Publish final Bagpiper-TTS SFT dataset card

Updates direct Parquet loading, exact final counts, no-system-prompt schema, transformations, limitations, and release status. No data files in this commit.

README.md

@@ -12,13 +12,14 @@ tags:
 - text-to-speech
 - multimodal
 - parquet
+size_categories:
+- 100K<n<1M
 ---
 
 # Bagpiper-TTS SFT Data
 
-> **Release status:** the repository currently contains documentation only.
-> The local Parquet release candidate is complete and has passed validation,
-> but no training data has been uploaded. Data upload awaits owner approval.
+> **Release status:** the validated Parquet release is being uploaded. The
+> homepage and metadata may appear before every large shard is committed.
 
 Bagpiper-TTS SFT Data supports
 [Bagpiper-TTS](https://bagpipertts.github.io/bagpiper_tts_demo/), a universal
@@ -40,18 +41,23 @@ The release is organized into the six applications used by the paper:
 Individual source corpora remain provenance fields and do not create extra
 public configurations.
 
-## Planned usage
+## Loading the data
 
-After the data release:
+Named Hugging Face configurations are deferred. Load an application with its
+explicit Parquet glob; streaming is recommended:
 
 ```python
 from datasets import load_dataset
 
 dataset = load_dataset(
-    "espnet/Bagpiper_TTS_SFT_Data",
-    "intent_to_speech",
+    "parquet",
+    data_files={"train": "hf://datasets/espnet/Bagpiper_TTS_SFT_Data/intent_to_speech/*.parquet"},
     split="train",
+    streaming=True,
 )
+
+example = next(iter(dataset))
+audio_bytes = example["audio"]["bytes"]
 ```
 
 Every example will be self-contained: its request, planning/caption text, and
@@ -77,9 +83,21 @@ three-part planning trace, and applies consistency filtering. The paper uses
 Qwen3-235B-A22B-Instruct-FP8 as its primary text processor and optionally uses
 Gemini audio validation.
 
-The validated local candidate contains 738,123 rows in six configurations. It
-is split into 1,151 Parquet shards totaling 252,981,076,979 bytes. These files
-remain local pending approval.
+## Release statistics
+
+| Application partition | Rows | Shards | Parquet bytes | Embedded audio bytes |
+|---|---:|---:|---:|---:|
+| `classical_tts` | 235,279 | 241 | 52,899,149,925 | 57,191,194,377 |
+| `multi_talker` | 64,659 | 170 | 38,632,502,383 | 44,989,933,724 |
+| `intent_to_speech` | 153,568 | 325 | 68,640,618,698 | 85,622,064,832 |
+| `role_play` | 47,116 | 107 | 22,456,138,107 | 28,130,403,188 |
+| `svs` | 101,887 | 120 | 29,503,615,421 | 31,274,125,012 |
+| `general_purpose` | 135,614 | 188 | 40,849,052,445 | 48,322,109,074 |
+| **Total** | **738,123** | **1,151** | **252,981,076,979** | **295,529,830,207** |
+
+`classical_tts` contains 17,656 LibriTTS-R clean-100, 60,657 clean-360,
+103,267 other-500, 40,793 Genshin, and 12,906 Star Rail rows. Intended reuse
+of valid audio across rows is retained.
 
 ## Intended use and non-use
 
@@ -98,6 +116,8 @@ release should not be described as a voice-cloning dataset or system.
   not itself redistribution permission.
 - The dataset may contain synthetic or transformed audio and machine-generated
   text rather than expert annotation.
+- The corpus is not exhaustively moderated for personal information, unsafe
+  content, offensive language, or copyrighted text/lyrics.
 
 ## Related resources