",
+          "low"
+        ]
+      }
+    },
+    {
+      "name": "maint",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "vhigh",
+          "low",
+          "med",
+          "high"
+        ]
+      }
+    },
+    {
+      "name": "doors",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "2",
+          "5more",
+          "3",
+          "4"
+        ]
+      }
+    },
+    {
+      "name": "persons",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "2",
+          "4",
+          "more"
+        ]
+      }
+    },
+    {
+      "name": "lug_boot",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "small",
+          "big",
+          "med"
+        ]
+      }
+    },
+    {
+      "name": "safety",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "low",
+          "high",
+          "med"
+        ]
+      }
+    },
+    {
+      "name": "class",
+      "role": "target",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "unacc",
+          "good",
+          "acc",
+          "vgood"
+        ]
+      }
+    }
+  ],
+  "public_manifest": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174121/public_gate/staged_input_manifest.json",
+  "train_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174121/staged/public/train.csv",
+  "val_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174121/staged/public/val.csv",
+  "test_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174121/staged/public/test.csv",
+  "features_json": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174121/staged/public/staged_features.json",
+  "public_gate_report": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174121/public_gate/public_gate_report.json"
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174121/tvae-c2-1382-20260504_174137.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b2dc7b4aac308837c0d16c501f82943b483cde999544304161ff98921ed6e4c
+size 41505

SynthData0523/main/c2/tvae/tvae-c2-20260504_174121/tvae_metadata.json

@@ -0,0 +1,32 @@
+{
+  "columns": [
+    {
+      "name": "buying",
+      "type": "categorical"
+    },
+    {
+      "name": "maint",
+      "type": "categorical"
+    },
+    {
+      "name": "doors",
+      "type": "categorical"
+    },
+    {
+      "name": "persons",
+      "type": "categorical"
+    },
+    {
+      "name": "lug_boot",
+      "type": "categorical"
+    },
+    {
+      "name": "safety",
+      "type": "categorical"
+    },
+    {
+      "name": "class",
+      "type": "categorical"
+    }
+  ]
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/_tvae_generate.py

@@ -0,0 +1,23 @@
+import os, sys
+sys.path.insert(0, "/work")
+from src.SpecificModels.ctgan_joblib_parallel_cap import apply_parallel_cap_from_env
+apply_parallel_cap_from_env()
+from src.SpecificModels.ctgan_rdt_inverse_fix import apply_ctgan_inverse_fix
+apply_ctgan_inverse_fix()
+import pandas as pd
+from ctgan.synthesizers.tvae import TVAE
+os.environ.setdefault("LOKY_MAX_CPU_COUNT", "8")
+os.environ.setdefault("OPENBLAS_NUM_THREADS", "8")
+os.environ.setdefault("MKL_NUM_THREADS", "8")
+model = TVAE.load("/work/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/models_300epochs/tvae_300epochs.pt")
+total = 1382
+chunk = min(50000, total) if total > 50000 else total
+parts = []
+left = total
+while left > 0:
+    take = min(chunk, left)
+    parts.append(model.sample(take))
+    left -= take
+samples = pd.concat(parts, ignore_index=True) if len(parts) > 1 else parts[0]
+samples.to_csv("/work/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/tvae-c2-1382-20260504_174210.csv", index=False)
+print(f"[TVAE] Generated {total} rows (chunks={len(parts)}) -> /work/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/tvae-c2-1382-20260504_174210.csv")

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/_tvae_train.py

@@ -0,0 +1,39 @@
+import json, os, sys
+sys.path.insert(0, "/work")
+from src.SpecificModels.ctgan_joblib_parallel_cap import apply_parallel_cap_from_env
+apply_parallel_cap_from_env()
+import pandas as pd
+from ctgan.data import read_csv
+from ctgan.synthesizers.tvae import TVAE
+
+# Keep transform stage parallelism bounded for stability on shared host.
+os.environ.setdefault("LOKY_MAX_CPU_COUNT", "8")
+os.environ.setdefault("OPENBLAS_NUM_THREADS", "8")
+os.environ.setdefault("MKL_NUM_THREADS", "8")
+_nj = (os.environ.get("TVAE_CTGAN_JOBTRANS_N_JOBS") or "").strip()
+if _nj:
+    print("[TVAE] joblib Parallel cap ON, TVAE_CTGAN_JOBTRANS_N_JOBS=" + _nj)
+else:
+    print("[TVAE] joblib Parallel cap OFF (unset TVAE_CTGAN_JOBTRANS_N_JOBS)")
+print("[TVAE] LOKY_MAX_CPU_COUNT=" + str(os.environ.get("LOKY_MAX_CPU_COUNT", "")))
+
+csv_path = "/work/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/staged/public/train.csv"
+meta_path = "/work/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/tvae_metadata.json"
+save_path = "/work/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/models_300epochs/tvae_300epochs.pt"
+epochs = int(os.environ.get("TVAE_EPOCHS", 300))
+batch_size = int(os.environ.get("TVAE_BATCH_SIZE", "500"))
+embedding_dim = int(os.environ.get("TVAE_EMBEDDING_DIM", "128"))
+def _parse_dims(name, default):
+    raw = (os.environ.get(name) or "").strip()
+    if not raw:
+        return default
+    return tuple(int(x.strip()) for x in raw.split(",") if x.strip())
+compress_dims = _parse_dims("TVAE_COMPRESS_DIMS", (128, 128))
+decompress_dims = _parse_dims("TVAE_DECOMPRESS_DIMS", (128, 128))
+
+data, discrete_columns = read_csv(csv_path, meta_path, header=True, discrete=None)
+print(f"[TVAE] Training on {len(data)} rows, {len(data.columns)} cols, epochs={epochs}, batch_size={batch_size}, embedding_dim={embedding_dim}, compress_dims={compress_dims}, decompress_dims={decompress_dims}")
+model = TVAE(epochs=epochs, batch_size=batch_size, embedding_dim=embedding_dim, compress_dims=compress_dims, decompress_dims=decompress_dims)
+model.fit(data, discrete_columns)
+model.save(save_path)
+print(f"[TVAE] Model saved -> {save_path}")

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/gen_20260504_174210.log

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9295c44c01b1fd97e9c805d7a481d72fcd0e87d969e89140bc0d39af40dfc04a
+size 404

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/input_snapshot.json

@@ -0,0 +1,36 @@
+{
+  "dataset_id": "c2",
+  "model": "tvae",
+  "inputs": {
+    "train_csv": {
+      "path": "/data/jialinzhang/SynthesizePipeline-server/DatasetNew/c2/c2-train.csv",
+      "exists": true,
+      "size": 42948,
+      "sha256": "17bc560fa96bd00fb3b526e1e65bc91210b701d0d0a4e8bb9b4c5196cab56def"
+    },
+    "val_csv": {
+      "path": "/data/jialinzhang/SynthesizePipeline-server/DatasetNew/c2/c2-val.csv",
+      "exists": true,
+      "size": 5349,
+      "sha256": "61e565eca62e65a7dccd9d51039a3170413379e10fc494e25870e7c4294863c9"
+    },
+    "test_csv": {
+      "path": "/data/jialinzhang/SynthesizePipeline-server/DatasetNew/c2/c2-test.csv",
+      "exists": true,
+      "size": 5448,
+      "sha256": "cbcbb062a1faf5fa44b66c80532baa229e05b94fc42137269761e6c6d84af20a"
+    },
+    "profile_json": {
+      "path": "/data/jialinzhang/SynthesizePipeline-server/DatasetNew/artifacts/data_core/tabular/c2/c2-dataset_profile.json",
+      "exists": true,
+      "size": 3240,
+      "sha256": "526b7163b2076c93c0bf4638438081ee8a6907065d5b608faa40d1a3dbc2a27b"
+    },
+    "contract_json": {
+      "path": "/data/jialinzhang/SynthesizePipeline-server/DatasetNew/artifacts/data_core/tabular/c2/c2-dataset_contract_v1.json",
+      "exists": true,
+      "size": 3731,
+      "sha256": "fb595a876054c2ee9b4e10cfe83a5691588de1d25466cbb9d473c18ad3604009"
+    }
+  }
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/models_300epochs/train_20260504_174154.log

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:545125874fecb8ddd76af96a5ee5e06601127a0dd2e37ce07f49191222e99dc3
+size 620

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/models_300epochs/tvae_300epochs.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c313f03bd4c66656f8c97200f3e68fa34fe32d3c2b0f3e6faccdd71d3f96f276
+size 344876

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/public_gate/normalized_schema_snapshot.json

@@ -0,0 +1,144 @@
+{
+  "dataset_id": "c2",
+  "target_column": "class",
+  "task_type": "classification",
+  "columns": [
+    {
+      "name": "buying",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "vhigh",
+          "med",
+          "high",
+          "low"
+        ]
+      }
+    },
+    {
+      "name": "maint",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "vhigh",
+          "low",
+          "med",
+          "high"
+        ]
+      }
+    },
+    {
+      "name": "doors",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "2",
+          "5more",
+          "3",
+          "4"
+        ]
+      }
+    },
+    {
+      "name": "persons",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "2",
+          "4",
+          "more"
+        ]
+      }
+    },
+    {
+      "name": "lug_boot",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "small",
+          "big",
+          "med"
+        ]
+      }
+    },
+    {
+      "name": "safety",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "low",
+          "high",
+          "med"
+        ]
+      }
+    },
+    {
+      "name": "class",
+      "role": "target",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "unacc",
+          "good",
+          "acc",
+          "vgood"
+        ]
+      }
+    }
+  ]
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/public_gate/public_gate_report.json

@@ -0,0 +1,37 @@
+{
+  "dataset_id": "c2",
+  "status": "pass",
+  "checks": [
+    {
+      "check_id": "PG001_csv_parse_ok",
+      "status": "pass"
+    },
+    {
+      "check_id": "PG002_split_header_consistent",
+      "status": "pass"
+    },
+    {
+      "check_id": "PG003_profile_header_match",
+      "status": "pass"
+    },
+    {
+      "check_id": "PG004_missing_token_normalized",
+      "status": "pass"
+    },
+    {
+      "check_id": "PG005_semantic_type_validated",
+      "status": "pass"
+    },
+    {
+      "check_id": "PG006_target_defined_and_valid",
+      "status": "pass"
+    }
+  ],
+  "target_column": "class",
+  "task_type": "classification",
+  "input_splits": {
+    "train": "/data/jialinzhang/SynthesizePipeline-server/DatasetNew/c2/c2-train.csv",
+    "val": "/data/jialinzhang/SynthesizePipeline-server/DatasetNew/c2/c2-val.csv",
+    "test": "/data/jialinzhang/SynthesizePipeline-server/DatasetNew/c2/c2-test.csv"
+  }
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/public_gate/staged_input_manifest.json

@@ -0,0 +1,149 @@
+{
+  "dataset_id": "c2",
+  "target_column": "class",
+  "task_type": "classification",
+  "train_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/staged/public/train.csv",
+  "val_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/staged/public/val.csv",
+  "test_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/staged/public/test.csv",
+  "features_json": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/staged/public/staged_features.json",
+  "public_gate_report": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/public_gate/public_gate_report.json",
+  "column_schema": [
+    {
+      "name": "buying",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "vhigh",
+          "med",
+          "high",
+          "low"
+        ]
+      }
+    },
+    {
+      "name": "maint",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "vhigh",
+          "low",
+          "med",
+          "high"
+        ]
+      }
+    },
+    {
+      "name": "doors",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "2",
+          "5more",
+          "3",
+          "4"
+        ]
+      }
+    },
+    {
+      "name": "persons",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "2",
+          "4",
+          "more"
+        ]
+      }
+    },
+    {
+      "name": "lug_boot",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "small",
+          "big",
+          "med"
+        ]
+      }
+    },
+    {
+      "name": "safety",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "low",
+          "high",
+          "med"
+        ]
+      }
+    },
+    {
+      "name": "class",
+      "role": "target",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "unacc",
+          "good",
+          "acc",
+          "vgood"
+        ]
+      }
+    }
+  ]
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/run_config.json

@@ -0,0 +1,45 @@
+{
+  "schema_version": 1,
+  "recorded_at": "2026-05-04T17:41:54",
+  "dataset_id": "c2",
+  "model": "tvae",
+  "work_dir": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154",
+  "dataset_source_requested": "new",
+  "dataset_source_resolved": "new",
+  "cli_args": {
+    "model": "tvae",
+    "dataset": "c2",
+    "dataset_source": "new",
+    "train": true,
+    "generate": true,
+    "num_rows": 0,
+    "epochs": null,
+    "output_dir": null,
+    "model_dir": null,
+    "work_dir": null,
+    "resume": false,
+    "no_stats": false
+  },
+  "resolved": {
+    "num_rows": 1382,
+    "model_path": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/models_300epochs/tvae_300epochs.pt",
+    "output_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/tvae-c2-1382-20260504_174210.csv"
+  },
+  "input_artifacts": {
+    "public_gate_report": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/public_gate/public_gate_report.json",
+    "public_manifest": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/public_gate/staged_input_manifest.json",
+    "model_manifest": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/staged/tvae/model_input_manifest.json",
+    "train_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/staged/public/train.csv",
+    "features_json": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/staged/public/staged_features.json",
+    "target_column": "class",
+    "task_type": "classification"
+  },
+  "env_overrides": {
+    "BENCHMARK_TVAE_GPUS": "device=3",
+    "TVAE_BATCH_SIZE": "256",
+    "TVAE_COMPRESS_DIMS": "256,256",
+    "TVAE_DECOMPRESS_DIMS": "256,256",
+    "TVAE_EMBEDDING_DIM": "128",
+    "TVAE_EPOCHS": "500"
+  }
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/runtime_result.json

@@ -0,0 +1,27 @@
+{
+  "dataset_id": "c2",
+  "model": "tvae",
+  "run_id": "tvae-c2-20260504_174154",
+  "public_gate_status": "pass",
+  "adapter_ready_status": "pass",
+  "train_status": "success",
+  "generate_status": "success",
+  "reason_code": null,
+  "reason_detail": null,
+  "artifacts": {
+    "synthetic_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/tvae-c2-1382-20260504_174210.csv",
+    "model_path": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/models_300epochs/tvae_300epochs.pt"
+  },
+  "timings": {
+    "train": {
+      "started_at": "2026-05-04T17:41:54",
+      "ended_at": "2026-05-04T17:42:10",
+      "duration_sec": 15.509
+    },
+    "generate": {
+      "started_at": "2026-05-04T17:42:10",
+      "ended_at": "2026-05-04T17:42:15",
+      "duration_sec": 5.341
+    }
+  }
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/staged/public/staged_features.json

@@ -0,0 +1,37 @@
+[
+  {
+    "feature_name": "buying",
+    "data_type": "categorical",
+    "is_target": false
+  },
+  {
+    "feature_name": "maint",
+    "data_type": "categorical",
+    "is_target": false
+  },
+  {
+    "feature_name": "doors",
+    "data_type": "categorical",
+    "is_target": false
+  },
+  {
+    "feature_name": "persons",
+    "data_type": "categorical",
+    "is_target": false
+  },
+  {
+    "feature_name": "lug_boot",
+    "data_type": "categorical",
+    "is_target": false
+  },
+  {
+    "feature_name": "safety",
+    "data_type": "categorical",
+    "is_target": false
+  },
+  {
+    "feature_name": "class",
+    "data_type": "categorical",
+    "is_target": true
+  }
+]

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/staged/public/test.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b48114a7d0bc5bd9a07920f903c8d4aba8bf98bf2a66a050da03588b0245ca73
+size 5273

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/staged/public/train.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4aed00c2c2b3f88a55a7ebff31b2e1b5e0e32fb0a7267e0b9d2779cd23e434dd
+size 41565

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/staged/public/val.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:26e90c1170a57a14c05832ac88027722b1f3848f9662c7c09ef7c93dcba4cc01
+size 5176

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/staged/tvae/adapter_report.json

@@ -0,0 +1,7 @@
+{
+  "adapter_ready_status": "pass",
+  "adapter_fail_reason_code": null,
+  "adapter_fail_detail": null,
+  "adapter_transforms_applied": [],
+  "model_input_manifest": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/staged/tvae/model_input_manifest.json"
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/staged/tvae/adapter_transforms_applied.json

@@ -0,0 +1 @@
+[]

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/staged/tvae/model_input_manifest.json

@@ -0,0 +1,151 @@
+{
+  "dataset_id": "c2",
+  "model": "tvae",
+  "target_column": "class",
+  "task_type": "classification",
+  "column_schema": [
+    {
+      "name": "buying",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "vhigh",
+          "med",
+          "high",
+          "low"
+        ]
+      }
+    },
+    {
+      "name": "maint",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "vhigh",
+          "low",
+          "med",
+          "high"
+        ]
+      }
+    },
+    {
+      "name": "doors",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "2",
+          "5more",
+          "3",
+          "4"
+        ]
+      }
+    },
+    {
+      "name": "persons",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "2",
+          "4",
+          "more"
+        ]
+      }
+    },
+    {
+      "name": "lug_boot",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "small",
+          "big",
+          "med"
+        ]
+      }
+    },
+    {
+      "name": "safety",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "low",
+          "high",
+          "med"
+        ]
+      }
+    },
+    {
+      "name": "class",
+      "role": "target",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "unacc",
+          "good",
+          "acc",
+          "vgood"
+        ]
+      }
+    }
+  ],
+  "public_manifest": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/public_gate/staged_input_manifest.json",
+  "train_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/staged/public/train.csv",
+  "val_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/staged/public/val.csv",
+  "test_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/staged/public/test.csv",
+  "features_json": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/staged/public/staged_features.json",
+  "public_gate_report": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174154/public_gate/public_gate_report.json"
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/tvae-c2-1382-20260504_174210.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:94ee321408fa4dd67c768ba7f1522bec0d6ad27f49ecc50eb179d2ae844dd3f5
+size 41485

SynthData0523/main/c2/tvae/tvae-c2-20260504_174154/tvae_metadata.json

@@ -0,0 +1,32 @@
+{
+  "columns": [
+    {
+      "name": "buying",
+      "type": "categorical"
+    },
+    {
+      "name": "maint",
+      "type": "categorical"
+    },
+    {
+      "name": "doors",
+      "type": "categorical"
+    },
+    {
+      "name": "persons",
+      "type": "categorical"
+    },
+    {
+      "name": "lug_boot",
+      "type": "categorical"
+    },
+    {
+      "name": "safety",
+      "type": "categorical"
+    },
+    {
+      "name": "class",
+      "type": "categorical"
+    }
+  ]
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/_tvae_generate.py

@@ -0,0 +1,23 @@
+import os, sys
+sys.path.insert(0, "/work")
+from src.SpecificModels.ctgan_joblib_parallel_cap import apply_parallel_cap_from_env
+apply_parallel_cap_from_env()
+from src.SpecificModels.ctgan_rdt_inverse_fix import apply_ctgan_inverse_fix
+apply_ctgan_inverse_fix()
+import pandas as pd
+from ctgan.synthesizers.tvae import TVAE
+os.environ.setdefault("LOKY_MAX_CPU_COUNT", "8")
+os.environ.setdefault("OPENBLAS_NUM_THREADS", "8")
+os.environ.setdefault("MKL_NUM_THREADS", "8")
+model = TVAE.load("/work/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/models_300epochs/tvae_300epochs.pt")
+total = 1382
+chunk = min(50000, total) if total > 50000 else total
+parts = []
+left = total
+while left > 0:
+    take = min(chunk, left)
+    parts.append(model.sample(take))
+    left -= take
+samples = pd.concat(parts, ignore_index=True) if len(parts) > 1 else parts[0]
+samples.to_csv("/work/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/tvae-c2-1382-20260504_174245.csv", index=False)
+print(f"[TVAE] Generated {total} rows (chunks={len(parts)}) -> /work/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/tvae-c2-1382-20260504_174245.csv")

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/_tvae_train.py

@@ -0,0 +1,39 @@
+import json, os, sys
+sys.path.insert(0, "/work")
+from src.SpecificModels.ctgan_joblib_parallel_cap import apply_parallel_cap_from_env
+apply_parallel_cap_from_env()
+import pandas as pd
+from ctgan.data import read_csv
+from ctgan.synthesizers.tvae import TVAE
+
+# Keep transform stage parallelism bounded for stability on shared host.
+os.environ.setdefault("LOKY_MAX_CPU_COUNT", "8")
+os.environ.setdefault("OPENBLAS_NUM_THREADS", "8")
+os.environ.setdefault("MKL_NUM_THREADS", "8")
+_nj = (os.environ.get("TVAE_CTGAN_JOBTRANS_N_JOBS") or "").strip()
+if _nj:
+    print("[TVAE] joblib Parallel cap ON, TVAE_CTGAN_JOBTRANS_N_JOBS=" + _nj)
+else:
+    print("[TVAE] joblib Parallel cap OFF (unset TVAE_CTGAN_JOBTRANS_N_JOBS)")
+print("[TVAE] LOKY_MAX_CPU_COUNT=" + str(os.environ.get("LOKY_MAX_CPU_COUNT", "")))
+
+csv_path = "/work/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/staged/public/train.csv"
+meta_path = "/work/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/tvae_metadata.json"
+save_path = "/work/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/models_300epochs/tvae_300epochs.pt"
+epochs = int(os.environ.get("TVAE_EPOCHS", 300))
+batch_size = int(os.environ.get("TVAE_BATCH_SIZE", "500"))
+embedding_dim = int(os.environ.get("TVAE_EMBEDDING_DIM", "128"))
+def _parse_dims(name, default):
+    raw = (os.environ.get(name) or "").strip()
+    if not raw:
+        return default
+    return tuple(int(x.strip()) for x in raw.split(",") if x.strip())
+compress_dims = _parse_dims("TVAE_COMPRESS_DIMS", (128, 128))
+decompress_dims = _parse_dims("TVAE_DECOMPRESS_DIMS", (128, 128))
+
+data, discrete_columns = read_csv(csv_path, meta_path, header=True, discrete=None)
+print(f"[TVAE] Training on {len(data)} rows, {len(data.columns)} cols, epochs={epochs}, batch_size={batch_size}, embedding_dim={embedding_dim}, compress_dims={compress_dims}, decompress_dims={decompress_dims}")
+model = TVAE(epochs=epochs, batch_size=batch_size, embedding_dim=embedding_dim, compress_dims=compress_dims, decompress_dims=decompress_dims)
+model.fit(data, discrete_columns)
+model.save(save_path)
+print(f"[TVAE] Model saved -> {save_path}")

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/gen_20260504_174245.log

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:52660e570f2b1798f41853f45ac28d1c585f75b473dc0bafb63cea59cf9cd40d
+size 404

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/input_snapshot.json

@@ -0,0 +1,36 @@
+{
+  "dataset_id": "c2",
+  "model": "tvae",
+  "inputs": {
+    "train_csv": {
+      "path": "/data/jialinzhang/SynthesizePipeline-server/DatasetNew/c2/c2-train.csv",
+      "exists": true,
+      "size": 42948,
+      "sha256": "17bc560fa96bd00fb3b526e1e65bc91210b701d0d0a4e8bb9b4c5196cab56def"
+    },
+    "val_csv": {
+      "path": "/data/jialinzhang/SynthesizePipeline-server/DatasetNew/c2/c2-val.csv",
+      "exists": true,
+      "size": 5349,
+      "sha256": "61e565eca62e65a7dccd9d51039a3170413379e10fc494e25870e7c4294863c9"
+    },
+    "test_csv": {
+      "path": "/data/jialinzhang/SynthesizePipeline-server/DatasetNew/c2/c2-test.csv",
+      "exists": true,
+      "size": 5448,
+      "sha256": "cbcbb062a1faf5fa44b66c80532baa229e05b94fc42137269761e6c6d84af20a"
+    },
+    "profile_json": {
+      "path": "/data/jialinzhang/SynthesizePipeline-server/DatasetNew/artifacts/data_core/tabular/c2/c2-dataset_profile.json",
+      "exists": true,
+      "size": 3240,
+      "sha256": "526b7163b2076c93c0bf4638438081ee8a6907065d5b608faa40d1a3dbc2a27b"
+    },
+    "contract_json": {
+      "path": "/data/jialinzhang/SynthesizePipeline-server/DatasetNew/artifacts/data_core/tabular/c2/c2-dataset_contract_v1.json",
+      "exists": true,
+      "size": 3731,
+      "sha256": "fb595a876054c2ee9b4e10cfe83a5691588de1d25466cbb9d473c18ad3604009"
+    }
+  }
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/models_300epochs/train_20260504_174228.log

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:58c60bb9fe86a317efabacf81546eb2b25f680a81383de073a9feceb8729d822
+size 620

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/models_300epochs/tvae_300epochs.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:15fc84c398c90a6b1dfa00fc6c2395643c8ca9901216eb12f9c29c1389c2775b
+size 344876

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/public_gate/normalized_schema_snapshot.json

@@ -0,0 +1,144 @@
+{
+  "dataset_id": "c2",
+  "target_column": "class",
+  "task_type": "classification",
+  "columns": [
+    {
+      "name": "buying",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "vhigh",
+          "med",
+          "high",
+          "low"
+        ]
+      }
+    },
+    {
+      "name": "maint",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "vhigh",
+          "low",
+          "med",
+          "high"
+        ]
+      }
+    },
+    {
+      "name": "doors",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "2",
+          "5more",
+          "3",
+          "4"
+        ]
+      }
+    },
+    {
+      "name": "persons",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "2",
+          "4",
+          "more"
+        ]
+      }
+    },
+    {
+      "name": "lug_boot",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "small",
+          "big",
+          "med"
+        ]
+      }
+    },
+    {
+      "name": "safety",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "low",
+          "high",
+          "med"
+        ]
+      }
+    },
+    {
+      "name": "class",
+      "role": "target",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "unacc",
+          "good",
+          "acc",
+          "vgood"
+        ]
+      }
+    }
+  ]
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/public_gate/public_gate_report.json

@@ -0,0 +1,37 @@
+{
+  "dataset_id": "c2",
+  "status": "pass",
+  "checks": [
+    {
+      "check_id": "PG001_csv_parse_ok",
+      "status": "pass"
+    },
+    {
+      "check_id": "PG002_split_header_consistent",
+      "status": "pass"
+    },
+    {
+      "check_id": "PG003_profile_header_match",
+      "status": "pass"
+    },
+    {
+      "check_id": "PG004_missing_token_normalized",
+      "status": "pass"
+    },
+    {
+      "check_id": "PG005_semantic_type_validated",
+      "status": "pass"
+    },
+    {
+      "check_id": "PG006_target_defined_and_valid",
+      "status": "pass"
+    }
+  ],
+  "target_column": "class",
+  "task_type": "classification",
+  "input_splits": {
+    "train": "/data/jialinzhang/SynthesizePipeline-server/DatasetNew/c2/c2-train.csv",
+    "val": "/data/jialinzhang/SynthesizePipeline-server/DatasetNew/c2/c2-val.csv",
+    "test": "/data/jialinzhang/SynthesizePipeline-server/DatasetNew/c2/c2-test.csv"
+  }
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/public_gate/staged_input_manifest.json

@@ -0,0 +1,149 @@
+{
+  "dataset_id": "c2",
+  "target_column": "class",
+  "task_type": "classification",
+  "train_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/staged/public/train.csv",
+  "val_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/staged/public/val.csv",
+  "test_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/staged/public/test.csv",
+  "features_json": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/staged/public/staged_features.json",
+  "public_gate_report": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/public_gate/public_gate_report.json",
+  "column_schema": [
+    {
+      "name": "buying",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "vhigh",
+          "med",
+          "high",
+          "low"
+        ]
+      }
+    },
+    {
+      "name": "maint",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "vhigh",
+          "low",
+          "med",
+          "high"
+        ]
+      }
+    },
+    {
+      "name": "doors",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "2",
+          "5more",
+          "3",
+          "4"
+        ]
+      }
+    },
+    {
+      "name": "persons",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "2",
+          "4",
+          "more"
+        ]
+      }
+    },
+    {
+      "name": "lug_boot",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "small",
+          "big",
+          "med"
+        ]
+      }
+    },
+    {
+      "name": "safety",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "low",
+          "high",
+          "med"
+        ]
+      }
+    },
+    {
+      "name": "class",
+      "role": "target",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "unacc",
+          "good",
+          "acc",
+          "vgood"
+        ]
+      }
+    }
+  ]
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/run_config.json

@@ -0,0 +1,45 @@
+{
+  "schema_version": 1,
+  "recorded_at": "2026-05-04T17:42:28",
+  "dataset_id": "c2",
+  "model": "tvae",
+  "work_dir": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228",
+  "dataset_source_requested": "new",
+  "dataset_source_resolved": "new",
+  "cli_args": {
+    "model": "tvae",
+    "dataset": "c2",
+    "dataset_source": "new",
+    "train": true,
+    "generate": true,
+    "num_rows": 0,
+    "epochs": null,
+    "output_dir": null,
+    "model_dir": null,
+    "work_dir": null,
+    "resume": false,
+    "no_stats": false
+  },
+  "resolved": {
+    "num_rows": 1382,
+    "model_path": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/models_300epochs/tvae_300epochs.pt",
+    "output_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/tvae-c2-1382-20260504_174245.csv"
+  },
+  "input_artifacts": {
+    "public_gate_report": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/public_gate/public_gate_report.json",
+    "public_manifest": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/public_gate/staged_input_manifest.json",
+    "model_manifest": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/staged/tvae/model_input_manifest.json",
+    "train_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/staged/public/train.csv",
+    "features_json": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/staged/public/staged_features.json",
+    "target_column": "class",
+    "task_type": "classification"
+  },
+  "env_overrides": {
+    "BENCHMARK_TVAE_GPUS": "device=3",
+    "TVAE_BATCH_SIZE": "256",
+    "TVAE_COMPRESS_DIMS": "256,256",
+    "TVAE_DECOMPRESS_DIMS": "256,256",
+    "TVAE_EMBEDDING_DIM": "256",
+    "TVAE_EPOCHS": "500"
+  }
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/runtime_result.json

@@ -0,0 +1,27 @@
+{
+  "dataset_id": "c2",
+  "model": "tvae",
+  "run_id": "tvae-c2-20260504_174228",
+  "public_gate_status": "pass",
+  "adapter_ready_status": "pass",
+  "train_status": "success",
+  "generate_status": "success",
+  "reason_code": null,
+  "reason_detail": null,
+  "artifacts": {
+    "synthetic_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/tvae-c2-1382-20260504_174245.csv",
+    "model_path": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/models_300epochs/tvae_300epochs.pt"
+  },
+  "timings": {
+    "train": {
+      "started_at": "2026-05-04T17:42:28",
+      "ended_at": "2026-05-04T17:42:45",
+      "duration_sec": 16.727
+    },
+    "generate": {
+      "started_at": "2026-05-04T17:42:45",
+      "ended_at": "2026-05-04T17:42:50",
+      "duration_sec": 4.871
+    }
+  }
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/staged/public/staged_features.json

@@ -0,0 +1,37 @@
+[
+  {
+    "feature_name": "buying",
+    "data_type": "categorical",
+    "is_target": false
+  },
+  {
+    "feature_name": "maint",
+    "data_type": "categorical",
+    "is_target": false
+  },
+  {
+    "feature_name": "doors",
+    "data_type": "categorical",
+    "is_target": false
+  },
+  {
+    "feature_name": "persons",
+    "data_type": "categorical",
+    "is_target": false
+  },
+  {
+    "feature_name": "lug_boot",
+    "data_type": "categorical",
+    "is_target": false
+  },
+  {
+    "feature_name": "safety",
+    "data_type": "categorical",
+    "is_target": false
+  },
+  {
+    "feature_name": "class",
+    "data_type": "categorical",
+    "is_target": true
+  }
+]

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/staged/public/test.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b48114a7d0bc5bd9a07920f903c8d4aba8bf98bf2a66a050da03588b0245ca73
+size 5273

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/staged/public/train.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4aed00c2c2b3f88a55a7ebff31b2e1b5e0e32fb0a7267e0b9d2779cd23e434dd
+size 41565

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/staged/public/val.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:26e90c1170a57a14c05832ac88027722b1f3848f9662c7c09ef7c93dcba4cc01
+size 5176

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/staged/tvae/adapter_report.json

@@ -0,0 +1,7 @@
+{
+  "adapter_ready_status": "pass",
+  "adapter_fail_reason_code": null,
+  "adapter_fail_detail": null,
+  "adapter_transforms_applied": [],
+  "model_input_manifest": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/staged/tvae/model_input_manifest.json"
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/staged/tvae/adapter_transforms_applied.json

@@ -0,0 +1 @@
+[]

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/staged/tvae/model_input_manifest.json

@@ -0,0 +1,151 @@
+{
+  "dataset_id": "c2",
+  "model": "tvae",
+  "target_column": "class",
+  "task_type": "classification",
+  "column_schema": [
+    {
+      "name": "buying",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "vhigh",
+          "med",
+          "high",
+          "low"
+        ]
+      }
+    },
+    {
+      "name": "maint",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "vhigh",
+          "low",
+          "med",
+          "high"
+        ]
+      }
+    },
+    {
+      "name": "doors",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "2",
+          "5more",
+          "3",
+          "4"
+        ]
+      }
+    },
+    {
+      "name": "persons",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "2",
+          "4",
+          "more"
+        ]
+      }
+    },
+    {
+      "name": "lug_boot",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "small",
+          "big",
+          "med"
+        ]
+      }
+    },
+    {
+      "name": "safety",
+      "role": "feature",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 3,
+        "unique_ratio": 0.002171,
+        "example_values": [
+          "low",
+          "high",
+          "med"
+        ]
+      }
+    },
+    {
+      "name": "class",
+      "role": "target",
+      "semantic_type": "categorical",
+      "nullable": false,
+      "missing_tokens": [],
+      "parse_format": null,
+      "impute_strategy": "mode",
+      "profile_stats": {
+        "missing_rate": 0.0,
+        "unique_count": 4,
+        "unique_ratio": 0.002894,
+        "example_values": [
+          "unacc",
+          "good",
+          "acc",
+          "vgood"
+        ]
+      }
+    }
+  ],
+  "public_manifest": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/public_gate/staged_input_manifest.json",
+  "train_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/staged/public/train.csv",
+  "val_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/staged/public/val.csv",
+  "test_csv": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/staged/public/test.csv",
+  "features_json": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/staged/public/staged_features.json",
+  "public_gate_report": "/data/jialinzhang/SynthesizePipeline-server/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_174228/public_gate/public_gate_report.json"
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/tvae-c2-1382-20260504_174245.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ed8d81c7986042bdc6d46d880cf4b8ecbb4ab544eed3754733fc1ee8424db47d
+size 41641

SynthData0523/main/c2/tvae/tvae-c2-20260504_174228/tvae_metadata.json

@@ -0,0 +1,32 @@
+{
+  "columns": [
+    {
+      "name": "buying",
+      "type": "categorical"
+    },
+    {
+      "name": "maint",
+      "type": "categorical"
+    },
+    {
+      "name": "doors",
+      "type": "categorical"
+    },
+    {
+      "name": "persons",
+      "type": "categorical"
+    },
+    {
+      "name": "lug_boot",
+      "type": "categorical"
+    },
+    {
+      "name": "safety",
+      "type": "categorical"
+    },
+    {
+      "name": "class",
+      "type": "categorical"
+    }
+  ]
+}

SynthData0523/main/c2/tvae/tvae-c2-20260504_175027/_tvae_generate.py

@@ -0,0 +1,23 @@
+import os, sys
+sys.path.insert(0, "/work")
+from src.SpecificModels.ctgan_joblib_parallel_cap import apply_parallel_cap_from_env
+apply_parallel_cap_from_env()
+from src.SpecificModels.ctgan_rdt_inverse_fix import apply_ctgan_inverse_fix
+apply_ctgan_inverse_fix()
+import pandas as pd
+from ctgan.synthesizers.tvae import TVAE
+os.environ.setdefault("LOKY_MAX_CPU_COUNT", "8")
+os.environ.setdefault("OPENBLAS_NUM_THREADS", "8")
+os.environ.setdefault("MKL_NUM_THREADS", "8")
+model = TVAE.load("/work/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_175027/models_300epochs/tvae_300epochs.pt")
+total = 1382
+chunk = min(50000, total) if total > 50000 else total
+parts = []
+left = total
+while left > 0:
+    take = min(chunk, left)
+    parts.append(model.sample(take))
+    left -= take
+samples = pd.concat(parts, ignore_index=True) if len(parts) > 1 else parts[0]
+samples.to_csv("/work/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_175027/tvae-c2-1382-20260504_175041.csv", index=False)
+print(f"[TVAE] Generated {total} rows (chunks={len(parts)}) -> /work/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_175027/tvae-c2-1382-20260504_175041.csv")

SynthData0523/main/c2/tvae/tvae-c2-20260504_175027/_tvae_train.py

@@ -0,0 +1,39 @@
+import json, os, sys
+sys.path.insert(0, "/work")
+from src.SpecificModels.ctgan_joblib_parallel_cap import apply_parallel_cap_from_env
+apply_parallel_cap_from_env()
+import pandas as pd
+from ctgan.data import read_csv
+from ctgan.synthesizers.tvae import TVAE
+
+# Keep transform stage parallelism bounded for stability on shared host.
+os.environ.setdefault("LOKY_MAX_CPU_COUNT", "8")
+os.environ.setdefault("OPENBLAS_NUM_THREADS", "8")
+os.environ.setdefault("MKL_NUM_THREADS", "8")
+_nj = (os.environ.get("TVAE_CTGAN_JOBTRANS_N_JOBS") or "").strip()
+if _nj:
+    print("[TVAE] joblib Parallel cap ON, TVAE_CTGAN_JOBTRANS_N_JOBS=" + _nj)
+else:
+    print("[TVAE] joblib Parallel cap OFF (unset TVAE_CTGAN_JOBTRANS_N_JOBS)")
+print("[TVAE] LOKY_MAX_CPU_COUNT=" + str(os.environ.get("LOKY_MAX_CPU_COUNT", "")))
+
+csv_path = "/work/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_175027/staged/public/train.csv"
+meta_path = "/work/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_175027/tvae_metadata.json"
+save_path = "/work/output-Benchmark-trainonly-v1/c2/tvae/tvae-c2-20260504_175027/models_300epochs/tvae_300epochs.pt"
+epochs = int(os.environ.get("TVAE_EPOCHS", 300))
+batch_size = int(os.environ.get("TVAE_BATCH_SIZE", "500"))
+embedding_dim = int(os.environ.get("TVAE_EMBEDDING_DIM", "128"))
+def _parse_dims(name, default):
+    raw = (os.environ.get(name) or "").strip()
+    if not raw:
+        return default
+    return tuple(int(x.strip()) for x in raw.split(",") if x.strip())
+compress_dims = _parse_dims("TVAE_COMPRESS_DIMS", (128, 128))
+decompress_dims = _parse_dims("TVAE_DECOMPRESS_DIMS", (128, 128))
+
+data, discrete_columns = read_csv(csv_path, meta_path, header=True, discrete=None)
+print(f"[TVAE] Training on {len(data)} rows, {len(data.columns)} cols, epochs={epochs}, batch_size={batch_size}, embedding_dim={embedding_dim}, compress_dims={compress_dims}, decompress_dims={decompress_dims}")
+model = TVAE(epochs=epochs, batch_size=batch_size, embedding_dim=embedding_dim, compress_dims=compress_dims, decompress_dims=decompress_dims)
+model.fit(data, discrete_columns)
+model.save(save_path)
+print(f"[TVAE] Model saved -> {save_path}")