Add files using upload-large-folder tool

README.md

@@ -0,0 +1,58 @@
+---
+license: other
+language:
+  - en
+tags:
+  - audio
+  - speech
+  - speaker
+  - evaluation
+configs:
+- config_name: default
+  data_files:
+  - split: test
+    path: data.parquet
+task_categories:
+  - audio-classification
+pretty_name: Seamless Interaction Pairs
+---
+
+# Seamless Interaction Pairs
+
+This dataset contains paired query and document audio clips for interaction-based
+speaker evaluation. Each row describes a query clip and a related document clip,
+with segment metadata and durations for analysis.
+
+## Data structure
+
+The dataset uses a single split stored in `data.parquet`.
+
+Audio files are stored under `audio/` and referenced by relative paths in the
+parquet file.
+
+### Columns
+
+- `pair_id` (string): Pair identifier.
+- `interaction` (string): Interaction/session identifier.
+- `query_audio` (Audio): Path to query audio file.
+- `doc_audio` (Audio): Path to document audio file.
+- `query_speaker` (string): Speaker identifier for the query segment.
+- `query_file_id` (string): File identifier for the query speaker segment.
+- `query_segments` (string): JSON list of query segment start/end times.
+- `doc_segments` (string): JSON list of document segments with speaker/file IDs.
+- `query_duration` (float): Duration in seconds for the query clip.
+- `doc_duration` (float): Duration in seconds for the document clip.
+
+## Usage
+
+```python
+from datasets import load_dataset
+
+ds = load_dataset("frankie137/seamless_interaction_pairs")
+print(ds["test"][0])
+```
+
+## Notes
+
+- Audio columns are configured as `Audio` features so the Hugging Face viewer
+  can render playback controls.

audio/seamless_interaction_pair00106-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e5a63563406271e105f77757673f73a5731f0a46bb152202cc57d5b0e7c612a3
+size 66476

audio/seamless_interaction_pair00109-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1077dea4cb53f3bfa206d7086f06db838a4b921d60def288eda310cfbb581132
+size 65452

audio/seamless_interaction_pair00128-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:736ef28e1b59eddd8d8006671fa447a284be2a3e54fcb475b6ea06572295e549
+size 50092

audio/seamless_interaction_pair00132-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6df11615eab449fb2780c75df5fc2d41f2c76e8866c07c049aecac6476993192
+size 97196

audio/seamless_interaction_pair00133-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:49dc45b8460fc52b9ef6151d556d5a60f55d53530e5019e36da4ddf401fb9d4a
+size 42924

audio/seamless_interaction_pair00135-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b38d863ccba2cc1f85ae3043923b2d9786257cd87b36d795c9aebaa35a8ef3a4
+size 41900

audio/seamless_interaction_pair00136-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b52c0052bbc004b3069e5e36fcb93752ff7c728576fbc7bd0129fdec4cfdacd3
+size 137134

audio/seamless_interaction_pair00139-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2da361bd0be99869cfd87178c18f2fa643eed59336fd7ca7d071c83c54199964
+size 111532

audio/seamless_interaction_pair00143-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:58d4fa893b0adf5a8187de569fc0ee2d0537698ae64e8a643f6a8235d9736221
+size 40878

audio/seamless_interaction_pair00149-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0437185bedd211c0c098f7a62d4a74837ef272d38c40c4e60bc1f9f99b65dc7c
+size 51116

audio/seamless_interaction_pair00153-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d567a329d899847ab1577436956c35bcbb6311e7b34e0719453fe2bba8f93da
+size 74668

audio/seamless_interaction_pair00155-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4ce0d8443e029bf243e152477fc5ba097541e66b8d9b6de8d02ab45036db16c0
+size 120748

audio/seamless_interaction_pair00156-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:056b4dda7d45b4ca8d3409a752834dbe2b8b6e3852bb83dae5849f8ee2d39f3b
+size 149420

audio/seamless_interaction_pair00157-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c6f05f12a62dd15338b77b4c39c15f7e39fee6e46feaeb735730080f0dd704d8
+size 32684

audio/seamless_interaction_pair00158-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21d178ef6094717f8183d7d35269155ca2e8b95beef766dccde8fc5fb0e63c9a
+size 37804

audio/seamless_interaction_pair00159-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:24c6700d0e29eb3506942117b080e1e340fd204a68750f31e131f1a82d68c636
+size 37804

audio/seamless_interaction_pair00161-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5acee450ef4075f84469b1370ef11d191aa930caf01f634a175bd022ced56beb
+size 32684

audio/seamless_interaction_pair00162-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a5d0b4d443356756f1ac4ded924e11672c05d5f243146a9349d5b08ddf7f446
+size 74668

audio/seamless_interaction_pair00163-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:86dd94c25a64534a4dc06e40a56ff0d966d9185f857becb24a4ea0ba66c1daed
+size 78764

audio/seamless_interaction_pair00164-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aee1d049e1f03b211c0d861203008af8b77477c9362e23f3c199a904cca13c76
+size 34732

audio/seamless_interaction_pair00165-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:78be4adc69885239724d3cc787b43eff11abe7a5236a5e1dc756cfdacf293b46
+size 167852

audio/seamless_interaction_pair00166-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d0eb965442b8fd8a5993fdc563dd233916f3c66eb0525ad15033b817436c0875
+size 69548

audio/seamless_interaction_pair00168-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1ddf7de8dde6df7ffe63ec2e2b5af4f7f1b5cf137221e5011b8297b9b6808ac6
+size 87980

audio/seamless_interaction_pair00169-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b5aa24e7cf5761da83a796eb3b65859d5ae10006d4d4a736da8fa711870deb5b
+size 98220

audio/seamless_interaction_pair00175-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5395fe389d65db53e5034bc171a8c5bcc9b9d139bc479b217d415be05ddde0fc
+size 130988

audio/seamless_interaction_pair00177-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f51b9bb2566c7bf98de3fc82f58bdf434130c826e56705eed32cf9c6a7c0382d
+size 55212

audio/seamless_interaction_pair00179-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aeee0b76ebe2db190e43ce4fe8a470740112af12d8d5464da90584f7dc91d895
+size 67500

audio/seamless_interaction_pair00180-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c7f4bc7299aee269567c458e224c39e8b65f6c75ac124db7a3764fe08be27a9e
+size 65452

audio/seamless_interaction_pair00181-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6eea58ff5f06c9b8348962801dcbee280553c74d567b1c9f1c5c80633cb4e602
+size 41900

audio/seamless_interaction_pair00183-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ce3e4355e3821ee3593d4f67e537ee7247731e1cf8fae6fc4a7306c2d71fc654
+size 71596

audio/seamless_interaction_pair00184-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:03f85893bf7e52d268b16677d503dbd061c477e5f4568edffabcb36978dc2c27
+size 150444

audio/seamless_interaction_pair00185-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d7bd1641cace4954b2c1d46ecc7e7fe5a1fc26ab6abaa81c6d122abca6d260f0
+size 43948

audio/seamless_interaction_pair00186-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9807ec9f36715a3d6a1a51bd822f9c15015740045446069741084c3c17032003
+size 35758

audio/seamless_interaction_pair00187-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c03e84cd31ef098cacb1c81f88b1785808cae8eba8754df2eb0114db62d21c55
+size 105388

audio/seamless_interaction_pair00189-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cad1e1e0a98a39a98eaca313b08d53ed6bf75611335d2046ccfb4df683e50d45
+size 136108

audio/seamless_interaction_pair00192-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b46cab6a527a30d8aa1ac8470140a71ebdcae8dfd0f619438ce28f448d1aa685
+size 269228

audio/seamless_interaction_pair00194-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b1564ccd686a9a057bd407887462135512f8467edd7596312493cb401d935d2
+size 47022

audio/seamless_interaction_pair00195-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a83a3ec594b89e014644f0d0f0dfd2deccf8f199c0414ebd0a75827962c68c90
+size 67500

audio/seamless_interaction_pair00196-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a450a59f5c6c21dd1febd27357e635bea1beb1ac952c2d7d9300f38c55591018
+size 33708

audio/seamless_interaction_pair00197-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:515b6bdb0da8b6e058252af2f9a20fec2c51bb522d2238513e93528d4f8f8fc6
+size 123822

audio/seamless_interaction_pair00198-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eafa9e2e7f0b5eef8f6de3135c7bc529f4106e409badb329feb28edb7710b0a2
+size 181164

audio/seamless_interaction_pair00199-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47ed558a4c141403eb21601086fe70ba43905fec7bc4c25e1d3473212faad9d7
+size 94124

audio/seamless_interaction_pair00200-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ff1f4c6cdeb856fbd6a42ad1d5d0746ba24be377f352f36962770a32ca2a9bb
+size 100268

audio/seamless_interaction_pair00201-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:02976f11b8ca068e66cf18754d2c0a3d26897de3343c0d85c717a09d60dda1f6
+size 38828

audio/seamless_interaction_pair00202-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac912a1a79c501b1ad8002be061ce76c9db5177b5bc0de300c6101b226db1d0a
+size 35756

audio/seamless_interaction_pair00204-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f5b6c8aa342eaf788100afda234590a40b817f1e4ade23691215261121ed2712
+size 85932

audio/seamless_interaction_pair00207-query.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ce796ce39e5332fa53077805072cf8fde2ba129edd4c363df6bd5deb70a11e2f
+size 102316

convert.py

@@ -0,0 +1,115 @@
+import pandas as pd
+import os
+import torch
+import torchaudio
+import soundfile as sf
+from tqdm import tqdm
+
+target_sample_rate = 16000
+
+df = pd.read_parquet('data.parquet')
+
+# Rename columns
+df = df.rename(columns={
+    'query_audio': 'query_audio_path',
+    'doc_audio': 'document_audio_path',
+    'query_duration': 'query_audio_duration',
+    'doc_duration': 'document_audio_duration',
+})
+
+# Get base directory (where convert.py is located)
+base_dir = os.path.dirname(os.path.abspath(__file__)) if '__file__' in globals() else os.getcwd()
+
+def check_and_resample_audio(audio_path, base_dir):
+    """
+    检查音频文件的采样率，如果不是16kHz则重采样
+    
+    Args:
+        audio_path: 音频文件的相对路径（在parquet中存储的路径）
+        base_dir: 基础目录路径
+    
+    Returns:
+        (actual_sample_rate, needs_resample): 实际采样率和是否需要重采样
+    """
+    if pd.isna(audio_path) or not audio_path or not audio_path.strip():
+        return None, False
+    
+    # 构建完整路径
+    full_path = os.path.join(base_dir, audio_path)
+    
+    if not os.path.exists(full_path):
+        print(f"警告: 音频文件不存在: {full_path}")
+        return None, False
+    
+    try:
+        # 读取音频信息
+        info = sf.info(full_path)
+        actual_sample_rate = info.samplerate
+        
+        # 检查是否需要重采样
+        if actual_sample_rate != target_sample_rate:
+            # 需要重采样
+            print(f"重采样: {audio_path} ({actual_sample_rate}Hz -> {target_sample_rate}Hz)")
+            
+            # 加载音频
+            waveform, sample_rate = torchaudio.load(full_path)
+            
+            # 转换为单声道（如果是立体声）
+            if waveform.shape[0] > 1:
+                waveform = torch.mean(waveform, dim=0, keepdim=True)
+            
+            # 重采样
+            resampler = torchaudio.transforms.Resample(sample_rate, target_sample_rate)
+            waveform_resampled = resampler(waveform)
+            
+            # 转换为numpy并确保在[-1, 1]范围内
+            waveform_np = waveform_resampled.squeeze().numpy()
+            waveform_np = waveform_np.clip(-1.0, 1.0)
+            
+            # 保存为16kHz, 16-bit WAV
+            sf.write(full_path, waveform_np, target_sample_rate, subtype='PCM_16')
+            
+            return target_sample_rate, True
+        else:
+            # 采样率正确，不需要重采样
+            return actual_sample_rate, False
+            
+    except Exception as e:
+        print(f"错误: 处理音频文件 {full_path} 时出错: {e}")
+        return None, False
+
+# 处理query音频文件
+print("检查并重采样query音频文件...")
+query_sample_rates = []
+query_resampled_count = 0
+
+for idx, audio_path in enumerate(tqdm(df['query_audio_path'], desc="处理query音频")):
+    sample_rate, was_resampled = check_and_resample_audio(audio_path, base_dir)
+    query_sample_rates.append(sample_rate if sample_rate else target_sample_rate)
+    if was_resampled:
+        query_resampled_count += 1
+
+df['query_audio_sample_rate'] = query_sample_rates
+
+# 处理document音频文件
+print("检查并重采样document音频文件...")
+doc_sample_rates = []
+doc_resampled_count = 0
+
+for idx, audio_path in enumerate(tqdm(df['document_audio_path'], desc="处理document音频")):
+    sample_rate, was_resampled = check_and_resample_audio(audio_path, base_dir)
+    doc_sample_rates.append(sample_rate if sample_rate else target_sample_rate)
+    if was_resampled:
+        doc_resampled_count += 1
+
+df['document_audio_sample_rate'] = doc_sample_rates
+
+# 保存更新后的parquet
+df.to_parquet('data.parquet', index=False)
+
+print(f'\n完成!')
+print(f'Query音频: 重采样了 {query_resampled_count}/{len(df)} 个文件')
+print(f'Document音频: 重采样了 {doc_resampled_count}/{len(df)} 个文件')
+print(f'新列: {df.columns.tolist()}')
+print(f'形状: {df.shape}')
+print(df.head(2))