Upload folder using huggingface_hub

dataset.py

@@ -1,47 +1,47 @@
-import json
-import datasets
-
-_DESCRIPTION = """Titan CV Agent Benchmark：用于评测智能体在复杂视觉任务下的推理与决策能力。包含图片与视频样本，以及配套的问答任务。"""
-_HOMEPAGE = "https://huggingface.co/datasets/DataCanvasAILab/Titan-CV-Agent-Benchmark"
-_LICENSE = "MIT"
-_DATA_URL = "https://huggingface.co/datasets/DataCanvasAILab/Titan-CV-Agent-Benchmark/resolve/main/titan_agent_benchmark.json"
-
-class TitanCVAgentBenchmark(datasets.GeneratorBasedBuilder):
-    VERSION = datasets.Version("1.0.0")
-
-    def _info(self):
-        return datasets.DatasetInfo(
-            description=_DESCRIPTION,
-            homepage=_HOMEPAGE,
-            license=_LICENSE,
-            features=datasets.Features({
-                "id": datasets.Value("string"),
-                "media_path": datasets.Value("string"),
-                "media_type": datasets.ClassLabel(names=["image", "video"]),
-                "query": datasets.Value("string"),
-                "answer": datasets.Value("string"),
-                "note": datasets.Value("string"),
-            }),
-        )
-
-    def _split_generators(self, dl_manager):
-        data_path = dl_manager.download(_DATA_URL)
-        return [
-            datasets.SplitGenerator(
-                name=datasets.Split.TRAIN,
-                gen_kwargs={"filepath": data_path},
-            )
-        ]
-
-    def _generate_examples(self, filepath):
-        with open(filepath, encoding="utf-8") as f:
-            data = json.load(f)
-            for idx, row in enumerate(data):
-                yield idx, {
-                    "id": row["id"],
-                    "media_path": row["media_path"],
-                    "media_type": row["media_type"],
-                    "query": row["query"],
-                    "answer": row["answer"],
-                    "note": row["note"],
-                }
+import json
+import datasets
+
+_DESCRIPTION = """Titan CV Agent Benchmark：用于评测智能体在复杂视觉任务下的推理与决策能力。包含图片与视频样本，以及配套的问答任务。"""
+_HOMEPAGE = "https://huggingface.co/datasets/DataCanvasAILab/Titan-CV-Agent-Benchmark"
+_LICENSE = "MIT"
+_DATA_URL = "https://huggingface.co/datasets/DataCanvasAILab/Titan-CV-Agent-Benchmark/resolve/main/titan_agent_benchmark_v3.json"
+
+class Dataset(datasets.GeneratorBasedBuilder):  # ← 关键改成 Dataset
+    VERSION = datasets.Version("1.0.0")
+
+    def _info(self):
+        return datasets.DatasetInfo(
+            description=_DESCRIPTION,
+            homepage=_HOMEPAGE,
+            license=_LICENSE,
+            features=datasets.Features({
+                "id": datasets.Value("string"),
+                "media_path": datasets.Value("string"),
+                "media_type": datasets.ClassLabel(names=["image", "video"]),
+                "query": datasets.Value("string"),
+                "answer": datasets.Value("string"),
+                "note": datasets.Value("string"),
+            }),
+        )
+
+    def _split_generators(self, dl_manager):
+        data_path = dl_manager.download(_DATA_URL)
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                gen_kwargs={"filepath": data_path},
+            )
+        ]
+
+    def _generate_examples(self, filepath):
+        with open(filepath, encoding="utf-8") as f:
+            data = json.load(f)
+            for idx, row in enumerate(data):
+                yield idx, {
+                    "id": row["id"],
+                    "media_path": row["media_path"],
+                    "media_type": row["media_type"],
+                    "query": row["query"],
+                    "answer": row["answer"],
+                    "note": row["note"],
+                }

dataset_info.json

@@ -0,0 +1,16 @@
+{
+  "default": {
+    "description": "Titan CV Agent Benchmark：用于评测智能体在复杂视觉任务下的推理与决策能力。",
+    "features": {
+      "id": {"dtype": "string", "_type": "Value"},
+      "media_path": {"dtype": "string", "_type": "Value"},
+      "media_type": {"names": ["image", "video"], "_type": "ClassLabel"},
+      "query": {"dtype": "string", "_type": "Value"},
+      "answer": {"dtype": "string", "_type": "Value"},
+      "note": {"dtype": "string", "_type": "Value"}
+    },
+    "homepage": "https://huggingface.co/datasets/DataCanvasAILab/Titan-CV-Agent-Benchmark",
+    "license": "MIT",
+    "citation": ""
+  }
+}