diff --git a/README.md b/README.md
index 4959aee0476a602b2d762675016f4a658b2c0da5..6d22a388b07c6c591e2cfa67010e2dbec540826a 100644
--- a/README.md
+++ b/README.md
@@ -158,22 +158,22 @@ uv run python example_usage.py
 ### Sample JSONL Records
 
 ```json
-{"file_name": "data/audio/xin_chao.mp3", "word": "xin chào", "ipa": "sin˧˧ caːw˨˩"}
-{"file_name": "data/audio/cam_on.mp3", "word": "cảm ơn", "ipa": "kaːm˧˩ ʔəːn˧˧"}
-{"file_name": "data/audio/viet_nam.mp3", "word": "Việt Nam", "ipa": "viət˧˩ˀ naːm˧˧"}
-{"file_name": "data/audio/nuoc.mp3", "word": "nước", "ipa": "nɨək˧˥"}
-{"file_name": "data/audio/yeu.mp3", "word": "yêu", "ipa": "iəw˧˧"}
+{"file_name": "audio/xin_chao.mp3", "word": "xin chào", "ipa": "sin˧˧ caːw˨˩"}
+{"file_name": "audio/cam_on.mp3", "word": "cảm ơn", "ipa": "kaːm˧˩ ʔəːn˧˧"}
+{"file_name": "audio/viet_nam.mp3", "word": "Việt Nam", "ipa": "viət˧˩ˀ naːm˧˧"}
+{"file_name": "audio/nuoc.mp3", "word": "nước", "ipa": "nɨək˧˥"}
+{"file_name": "audio/yeu.mp3", "word": "yêu", "ipa": "iəw˧˧"}
 ```
 
 ### Data Table View
 
 | Audio File | Word | IPA |
 |------------|------|-----|
-| data/audio/xin_chao.mp3 | xin chào | sin˧˧ caːw˨˩ |
-| data/audio/cam_on.mp3 | cảm ơn | kaːm˧˩ ʔəːn˧˧ |
-| data/audio/viet_nam.mp3 | Việt Nam | viət˧˩ˀ naːm˧˧ |
-| data/audio/nuoc.mp3 | nước | nɨək˧˥ |
-| data/audio/yeu.mp3 | yêu | iəw˧˧ |
+| audio/xin_chao.mp3 | xin chào | sin˧˧ caːw˨˩ |
+| audio/cam_on.mp3 | cảm ơn | kaːm˧˩ ʔəːn˧˧ |
+| audio/viet_nam.mp3 | Việt Nam | viət˧˩ˀ naːm˧˧ |
+| audio/nuoc.mp3 | nước | nɨək˧˥ |
+| audio/yeu.mp3 | yêu | iəw˧˧ |
 
 ## Considerations for Using the Data
 
diff --git a/data/audio/an.mp3 b/audio/an.mp3
similarity index 100%
rename from data/audio/an.mp3
rename to audio/an.mp3
diff --git a/data/audio/anh.mp3 b/audio/anh.mp3
similarity index 100%
rename from data/audio/anh.mp3
rename to audio/anh.mp3
diff --git a/data/audio/ba.mp3 b/audio/ba.mp3
similarity index 100%
rename from data/audio/ba.mp3
rename to audio/ba.mp3
diff --git a/data/audio/ba_number.mp3 b/audio/ba_number.mp3
similarity index 100%
rename from data/audio/ba_number.mp3
rename to audio/ba_number.mp3
diff --git a/data/audio/ba_tone.mp3 b/audio/ba_tone.mp3
similarity index 100%
rename from data/audio/ba_tone.mp3
rename to audio/ba_tone.mp3
diff --git a/data/audio/ban.mp3 b/audio/ban.mp3
similarity index 100%
rename from data/audio/ban.mp3
rename to audio/ban.mp3
diff --git a/data/audio/bon.mp3 b/audio/bon.mp3
similarity index 100%
rename from data/audio/bon.mp3
rename to audio/bon.mp3
diff --git a/data/audio/cam_on.mp3 b/audio/cam_on.mp3
similarity index 100%
rename from data/audio/cam_on.mp3
rename to audio/cam_on.mp3
diff --git a/data/audio/cay.mp3 b/audio/cay.mp3
similarity index 100%
rename from data/audio/cay.mp3
rename to audio/cay.mp3
diff --git a/data/audio/chi.mp3 b/audio/chi.mp3
similarity index 100%
rename from data/audio/chi.mp3
rename to audio/chi.mp3
diff --git a/data/audio/chieu.mp3 b/audio/chieu.mp3
similarity index 100%
rename from data/audio/chieu.mp3
rename to audio/chieu.mp3
diff --git a/data/audio/cho.mp3 b/audio/cho.mp3
similarity index 100%
rename from data/audio/cho.mp3
rename to audio/cho.mp3
diff --git a/data/audio/co.mp3 b/audio/co.mp3
similarity index 100%
rename from data/audio/co.mp3
rename to audio/co.mp3
diff --git a/data/audio/con.mp3 b/audio/con.mp3
similarity index 100%
rename from data/audio/con.mp3
rename to audio/con.mp3
diff --git a/data/audio/dem.mp3 b/audio/dem.mp3
similarity index 100%
rename from data/audio/dem.mp3
rename to audio/dem.mp3
diff --git a/data/audio/dep.mp3 b/audio/dep.mp3
similarity index 100%
rename from data/audio/dep.mp3
rename to audio/dep.mp3
diff --git a/data/audio/duong.mp3 b/audio/duong.mp3
similarity index 100%
rename from data/audio/duong.mp3
rename to audio/duong.mp3
diff --git a/data/audio/em.mp3 b/audio/em.mp3
similarity index 100%
rename from data/audio/em.mp3
rename to audio/em.mp3
diff --git a/data/audio/gio.mp3 b/audio/gio.mp3
similarity index 100%
rename from data/audio/gio.mp3
rename to audio/gio.mp3
diff --git a/data/audio/hai.mp3 b/audio/hai.mp3
similarity index 100%
rename from data/audio/hai.mp3
rename to audio/hai.mp3
diff --git a/data/audio/hoa.mp3 b/audio/hoa.mp3
similarity index 100%
rename from data/audio/hoa.mp3
rename to audio/hoa.mp3
diff --git a/data/audio/hoc.mp3 b/audio/hoc.mp3
similarity index 100%
rename from data/audio/hoc.mp3
rename to audio/hoc.mp3
diff --git a/data/audio/la.mp3 b/audio/la.mp3
similarity index 100%
rename from data/audio/la.mp3
rename to audio/la.mp3
diff --git a/data/audio/lam.mp3 b/audio/lam.mp3
similarity index 100%
rename from data/audio/lam.mp3
rename to audio/lam.mp3
diff --git a/data/audio/me.mp3 b/audio/me.mp3
similarity index 100%
rename from data/audio/me.mp3
rename to audio/me.mp3
diff --git a/data/audio/meo.mp3 b/audio/meo.mp3
similarity index 100%
rename from data/audio/meo.mp3
rename to audio/meo.mp3
diff --git a/data/audio/mot.mp3 b/audio/mot.mp3
similarity index 100%
rename from data/audio/mot.mp3
rename to audio/mot.mp3
diff --git a/data/audio/mua.mp3 b/audio/mua.mp3
similarity index 100%
rename from data/audio/mua.mp3
rename to audio/mua.mp3
diff --git a/data/audio/nam.mp3 b/audio/nam.mp3
similarity index 100%
rename from data/audio/nam.mp3
rename to audio/nam.mp3
diff --git a/data/audio/nang.mp3 b/audio/nang.mp3
similarity index 100%
rename from data/audio/nang.mp3
rename to audio/nang.mp3
diff --git a/data/audio/ngay.mp3 b/audio/ngay.mp3
similarity index 100%
rename from data/audio/ngay.mp3
rename to audio/ngay.mp3
diff --git a/data/audio/nha.mp3 b/audio/nha.mp3
similarity index 100%
rename from data/audio/nha.mp3
rename to audio/nha.mp3
diff --git a/data/audio/nuoc.mp3 b/audio/nuoc.mp3
similarity index 100%
rename from data/audio/nuoc.mp3
rename to audio/nuoc.mp3
diff --git a/data/audio/ong.mp3 b/audio/ong.mp3
similarity index 100%
rename from data/audio/ong.mp3
rename to audio/ong.mp3
diff --git a/data/audio/pho.mp3 b/audio/pho.mp3
similarity index 100%
rename from data/audio/pho.mp3
rename to audio/pho.mp3
diff --git a/data/audio/sach.mp3 b/audio/sach.mp3
similarity index 100%
rename from data/audio/sach.mp3
rename to audio/sach.mp3
diff --git a/data/audio/sang.mp3 b/audio/sang.mp3
similarity index 100%
rename from data/audio/sang.mp3
rename to audio/sang.mp3
diff --git a/data/audio/sinh.mp3 b/audio/sinh.mp3
similarity index 100%
rename from data/audio/sinh.mp3
rename to audio/sinh.mp3
diff --git a/data/audio/thang.mp3 b/audio/thang.mp3
similarity index 100%
rename from data/audio/thang.mp3
rename to audio/thang.mp3
diff --git a/data/audio/thay.mp3 b/audio/thay.mp3
similarity index 100%
rename from data/audio/thay.mp3
rename to audio/thay.mp3
diff --git a/data/audio/toi.mp3 b/audio/toi.mp3
similarity index 100%
rename from data/audio/toi.mp3
rename to audio/toi.mp3
diff --git a/data/audio/toi_tone.mp3 b/audio/toi_tone.mp3
similarity index 100%
rename from data/audio/toi_tone.mp3
rename to audio/toi_tone.mp3
diff --git a/data/audio/troi.mp3 b/audio/troi.mp3
similarity index 100%
rename from data/audio/troi.mp3
rename to audio/troi.mp3
diff --git a/data/audio/trua.mp3 b/audio/trua.mp3
similarity index 100%
rename from data/audio/trua.mp3
rename to audio/trua.mp3
diff --git a/data/audio/truong.mp3 b/audio/truong.mp3
similarity index 100%
rename from data/audio/truong.mp3
rename to audio/truong.mp3
diff --git a/data/audio/uong.mp3 b/audio/uong.mp3
similarity index 100%
rename from data/audio/uong.mp3
rename to audio/uong.mp3
diff --git a/data/audio/viec.mp3 b/audio/viec.mp3
similarity index 100%
rename from data/audio/viec.mp3
rename to audio/viec.mp3
diff --git a/data/audio/viet_nam.mp3 b/audio/viet_nam.mp3
similarity index 100%
rename from data/audio/viet_nam.mp3
rename to audio/viet_nam.mp3
diff --git a/data/audio/xin_chao.mp3 b/audio/xin_chao.mp3
similarity index 100%
rename from data/audio/xin_chao.mp3
rename to audio/xin_chao.mp3
diff --git a/data/audio/yeu.mp3 b/audio/yeu.mp3
similarity index 100%
rename from data/audio/yeu.mp3
rename to audio/yeu.mp3
diff --git a/data/metadata.jsonl b/data/metadata.jsonl
deleted file mode 100644
index b424a73bc8b8888fab48f46d6250a6298438b24e..0000000000000000000000000000000000000000
--- a/data/metadata.jsonl
+++ /dev/null
@@ -1,50 +0,0 @@
-{"file_name": "data/audio/xin_chao.mp3", "word": "xin chào", "ipa": "sin˧˧ caːw˨˩"}
-{"file_name": "data/audio/cam_on.mp3", "word": "cảm ơn", "ipa": "kaːm˧˩ ʔəːn˧˧"}
-{"file_name": "data/audio/viet_nam.mp3", "word": "Việt Nam", "ipa": "viət˧˩ˀ naːm˧˧"}
-{"file_name": "data/audio/nuoc.mp3", "word": "nước", "ipa": "nɨək˧˥"}
-{"file_name": "data/audio/an.mp3", "word": "ăn", "ipa": "ăn˧˧"}
-{"file_name": "data/audio/uong.mp3", "word": "uống", "ipa": "ʔuəŋ˧˥"}
-{"file_name": "data/audio/nha.mp3", "word": "nhà", "ipa": "ɲaː˨˩"}
-{"file_name": "data/audio/truong.mp3", "word": "trường", "ipa": "ʈɨəŋ˨˩"}
-{"file_name": "data/audio/ban.mp3", "word": "bạn", "ipa": "ɓaːn˧˩ˀ"}
-{"file_name": "data/audio/yeu.mp3", "word": "yêu", "ipa": "iəw˧˧"}
-{"file_name": "data/audio/dep.mp3", "word": "đẹp", "ipa": "ɗɛp˧˩ˀ"}
-{"file_name": "data/audio/sach.mp3", "word": "sách", "ipa": "saːk˧˥"}
-{"file_name": "data/audio/hoa.mp3", "word": "hoa", "ipa": "hwaː˧˧"}
-{"file_name": "data/audio/me.mp3", "word": "mẹ", "ipa": "mɛ˧˩ˀ"}
-{"file_name": "data/audio/ba.mp3", "word": "ba", "ipa": "ɓaː˧˧"}
-{"file_name": "data/audio/con.mp3", "word": "con", "ipa": "kɔn˧˧"}
-{"file_name": "data/audio/cho.mp3", "word": "chó", "ipa": "cɔ˧˥"}
-{"file_name": "data/audio/meo.mp3", "word": "mèo", "ipa": "mɛːw˨˩"}
-{"file_name": "data/audio/cay.mp3", "word": "cây", "ipa": "kəj˧˧"}
-{"file_name": "data/audio/la.mp3", "word": "lá", "ipa": "laː˧˥"}
-{"file_name": "data/audio/troi.mp3", "word": "trời", "ipa": "ʈəːj˨˩"}
-{"file_name": "data/audio/mua.mp3", "word": "mưa", "ipa": "mɨə˧˧"}
-{"file_name": "data/audio/gio.mp3", "word": "gió", "ipa": "zɔ˧˥"}
-{"file_name": "data/audio/nang.mp3", "word": "nắng", "ipa": "năŋ˧˥"}
-{"file_name": "data/audio/duong.mp3", "word": "đường", "ipa": "ɗɨəŋ˨˩"}
-{"file_name": "data/audio/pho.mp3", "word": "phố", "ipa": "fo˧˥"}
-{"file_name": "data/audio/lam.mp3", "word": "làm", "ipa": "laːm˨˩"}
-{"file_name": "data/audio/viec.mp3", "word": "việc", "ipa": "viək˧˩ˀ"}
-{"file_name": "data/audio/hoc.mp3", "word": "học", "ipa": "hɔk˧˩ˀ"}
-{"file_name": "data/audio/sinh.mp3", "word": "sinh", "ipa": "siɲ˧˧"}
-{"file_name": "data/audio/thay.mp3", "word": "thầy", "ipa": "thəj˨˩"}
-{"file_name": "data/audio/co.mp3", "word": "cô", "ipa": "ko˧˧"}
-{"file_name": "data/audio/toi.mp3", "word": "tôi", "ipa": "toj˧˧"}
-{"file_name": "data/audio/anh.mp3", "word": "anh", "ipa": "aːɲ˧˧"}
-{"file_name": "data/audio/em.mp3", "word": "em", "ipa": "ɛm˧˧"}
-{"file_name": "data/audio/chi.mp3", "word": "chị", "ipa": "ci˧˩ˀ"}
-{"file_name": "data/audio/ong.mp3", "word": "ông", "ipa": "ʔoŋ˧˧"}
-{"file_name": "data/audio/ba_tone.mp3", "word": "bà", "ipa": "ɓaː˨˩"}
-{"file_name": "data/audio/nam.mp3", "word": "năm", "ipa": "năm˧˧"}
-{"file_name": "data/audio/thang.mp3", "word": "tháng", "ipa": "thaːŋ˧˥"}
-{"file_name": "data/audio/ngay.mp3", "word": "ngày", "ipa": "ŋaːj˨˩"}
-{"file_name": "data/audio/sang.mp3", "word": "sáng", "ipa": "saːŋ˧˥"}
-{"file_name": "data/audio/trua.mp3", "word": "trưa", "ipa": "ʈɨə˧˧"}
-{"file_name": "data/audio/chieu.mp3", "word": "chiều", "ipa": "ciəw˨˩"}
-{"file_name": "data/audio/toi_tone.mp3", "word": "tối", "ipa": "toj˧˥"}
-{"file_name": "data/audio/dem.mp3", "word": "đêm", "ipa": "ɗem˧˧"}
-{"file_name": "data/audio/mot.mp3", "word": "một", "ipa": "mot˧˩ˀ"}
-{"file_name": "data/audio/hai.mp3", "word": "hai", "ipa": "haːj˧˧"}
-{"file_name": "data/audio/ba_number.mp3", "word": "ba", "ipa": "ɓaː˧˧"}
-{"file_name": "data/audio/bon.mp3", "word": "bốn", "ipa": "ɓon˧˥"}
diff --git a/fix_root_paths.py b/fix_root_paths.py
new file mode 100644
index 0000000000000000000000000000000000000000..96c8266796a22995037fbe149cb8f4fedb8652ef
--- /dev/null
+++ b/fix_root_paths.py
@@ -0,0 +1,42 @@
+#!/usr/bin/env python3
+
+import json
+
+
+def fix_root_paths():
+    """Update metadata.jsonl to use root-level audio paths"""
+
+    input_file = "metadata.jsonl"
+    output_file = "metadata.jsonl"
+
+    records = []
+
+    # Read current metadata
+    with open(input_file, "r", encoding="utf-8") as f:
+        for line in f:
+            record = json.loads(line)
+
+            # Update path to remove data/ prefix
+            if record["file_name"].startswith("data/audio/"):
+                record["file_name"] = record["file_name"].replace("data/audio/", "audio/")
+
+            records.append(record)
+
+    # Write updated metadata
+    with open(output_file, "w", encoding="utf-8") as f:
+        for record in records:
+            f.write(json.dumps(record, ensure_ascii=False) + "\n")
+
+    print(f"Fixed {len(records)} audio paths in metadata.jsonl")
+    print("Updated paths to use root-level audio/ directory")
+
+    # Show first 3 records
+    print("\nFirst 3 records:")
+    for i, record in enumerate(records[:3], 1):
+        print(f"{i}. {json.dumps(record, ensure_ascii=False)}")
+
+    return records
+
+
+if __name__ == "__main__":
+    fix_root_paths()
\ No newline at end of file
diff --git a/metadata.jsonl b/metadata.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..e2a41c0ef53dd0231604f76a75868202c06d24b4
--- /dev/null
+++ b/metadata.jsonl
@@ -0,0 +1,50 @@
+{"file_name": "audio/xin_chao.mp3", "word": "xin chào", "ipa": "sin˧˧ caːw˨˩"}
+{"file_name": "audio/cam_on.mp3", "word": "cảm ơn", "ipa": "kaːm˧˩ ʔəːn˧˧"}
+{"file_name": "audio/viet_nam.mp3", "word": "Việt Nam", "ipa": "viət˧˩ˀ naːm˧˧"}
+{"file_name": "audio/nuoc.mp3", "word": "nước", "ipa": "nɨək˧˥"}
+{"file_name": "audio/an.mp3", "word": "ăn", "ipa": "ăn˧˧"}
+{"file_name": "audio/uong.mp3", "word": "uống", "ipa": "ʔuəŋ˧˥"}
+{"file_name": "audio/nha.mp3", "word": "nhà", "ipa": "ɲaː˨˩"}
+{"file_name": "audio/truong.mp3", "word": "trường", "ipa": "ʈɨəŋ˨˩"}
+{"file_name": "audio/ban.mp3", "word": "bạn", "ipa": "ɓaːn˧˩ˀ"}
+{"file_name": "audio/yeu.mp3", "word": "yêu", "ipa": "iəw˧˧"}
+{"file_name": "audio/dep.mp3", "word": "đẹp", "ipa": "ɗɛp˧˩ˀ"}
+{"file_name": "audio/sach.mp3", "word": "sách", "ipa": "saːk˧˥"}
+{"file_name": "audio/hoa.mp3", "word": "hoa", "ipa": "hwaː˧˧"}
+{"file_name": "audio/me.mp3", "word": "mẹ", "ipa": "mɛ˧˩ˀ"}
+{"file_name": "audio/ba.mp3", "word": "ba", "ipa": "ɓaː˧˧"}
+{"file_name": "audio/con.mp3", "word": "con", "ipa": "kɔn˧˧"}
+{"file_name": "audio/cho.mp3", "word": "chó", "ipa": "cɔ˧˥"}
+{"file_name": "audio/meo.mp3", "word": "mèo", "ipa": "mɛːw˨˩"}
+{"file_name": "audio/cay.mp3", "word": "cây", "ipa": "kəj˧˧"}
+{"file_name": "audio/la.mp3", "word": "lá", "ipa": "laː˧˥"}
+{"file_name": "audio/troi.mp3", "word": "trời", "ipa": "ʈəːj˨˩"}
+{"file_name": "audio/mua.mp3", "word": "mưa", "ipa": "mɨə˧˧"}
+{"file_name": "audio/gio.mp3", "word": "gió", "ipa": "zɔ˧˥"}
+{"file_name": "audio/nang.mp3", "word": "nắng", "ipa": "năŋ˧˥"}
+{"file_name": "audio/duong.mp3", "word": "đường", "ipa": "ɗɨəŋ˨˩"}
+{"file_name": "audio/pho.mp3", "word": "phố", "ipa": "fo˧˥"}
+{"file_name": "audio/lam.mp3", "word": "làm", "ipa": "laːm˨˩"}
+{"file_name": "audio/viec.mp3", "word": "việc", "ipa": "viək˧˩ˀ"}
+{"file_name": "audio/hoc.mp3", "word": "học", "ipa": "hɔk˧˩ˀ"}
+{"file_name": "audio/sinh.mp3", "word": "sinh", "ipa": "siɲ˧˧"}
+{"file_name": "audio/thay.mp3", "word": "thầy", "ipa": "thəj˨˩"}
+{"file_name": "audio/co.mp3", "word": "cô", "ipa": "ko˧˧"}
+{"file_name": "audio/toi.mp3", "word": "tôi", "ipa": "toj˧˧"}
+{"file_name": "audio/anh.mp3", "word": "anh", "ipa": "aːɲ˧˧"}
+{"file_name": "audio/em.mp3", "word": "em", "ipa": "ɛm˧˧"}
+{"file_name": "audio/chi.mp3", "word": "chị", "ipa": "ci˧˩ˀ"}
+{"file_name": "audio/ong.mp3", "word": "ông", "ipa": "ʔoŋ˧˧"}
+{"file_name": "audio/ba_tone.mp3", "word": "bà", "ipa": "ɓaː˨˩"}
+{"file_name": "audio/nam.mp3", "word": "năm", "ipa": "năm˧˧"}
+{"file_name": "audio/thang.mp3", "word": "tháng", "ipa": "thaːŋ˧˥"}
+{"file_name": "audio/ngay.mp3", "word": "ngày", "ipa": "ŋaːj˨˩"}
+{"file_name": "audio/sang.mp3", "word": "sáng", "ipa": "saːŋ˧˥"}
+{"file_name": "audio/trua.mp3", "word": "trưa", "ipa": "ʈɨə˧˧"}
+{"file_name": "audio/chieu.mp3", "word": "chiều", "ipa": "ciəw˨˩"}
+{"file_name": "audio/toi_tone.mp3", "word": "tối", "ipa": "toj˧˥"}
+{"file_name": "audio/dem.mp3", "word": "đêm", "ipa": "ɗem˧˧"}
+{"file_name": "audio/mot.mp3", "word": "một", "ipa": "mot˧˩ˀ"}
+{"file_name": "audio/hai.mp3", "word": "hai", "ipa": "haːj˧˧"}
+{"file_name": "audio/ba_number.mp3", "word": "ba", "ipa": "ɓaː˧˧"}
+{"file_name": "audio/bon.mp3", "word": "bốn", "ipa": "ɓon˧˥"}
diff --git a/test_final_structure.py b/test_final_structure.py
new file mode 100644
index 0000000000000000000000000000000000000000..38a1e9c2a9ea73bda7b2ea31c33bf3657d5851e2
--- /dev/null
+++ b/test_final_structure.py
@@ -0,0 +1,94 @@
+#!/usr/bin/env python3
+
+import json
+import os
+from pathlib import Path
+from datasets import Dataset, Audio
+
+
+def test_final_structure():
+    """Test the final dataset structure for HF compatibility"""
+
+    print("Testing final dataset structure...")
+    print("=" * 50)
+
+    # Check for required files
+    files_to_check = [
+        "metadata.jsonl",
+        "audio/xin_chao.mp3",
+        "audio/cam_on.mp3",
+        "audio/viet_nam.mp3"
+    ]
+
+    print("1. Checking file structure:")
+    for file_path in files_to_check:
+        if os.path.exists(file_path):
+            print(f"   ✓ {file_path}")
+        else:
+            print(f"   ✗ {file_path} - MISSING")
+            return False
+
+    # Load and validate metadata
+    print("\n2. Loading metadata.jsonl:")
+    try:
+        records = []
+        with open("metadata.jsonl", "r", encoding="utf-8") as f:
+            for line_num, line in enumerate(f, 1):
+                try:
+                    record = json.loads(line)
+                    records.append(record)
+                except json.JSONDecodeError as e:
+                    print(f"   ✗ Line {line_num}: Invalid JSON - {e}")
+                    return False
+
+        print(f"   ✓ Loaded {len(records)} records")
+
+        # Validate required fields
+        required_fields = ["file_name", "word", "ipa"]
+        for i, record in enumerate(records[:3], 1):
+            missing_fields = [field for field in required_fields if field not in record]
+            if missing_fields:
+                print(f"   ✗ Record {i} missing fields: {missing_fields}")
+                return False
+
+        print(f"   ✓ All records have required fields: {required_fields}")
+
+    except Exception as e:
+        print(f"   ✗ Error loading metadata: {e}")
+        return False
+
+    # Test with datasets library
+    print("\n3. Testing with datasets library:")
+    try:
+        dataset = Dataset.from_list(records)
+        dataset = dataset.cast_column("file_name", Audio(decode=False))
+
+        print(f"   ✓ Dataset created successfully")
+        print(f"   ✓ Features: {dataset.features}")
+        print(f"   ✓ Number of examples: {len(dataset)}")
+
+        # Test first example
+        first_example = dataset[0]
+        print(f"   ✓ First example loaded: {first_example['word']} ({first_example['ipa']})")
+
+    except Exception as e:
+        print(f"   ✗ Error with datasets library: {e}")
+        return False
+
+    print("\n4. Summary:")
+    print("   ✓ Files in root directory")
+    print("   ✓ metadata.jsonl with file_name column")
+    print("   ✓ Audio files in audio/ subdirectory")
+    print("   ✓ Compatible with datasets library")
+    print("   ✓ Ready for Hugging Face Dataset Viewer")
+
+    return True
+
+
+if __name__ == "__main__":
+    success = test_final_structure()
+
+    if success:
+        print("\n🎉 Dataset structure is correct and ready for upload!")
+    else:
+        print("\n❌ Dataset structure needs fixes")
\ No newline at end of file