Hugging Face's logo

Datasets:
Dongchao
/
tmp_data

Dataset card Files
xet
Community
Dongchao commited on
Commit
b439acc
·
verified ·
1 Parent(s): f54f880

Upload 231 files

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +2 -0
  2. music/.DS_Store +0 -0
  3. music/get_scp.py +93 -0
  4. music/test/16splits/hugg_tokens.1.pt +3 -0
  5. music/test/16splits/hugg_tokens.10.pt +3 -0
  6. music/test/16splits/hugg_tokens.11.pt +3 -0
  7. music/test/16splits/hugg_tokens.12.pt +3 -0
  8. music/test/16splits/hugg_tokens.13.pt +3 -0
  9. music/test/16splits/hugg_tokens.14.pt +3 -0
  10. music/test/16splits/hugg_tokens.15.pt +3 -0
  11. music/test/16splits/hugg_tokens.16.pt +3 -0
  12. music/test/16splits/hugg_tokens.2.pt +3 -0
  13. music/test/16splits/hugg_tokens.3.pt +3 -0
  14. music/test/16splits/hugg_tokens.4.pt +3 -0
  15. music/test/16splits/hugg_tokens.5.pt +3 -0
  16. music/test/16splits/hugg_tokens.6.pt +3 -0
  17. music/test/16splits/hugg_tokens.7.pt +3 -0
  18. music/test/16splits/hugg_tokens.8.pt +3 -0
  19. music/test/16splits/hugg_tokens.9.pt +3 -0
  20. music/test/16splits/log/audio_tokenizer.1.log +0 -0
  21. music/test/16splits/log/audio_tokenizer.10.log +0 -0
  22. music/test/16splits/log/audio_tokenizer.11.log +0 -0
  23. music/test/16splits/log/audio_tokenizer.12.log +0 -0
  24. music/test/16splits/log/audio_tokenizer.13.log +0 -0
  25. music/test/16splits/log/audio_tokenizer.14.log +0 -0
  26. music/test/16splits/log/audio_tokenizer.15.log +0 -0
  27. music/test/16splits/log/audio_tokenizer.16.log +0 -0
  28. music/test/16splits/log/audio_tokenizer.2.log +0 -0
  29. music/test/16splits/log/audio_tokenizer.3.log +0 -0
  30. music/test/16splits/log/audio_tokenizer.4.log +0 -0
  31. music/test/16splits/log/audio_tokenizer.5.log +0 -0
  32. music/test/16splits/log/audio_tokenizer.6.log +0 -0
  33. music/test/16splits/log/audio_tokenizer.7.log +0 -0
  34. music/test/16splits/log/audio_tokenizer.8.log +0 -0
  35. music/test/16splits/log/audio_tokenizer.9.log +0 -0
  36. music/test/16splits/log/audio_tokenizer_higg.1.log +37 -0
  37. music/test/16splits/log/audio_tokenizer_higg.10.log +37 -0
  38. music/test/16splits/log/audio_tokenizer_higg.11.log +37 -0
  39. music/test/16splits/log/audio_tokenizer_higg.12.log +37 -0
  40. music/test/16splits/log/audio_tokenizer_higg.13.log +37 -0
  41. music/test/16splits/log/audio_tokenizer_higg.14.log +37 -0
  42. music/test/16splits/log/audio_tokenizer_higg.15.log +37 -0
  43. music/test/16splits/log/audio_tokenizer_higg.16.log +37 -0
  44. music/test/16splits/log/audio_tokenizer_higg.2.log +37 -0
  45. music/test/16splits/log/audio_tokenizer_higg.3.log +37 -0
  46. music/test/16splits/log/audio_tokenizer_higg.4.log +37 -0
  47. music/test/16splits/log/audio_tokenizer_higg.5.log +37 -0
  48. music/test/16splits/log/audio_tokenizer_higg.6.log +37 -0
  49. music/test/16splits/log/audio_tokenizer_higg.7.log +37 -0
  50. music/test/16splits/log/audio_tokenizer_higg.8.log +37 -0
.gitattributes CHANGED
@@ -57,3 +57,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
57
  # Video files - compressed
58
  *.mp4 filter=lfs diff=lfs merge=lfs -text
59
  *.webm filter=lfs diff=lfs merge=lfs -text
 
 
 
57
  # Video files - compressed
58
  *.mp4 filter=lfs diff=lfs merge=lfs -text
59
  *.webm filter=lfs diff=lfs merge=lfs -text
60
+ music/train/wav.scp filter=lfs diff=lfs merge=lfs -text
61
+ music/train/wav.scp.shuf filter=lfs diff=lfs merge=lfs -text
music/.DS_Store ADDED
Binary file (6.15 kB). View file
 
music/get_scp.py ADDED
@@ -0,0 +1,93 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # import os
2
+ # import json
3
+ # from pathlib import Path
4
+ # import argparse
5
+
6
+ # def generate_wav_scp(folder_path, output_file="wav.scp"):
7
+ # """
8
+ # 生成wav.scp文件，格式: name wav_path
9
+
10
+ # 参数:
11
+ # folder_path: 包含子文件夹的根目录（如QQ_Music_Sliced_New）
12
+ # output_file: 输出的wav.scp文件名
13
+ # """
14
+ # folder_path = Path(folder_path)
15
+
16
+ # if not folder_path.exists():
17
+ # print(f"错误: 文件夹 {folder_path} 不存在")
18
+ # return
19
+
20
+ # # 查找所有子文件夹（如0_12）
21
+ # subfolders = [f for f in folder_path.iterdir() if f.is_dir()]
22
+
23
+ # if not subfolders:
24
+ # print(f"在 {folder_path} 中未找到子文件夹")
25
+ # return
26
+
27
+ # print(f"找到 {len(subfolders)} 个子文件夹")
28
+
29
+ # wav_entries = []
30
+
31
+ # for subfolder in subfolders:
32
+ # print(f"处理子文件夹: {subfolder.name}")
33
+
34
+ # # 查找所有.mp3文件
35
+ # mp3_files = list(subfolder.glob("segment_*.mp3"))
36
+
37
+ # if not mp3_files:
38
+ # print(f" 在 {subfolder.name} 中未找到segment_*.mp3文件")
39
+ # continue
40
+
41
+ # # 按数字排序（segment_1.mp3, segment_2.mp3, ...）
42
+ # mp3_files.sort(key=lambda x: int(x.stem.split('_')[1]))
43
+
44
+ # for mp3_file in mp3_files:
45
+ # # 生成唯一名称：子文件夹名_音频文件名（不含扩展名）
46
+ # name = f"{subfolder.name}_{mp3_file.stem}"
47
+ # wav_path = mp3_file.resolve() # 获取绝对路径
48
+
49
+ # wav_entries.append((name, str(wav_path)))
50
+ # print(f" 添加: {name} -> {mp3_file.name}")
51
+
52
+ # if not wav_entries:
53
+ # print("未找到任何音频文件")
54
+ # return
55
+
56
+ # # 写入wav.scp文件
57
+ # with open(output_file, 'w', encoding='utf-8') as f:
58
+ # for name, wav_path in wav_entries:
59
+ # f.write(f"{name} {wav_path}\n")
60
+
61
+ # print(f"\n✅ 成功生成 {len(wav_entries)} 个条目")
62
+ # print(f"📁 输出文件: {output_file}")
63
+
64
+ # def main():
65
+ # parser = argparse.ArgumentParser(description="生成wav.scp文件")
66
+ # parser.add_argument("folder_path", help="包含子文件夹的根目录路径")
67
+ # parser.add_argument("--output", "-o", default="wav.scp", help="输出文件名")
68
+ # parser.add_argument("--detailed", "-d", action="store_true", help="生成详细的数据映射")
69
+
70
+ # args = parser.parse_args()
71
+
72
+ # if args.detailed:
73
+ # generate_detailed_mapping(args.folder_path)
74
+ # else:
75
+ # generate_wav_scp(args.folder_path, args.output)
76
+
77
+ # if __name__ == "__main__":
78
+ # # 直接使用示例
79
+ # folder_path = "/turing_music_fs/music_data/ydc/QQ_Music_Sliced_New" # 修改为你的实际路径
80
+
81
+ # # 简单版本：只生成wav.scp
82
+ # generate_wav_scp(folder_path)
83
+
84
+ import os
85
+ import glob
86
+
87
+ names = glob.glob("/turing_music_fs/music_data/ydc/speech_data/codec/ours_test/music_200/*.wav")
88
+
89
+ f = open('/turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/wav.scp', 'w')
90
+
91
+ for name in names:
92
+ bs_name = os.path.basename(name)
93
+ f.write(bs_name+' '+name+'\n')
music/test/16splits/hugg_tokens.1.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8c326e40492e15b0974c982817dcacd60ff0ccec1f380603bcee7ee6f06c6982
3
+ size 56213
music/test/16splits/hugg_tokens.10.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:573c15a09092e40403b610efdd7cbf8b0a07408a9e8c670b1f27fbc61081abcf
3
+ size 52002
music/test/16splits/hugg_tokens.11.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ed227d95bc7df4e0a2458e8b4700fe6b69d5da0911fc07a9507c16699a5a147a
3
+ size 51938
music/test/16splits/hugg_tokens.12.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1194b9305230aa3942dac34fa030647e548a323a7d7b3fcd1cc7850437add7ea
3
+ size 52002
music/test/16splits/hugg_tokens.13.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cff8a8b3d1a7735b7e3c827299beb849e77a6fea5b6418745ef9610ec41768d4
3
+ size 51938
music/test/16splits/hugg_tokens.14.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:34637bf2c5f9aa87250c005f9b2cc4131a746e82018407ab7b3e708eb6670067
3
+ size 51938
music/test/16splits/hugg_tokens.15.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dfec13aafc913594e9253db5b3460cd739e70a3316f011589a2f2b4ac7d04706
3
+ size 52002
music/test/16splits/hugg_tokens.16.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:33c75dabbf4f7d429c8f3a3a97f7e8bb63e55b73ca965cf4fd6bc24cd13dbc6e
3
+ size 52002
music/test/16splits/hugg_tokens.2.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:744cb558024b6a301104e14c495ecc95791304e4f6e3a3cf03f248c721d5b262
3
+ size 56213
music/test/16splits/hugg_tokens.3.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:984030b2ccb236c6be982c4eaf1bac96abf474ecc41b6a02bbd4877e75fbbc86
3
+ size 56213
music/test/16splits/hugg_tokens.4.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e0332c4f5f8b1a6fab94dc31487a14f17ebb2cc63f040787d6e0ae6124c76d0d
3
+ size 56213
music/test/16splits/hugg_tokens.5.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9fc6df9f6dd3564fc9ecaa6817a523391c0f715b35f47432dacab30b485b0ecd
3
+ size 56213
music/test/16splits/hugg_tokens.6.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4d5e2aaf49839da87f9f7ef5dd076307a1e9e7e42ec5b6a70dfe6641a1919354
3
+ size 56213
music/test/16splits/hugg_tokens.7.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:459b822ca192a6f2bc8ce549aa09389be0feb2b499a30db8952bb4fde3b811a4
3
+ size 56213
music/test/16splits/hugg_tokens.8.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5313e7af6983685d33ccb51aacbfebaf6c59c5db330b1911d586880ebfa1c927
3
+ size 56213
music/test/16splits/hugg_tokens.9.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0547c449d3ffb55967a993c91305b8565e61a2de56cf5dbc29f9b4f762abb97d
3
+ size 51986
music/test/16splits/log/audio_tokenizer.1.log ADDED
The diff for this file is too large to render. See raw diff
 
music/test/16splits/log/audio_tokenizer.10.log ADDED
The diff for this file is too large to render. See raw diff
 
music/test/16splits/log/audio_tokenizer.11.log ADDED
The diff for this file is too large to render. See raw diff
 
music/test/16splits/log/audio_tokenizer.12.log ADDED
The diff for this file is too large to render. See raw diff
 
music/test/16splits/log/audio_tokenizer.13.log ADDED
The diff for this file is too large to render. See raw diff
 
music/test/16splits/log/audio_tokenizer.14.log ADDED
The diff for this file is too large to render. See raw diff
 
music/test/16splits/log/audio_tokenizer.15.log ADDED
The diff for this file is too large to render. See raw diff
 
music/test/16splits/log/audio_tokenizer.16.log ADDED
The diff for this file is too large to render. See raw diff
 
music/test/16splits/log/audio_tokenizer.2.log ADDED
The diff for this file is too large to render. See raw diff
 
music/test/16splits/log/audio_tokenizer.3.log ADDED
The diff for this file is too large to render. See raw diff
 
music/test/16splits/log/audio_tokenizer.4.log ADDED
The diff for this file is too large to render. See raw diff
 
music/test/16splits/log/audio_tokenizer.5.log ADDED
The diff for this file is too large to render. See raw diff
 
music/test/16splits/log/audio_tokenizer.6.log ADDED
The diff for this file is too large to render. See raw diff
 
music/test/16splits/log/audio_tokenizer.7.log ADDED
The diff for this file is too large to render. See raw diff
 
music/test/16splits/log/audio_tokenizer.8.log ADDED
The diff for this file is too large to render. See raw diff
 
music/test/16splits/log/audio_tokenizer.9.log ADDED
The diff for this file is too large to render. See raw diff
 
music/test/16splits/log/audio_tokenizer_higg.1.log ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # python egs/pretraining/data_scripts/offline_tokenization_codec.py --rank 1 --input-file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/wav.1.scp --output_file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.1.pt --model_path /turing_music_fs/music_data/ydc/exp2/reason_ckpt/ReasoningCodec.checkpoint --train_config /turing_music_fs/music_data/ydc/exp2/reason_ckpt/infer_config.yaml --tokenizer hugg
2
+ # Started at Tue Nov 11 07:47:12 UTC 2025
3
+ #
4
+ 2025-11-11 07:47:14,313 INFO [offline_tokenization_codec.py:44] max gpu 8
5
+ 2025-11-11 07:47:14,313 INFO [offline_tokenization_codec.py:47] Using device: cuda:0
6
+ /root/miniconda3/envs/uniaudio2/lib/python3.10/site-packages/torch/nn/utils/weight_norm.py:134: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.
7
+ WeightNorm.apply(module, name, dim)
8
+ 2025-11-11 07:47:24,125 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
9
+ 2025-11-11 07:47:24,127 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
10
+ 2025-11-11 07:47:24,405 DEBUG [auth.py:50] Trying paths: ['/root/.docker/config.json', '/root/.dockercfg']
11
+ 2025-11-11 07:47:24,405 DEBUG [auth.py:57] No config file found
12
+ 2025-11-11 07:47:24,887 DEBUG [__init__.py:44] Skipping import of cpp extensions
13
+ [2025-11-11 07:47:25,437] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
14
+ 2025-11-11 07:47:25,515 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpaodejrnv/test.c -o /tmp/tmpaodejrnv/test.o
15
+ 2025-11-11 07:47:25,531 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpaodejrnv/test.o -laio -o /tmp/tmpaodejrnv/a.out
16
+ 2025-11-11 07:47:26,117 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpcfcqrtjp/test.c -o /tmp/tmpcfcqrtjp/test.o
17
+ 2025-11-11 07:47:26,132 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpcfcqrtjp/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpcfcqrtjp/a.out
18
+ /turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single/tools/tokenizer/higgCodec/audio_processing/higgs_audio_tokenizer.py:325: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
19
+ parameter_dict = torch.load(model_path, map_location=device)
20
+ 2025-11-11 07:47:30,294 INFO [offline_tokenization_codec.py:73] tokenizer built
21
+ 2025-11-11 07:47:30,295 INFO [offline_tokenization_codec.py:85] [Rank 0] Checkpoint path is: /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.1.pt.checkpoint_rec.pth
22
+ 2025-11-11 07:47:30,297 INFO [offline_tokenization_codec.py:102] [Rank 0] No checkpoint found. Starting from scratch.
23
+
24
+ audio_signal = torch.tensor(raw_audio)
25
+
26
+ 2025-11-11 07:47:32,138 INFO [offline_tokenization_codec.py:170] [Rank 0] Processing complete. Saving final data...
27
+ 2025-11-11 07:47:32,699 INFO [offline_tokenization_codec.py:188] [Rank 0] Final GPU cache cleanup completed
28
+ 2025-11-11 07:47:32,709 DEBUG [_api.py:331] Attempting to acquire lock 140165284418832 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
29
+ 2025-11-11 07:47:32,709 DEBUG [_api.py:334] Lock 140165284418832 acquired on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
30
+ 2025-11-11 07:47:32,709 DEBUG [_api.py:364] Attempting to release lock 140165284418832 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
31
+ 2025-11-11 07:47:32,709 DEBUG [_api.py:367] Lock 140165284418832 released on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
32
+ 2025-11-11 07:47:32,712 DEBUG [_api.py:331] Attempting to acquire lock 140165284426080 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
33
+ 2025-11-11 07:47:32,712 DEBUG [_api.py:334] Lock 140165284426080 acquired on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
34
+ 2025-11-11 07:47:32,712 DEBUG [_api.py:364] Attempting to release lock 140165284426080 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
35
+ 2025-11-11 07:47:32,712 DEBUG [_api.py:367] Lock 140165284426080 released on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
36
+ # Accounting: time=22 threads=1
37
+ # Ended (code 0) at Tue Nov 11 07:47:34 UTC 2025, elapsed time 22 seconds
music/test/16splits/log/audio_tokenizer_higg.10.log ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # python egs/pretraining/data_scripts/offline_tokenization_codec.py --rank 10 --input-file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/wav.10.scp --output_file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.10.pt --model_path /turing_music_fs/music_data/ydc/exp2/reason_ckpt/ReasoningCodec.checkpoint --train_config /turing_music_fs/music_data/ydc/exp2/reason_ckpt/infer_config.yaml --tokenizer hugg
2
+ # Started at Tue Nov 11 07:47:12 UTC 2025
3
+ #
4
+ 2025-11-11 07:47:14,234 INFO [offline_tokenization_codec.py:44] max gpu 8
5
+ 2025-11-11 07:47:14,234 INFO [offline_tokenization_codec.py:47] Using device: cuda:1
6
+ /root/miniconda3/envs/uniaudio2/lib/python3.10/site-packages/torch/nn/utils/weight_norm.py:134: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.
7
+ WeightNorm.apply(module, name, dim)
8
+ 2025-11-11 07:47:24,012 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
9
+ 2025-11-11 07:47:24,014 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
10
+ 2025-11-11 07:47:24,282 DEBUG [auth.py:50] Trying paths: ['/root/.docker/config.json', '/root/.dockercfg']
11
+ 2025-11-11 07:47:24,282 DEBUG [auth.py:57] No config file found
12
+ 2025-11-11 07:47:24,757 DEBUG [__init__.py:44] Skipping import of cpp extensions
13
+ [2025-11-11 07:47:25,240] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
14
+ 2025-11-11 07:47:25,317 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmp4pqc_7il/test.c -o /tmp/tmp4pqc_7il/test.o
15
+ 2025-11-11 07:47:25,333 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmp4pqc_7il/test.o -laio -o /tmp/tmp4pqc_7il/a.out
16
+ 2025-11-11 07:47:25,865 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpbhul5ff_/test.c -o /tmp/tmpbhul5ff_/test.o
17
+ 2025-11-11 07:47:25,880 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpbhul5ff_/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpbhul5ff_/a.out
18
+ /turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single/tools/tokenizer/higgCodec/audio_processing/higgs_audio_tokenizer.py:325: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
19
+ parameter_dict = torch.load(model_path, map_location=device)
20
+ 2025-11-11 07:47:30,348 INFO [offline_tokenization_codec.py:73] tokenizer built
21
+ 2025-11-11 07:47:30,348 INFO [offline_tokenization_codec.py:85] [Rank 1] Checkpoint path is: /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.10.pt.checkpoint_rec.pth
22
+ 2025-11-11 07:47:30,349 INFO [offline_tokenization_codec.py:102] [Rank 1] No checkpoint found. Starting from scratch.
23
+
24
+ audio_signal = torch.tensor(raw_audio)
25
+
26
+ 2025-11-11 07:47:34,799 INFO [offline_tokenization_codec.py:170] [Rank 1] Processing complete. Saving final data...
27
+ 2025-11-11 07:47:34,804 INFO [offline_tokenization_codec.py:188] [Rank 1] Final GPU cache cleanup completed
28
+ 2025-11-11 07:47:34,814 DEBUG [_api.py:331] Attempting to acquire lock 139863785198864 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
29
+ 2025-11-11 07:47:34,814 DEBUG [_api.py:334] Lock 139863785198864 acquired on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
30
+ 2025-11-11 07:47:34,814 DEBUG [_api.py:364] Attempting to release lock 139863785198864 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
31
+ 2025-11-11 07:47:34,814 DEBUG [_api.py:367] Lock 139863785198864 released on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
32
+ 2025-11-11 07:47:34,817 DEBUG [_api.py:331] Attempting to acquire lock 139863785206112 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
33
+ 2025-11-11 07:47:34,817 DEBUG [_api.py:334] Lock 139863785206112 acquired on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
34
+ 2025-11-11 07:47:34,817 DEBUG [_api.py:364] Attempting to release lock 139863785206112 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
35
+ 2025-11-11 07:47:34,817 DEBUG [_api.py:367] Lock 139863785206112 released on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
36
+ # Accounting: time=25 threads=1
37
+ # Ended (code 0) at Tue Nov 11 07:47:37 UTC 2025, elapsed time 25 seconds
music/test/16splits/log/audio_tokenizer_higg.11.log ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # python egs/pretraining/data_scripts/offline_tokenization_codec.py --rank 11 --input-file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/wav.11.scp --output_file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.11.pt --model_path /turing_music_fs/music_data/ydc/exp2/reason_ckpt/ReasoningCodec.checkpoint --train_config /turing_music_fs/music_data/ydc/exp2/reason_ckpt/infer_config.yaml --tokenizer hugg
2
+ # Started at Tue Nov 11 07:47:12 UTC 2025
3
+ #
4
+ 2025-11-11 07:47:14,201 INFO [offline_tokenization_codec.py:44] max gpu 8
5
+ 2025-11-11 07:47:14,202 INFO [offline_tokenization_codec.py:47] Using device: cuda:2
6
+ /root/miniconda3/envs/uniaudio2/lib/python3.10/site-packages/torch/nn/utils/weight_norm.py:134: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.
7
+ WeightNorm.apply(module, name, dim)
8
+ 2025-11-11 07:47:23,632 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
9
+ 2025-11-11 07:47:23,634 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
10
+ 2025-11-11 07:47:23,898 DEBUG [auth.py:50] Trying paths: ['/root/.docker/config.json', '/root/.dockercfg']
11
+ 2025-11-11 07:47:23,898 DEBUG [auth.py:57] No config file found
12
+ 2025-11-11 07:47:24,378 DEBUG [__init__.py:44] Skipping import of cpp extensions
13
+ [2025-11-11 07:47:24,821] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
14
+ 2025-11-11 07:47:24,899 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpt41kbyo9/test.c -o /tmp/tmpt41kbyo9/test.o
15
+ 2025-11-11 07:47:24,915 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpt41kbyo9/test.o -laio -o /tmp/tmpt41kbyo9/a.out
16
+ 2025-11-11 07:47:25,447 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpki1dsi7v/test.c -o /tmp/tmpki1dsi7v/test.o
17
+ 2025-11-11 07:47:25,464 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpki1dsi7v/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpki1dsi7v/a.out
18
+ /turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single/tools/tokenizer/higgCodec/audio_processing/higgs_audio_tokenizer.py:325: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
19
+ parameter_dict = torch.load(model_path, map_location=device)
20
+ 2025-11-11 07:47:30,305 INFO [offline_tokenization_codec.py:73] tokenizer built
21
+ 2025-11-11 07:47:30,305 INFO [offline_tokenization_codec.py:85] [Rank 2] Checkpoint path is: /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.11.pt.checkpoint_rec.pth
22
+ 2025-11-11 07:47:30,306 INFO [offline_tokenization_codec.py:102] [Rank 2] No checkpoint found. Starting from scratch.
23
+
24
+ audio_signal = torch.tensor(raw_audio)
25
+
26
+ 2025-11-11 07:47:34,936 INFO [offline_tokenization_codec.py:170] [Rank 2] Processing complete. Saving final data...
27
+ 2025-11-11 07:47:34,941 INFO [offline_tokenization_codec.py:188] [Rank 2] Final GPU cache cleanup completed
28
+ 2025-11-11 07:47:34,953 DEBUG [_api.py:331] Attempting to acquire lock 139876916040976 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
29
+ 2025-11-11 07:47:34,953 DEBUG [_api.py:334] Lock 139876916040976 acquired on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
30
+ 2025-11-11 07:47:34,953 DEBUG [_api.py:364] Attempting to release lock 139876916040976 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
31
+ 2025-11-11 07:47:34,953 DEBUG [_api.py:367] Lock 139876916040976 released on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
32
+ 2025-11-11 07:47:34,956 DEBUG [_api.py:331] Attempting to acquire lock 139876916048224 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
33
+ 2025-11-11 07:47:34,956 DEBUG [_api.py:334] Lock 139876916048224 acquired on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
34
+ 2025-11-11 07:47:34,956 DEBUG [_api.py:364] Attempting to release lock 139876916048224 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
35
+ 2025-11-11 07:47:34,957 DEBUG [_api.py:367] Lock 139876916048224 released on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
36
+ # Accounting: time=25 threads=1
37
+ # Ended (code 0) at Tue Nov 11 07:47:37 UTC 2025, elapsed time 25 seconds
music/test/16splits/log/audio_tokenizer_higg.12.log ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # python egs/pretraining/data_scripts/offline_tokenization_codec.py --rank 12 --input-file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/wav.12.scp --output_file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.12.pt --model_path /turing_music_fs/music_data/ydc/exp2/reason_ckpt/ReasoningCodec.checkpoint --train_config /turing_music_fs/music_data/ydc/exp2/reason_ckpt/infer_config.yaml --tokenizer hugg
2
+ # Started at Tue Nov 11 07:47:12 UTC 2025
3
+ #
4
+ 2025-11-11 07:47:14,253 INFO [offline_tokenization_codec.py:44] max gpu 8
5
+ 2025-11-11 07:47:14,253 INFO [offline_tokenization_codec.py:47] Using device: cuda:3
6
+ /root/miniconda3/envs/uniaudio2/lib/python3.10/site-packages/torch/nn/utils/weight_norm.py:134: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.
7
+ WeightNorm.apply(module, name, dim)
8
+ 2025-11-11 07:47:23,766 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
9
+ 2025-11-11 07:47:23,768 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
10
+ 2025-11-11 07:47:24,039 DEBUG [auth.py:50] Trying paths: ['/root/.docker/config.json', '/root/.dockercfg']
11
+ 2025-11-11 07:47:24,039 DEBUG [auth.py:57] No config file found
12
+ 2025-11-11 07:47:24,517 DEBUG [__init__.py:44] Skipping import of cpp extensions
13
+ [2025-11-11 07:47:24,946] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
14
+ 2025-11-11 07:47:25,026 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpzgfiq8qa/test.c -o /tmp/tmpzgfiq8qa/test.o
15
+ 2025-11-11 07:47:25,042 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpzgfiq8qa/test.o -laio -o /tmp/tmpzgfiq8qa/a.out
16
+ 2025-11-11 07:47:25,574 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpwayf2fx5/test.c -o /tmp/tmpwayf2fx5/test.o
17
+ 2025-11-11 07:47:25,589 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpwayf2fx5/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpwayf2fx5/a.out
18
+ /turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single/tools/tokenizer/higgCodec/audio_processing/higgs_audio_tokenizer.py:325: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
19
+ parameter_dict = torch.load(model_path, map_location=device)
20
+ 2025-11-11 07:47:30,337 INFO [offline_tokenization_codec.py:73] tokenizer built
21
+ 2025-11-11 07:47:30,337 INFO [offline_tokenization_codec.py:85] [Rank 3] Checkpoint path is: /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.12.pt.checkpoint_rec.pth
22
+ 2025-11-11 07:47:30,338 INFO [offline_tokenization_codec.py:102] [Rank 3] No checkpoint found. Starting from scratch.
23
+
24
+ audio_signal = torch.tensor(raw_audio)
25
+
26
+ 2025-11-11 07:47:34,763 INFO [offline_tokenization_codec.py:170] [Rank 3] Processing complete. Saving final data...
27
+ 2025-11-11 07:47:34,769 INFO [offline_tokenization_codec.py:188] [Rank 3] Final GPU cache cleanup completed
28
+ 2025-11-11 07:47:34,780 DEBUG [_api.py:331] Attempting to acquire lock 139893332497680 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
29
+ 2025-11-11 07:47:34,780 DEBUG [_api.py:334] Lock 139893332497680 acquired on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
30
+ 2025-11-11 07:47:34,780 DEBUG [_api.py:364] Attempting to release lock 139893332497680 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
31
+ 2025-11-11 07:47:34,780 DEBUG [_api.py:367] Lock 139893332497680 released on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
32
+ 2025-11-11 07:47:34,783 DEBUG [_api.py:331] Attempting to acquire lock 139893332504928 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
33
+ 2025-11-11 07:47:34,783 DEBUG [_api.py:334] Lock 139893332504928 acquired on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
34
+ 2025-11-11 07:47:34,783 DEBUG [_api.py:364] Attempting to release lock 139893332504928 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
35
+ 2025-11-11 07:47:34,783 DEBUG [_api.py:367] Lock 139893332504928 released on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
36
+ # Accounting: time=26 threads=1
37
+ # Ended (code 0) at Tue Nov 11 07:47:38 UTC 2025, elapsed time 26 seconds
music/test/16splits/log/audio_tokenizer_higg.13.log ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # python egs/pretraining/data_scripts/offline_tokenization_codec.py --rank 13 --input-file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/wav.13.scp --output_file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.13.pt --model_path /turing_music_fs/music_data/ydc/exp2/reason_ckpt/ReasoningCodec.checkpoint --train_config /turing_music_fs/music_data/ydc/exp2/reason_ckpt/infer_config.yaml --tokenizer hugg
2
+ # Started at Tue Nov 11 07:47:12 UTC 2025
3
+ #
4
+ 2025-11-11 07:47:14,302 INFO [offline_tokenization_codec.py:44] max gpu 8
5
+ 2025-11-11 07:47:14,303 INFO [offline_tokenization_codec.py:47] Using device: cuda:4
6
+ /root/miniconda3/envs/uniaudio2/lib/python3.10/site-packages/torch/nn/utils/weight_norm.py:134: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.
7
+ WeightNorm.apply(module, name, dim)
8
+ 2025-11-11 07:47:23,927 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
9
+ 2025-11-11 07:47:23,929 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
10
+ 2025-11-11 07:47:24,203 DEBUG [auth.py:50] Trying paths: ['/root/.docker/config.json', '/root/.dockercfg']
11
+ 2025-11-11 07:47:24,203 DEBUG [auth.py:57] No config file found
12
+ 2025-11-11 07:47:24,687 DEBUG [__init__.py:44] Skipping import of cpp extensions
13
+ [2025-11-11 07:47:25,128] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
14
+ 2025-11-11 07:47:25,207 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmp4w5nw9t6/test.c -o /tmp/tmp4w5nw9t6/test.o
15
+ 2025-11-11 07:47:25,223 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmp4w5nw9t6/test.o -laio -o /tmp/tmp4w5nw9t6/a.out
16
+ 2025-11-11 07:47:25,767 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpimey4pft/test.c -o /tmp/tmpimey4pft/test.o
17
+ 2025-11-11 07:47:25,784 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpimey4pft/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpimey4pft/a.out
18
+ /turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single/tools/tokenizer/higgCodec/audio_processing/higgs_audio_tokenizer.py:325: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
19
+ parameter_dict = torch.load(model_path, map_location=device)
20
+ 2025-11-11 07:47:30,364 INFO [offline_tokenization_codec.py:73] tokenizer built
21
+ 2025-11-11 07:47:30,364 INFO [offline_tokenization_codec.py:85] [Rank 4] Checkpoint path is: /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.13.pt.checkpoint_rec.pth
22
+ 2025-11-11 07:47:30,364 INFO [offline_tokenization_codec.py:102] [Rank 4] No checkpoint found. Starting from scratch.
23
+
24
+ audio_signal = torch.tensor(raw_audio)
25
+
26
+ 2025-11-11 07:47:34,776 INFO [offline_tokenization_codec.py:170] [Rank 4] Processing complete. Saving final data...
27
+ 2025-11-11 07:47:34,782 INFO [offline_tokenization_codec.py:188] [Rank 4] Final GPU cache cleanup completed
28
+ 2025-11-11 07:47:34,793 DEBUG [_api.py:331] Attempting to acquire lock 140377038427408 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
29
+ 2025-11-11 07:47:34,793 DEBUG [_api.py:334] Lock 140377038427408 acquired on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
30
+ 2025-11-11 07:47:34,793 DEBUG [_api.py:364] Attempting to release lock 140377038427408 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
31
+ 2025-11-11 07:47:34,793 DEBUG [_api.py:367] Lock 140377038427408 released on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
32
+ 2025-11-11 07:47:34,796 DEBUG [_api.py:331] Attempting to acquire lock 140377038434656 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
33
+ 2025-11-11 07:47:34,796 DEBUG [_api.py:334] Lock 140377038434656 acquired on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
34
+ 2025-11-11 07:47:34,796 DEBUG [_api.py:364] Attempting to release lock 140377038434656 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
35
+ 2025-11-11 07:47:34,797 DEBUG [_api.py:367] Lock 140377038434656 released on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
36
+ # Accounting: time=26 threads=1
37
+ # Ended (code 0) at Tue Nov 11 07:47:38 UTC 2025, elapsed time 26 seconds
music/test/16splits/log/audio_tokenizer_higg.14.log ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # python egs/pretraining/data_scripts/offline_tokenization_codec.py --rank 14 --input-file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/wav.14.scp --output_file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.14.pt --model_path /turing_music_fs/music_data/ydc/exp2/reason_ckpt/ReasoningCodec.checkpoint --train_config /turing_music_fs/music_data/ydc/exp2/reason_ckpt/infer_config.yaml --tokenizer hugg
2
+ # Started at Tue Nov 11 07:47:12 UTC 2025
3
+ #
4
+ 2025-11-11 07:47:14,303 INFO [offline_tokenization_codec.py:44] max gpu 8
5
+ 2025-11-11 07:47:14,304 INFO [offline_tokenization_codec.py:47] Using device: cuda:5
6
+ /root/miniconda3/envs/uniaudio2/lib/python3.10/site-packages/torch/nn/utils/weight_norm.py:134: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.
7
+ WeightNorm.apply(module, name, dim)
8
+ 2025-11-11 07:47:24,091 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
9
+ 2025-11-11 07:47:24,093 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
10
+ 2025-11-11 07:47:24,364 DEBUG [auth.py:50] Trying paths: ['/root/.docker/config.json', '/root/.dockercfg']
11
+ 2025-11-11 07:47:24,364 DEBUG [auth.py:57] No config file found
12
+ 2025-11-11 07:47:24,845 DEBUG [__init__.py:44] Skipping import of cpp extensions
13
+ [2025-11-11 07:47:25,341] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
14
+ 2025-11-11 07:47:25,422 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpl739ohyg/test.c -o /tmp/tmpl739ohyg/test.o
15
+ 2025-11-11 07:47:25,438 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpl739ohyg/test.o -laio -o /tmp/tmpl739ohyg/a.out
16
+ 2025-11-11 07:47:25,982 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmp5t52s4ji/test.c -o /tmp/tmp5t52s4ji/test.o
17
+ 2025-11-11 07:47:25,998 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmp5t52s4ji/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp5t52s4ji/a.out
18
+ /turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single/tools/tokenizer/higgCodec/audio_processing/higgs_audio_tokenizer.py:325: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
19
+ parameter_dict = torch.load(model_path, map_location=device)
20
+ 2025-11-11 07:47:30,361 INFO [offline_tokenization_codec.py:73] tokenizer built
21
+ 2025-11-11 07:47:30,361 INFO [offline_tokenization_codec.py:85] [Rank 5] Checkpoint path is: /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.14.pt.checkpoint_rec.pth
22
+ 2025-11-11 07:47:30,361 INFO [offline_tokenization_codec.py:102] [Rank 5] No checkpoint found. Starting from scratch.
23
+
24
+ audio_signal = torch.tensor(raw_audio)
25
+
26
+ 2025-11-11 07:47:34,757 INFO [offline_tokenization_codec.py:170] [Rank 5] Processing complete. Saving final data...
27
+ 2025-11-11 07:47:34,762 INFO [offline_tokenization_codec.py:188] [Rank 5] Final GPU cache cleanup completed
28
+ 2025-11-11 07:47:34,773 DEBUG [_api.py:331] Attempting to acquire lock 140597400458512 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
29
+ 2025-11-11 07:47:34,774 DEBUG [_api.py:334] Lock 140597400458512 acquired on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
30
+ 2025-11-11 07:47:34,774 DEBUG [_api.py:364] Attempting to release lock 140597400458512 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
31
+ 2025-11-11 07:47:34,774 DEBUG [_api.py:367] Lock 140597400458512 released on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
32
+ 2025-11-11 07:47:34,777 DEBUG [_api.py:331] Attempting to acquire lock 140597400465760 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
33
+ 2025-11-11 07:47:34,777 DEBUG [_api.py:334] Lock 140597400465760 acquired on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
34
+ 2025-11-11 07:47:34,777 DEBUG [_api.py:364] Attempting to release lock 140597400465760 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
35
+ 2025-11-11 07:47:34,777 DEBUG [_api.py:367] Lock 140597400465760 released on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
36
+ # Accounting: time=26 threads=1
37
+ # Ended (code 0) at Tue Nov 11 07:47:38 UTC 2025, elapsed time 26 seconds
music/test/16splits/log/audio_tokenizer_higg.15.log ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # python egs/pretraining/data_scripts/offline_tokenization_codec.py --rank 15 --input-file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/wav.15.scp --output_file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.15.pt --model_path /turing_music_fs/music_data/ydc/exp2/reason_ckpt/ReasoningCodec.checkpoint --train_config /turing_music_fs/music_data/ydc/exp2/reason_ckpt/infer_config.yaml --tokenizer hugg
2
+ # Started at Tue Nov 11 07:47:12 UTC 2025
3
+ #
4
+ 2025-11-11 07:47:14,279 INFO [offline_tokenization_codec.py:44] max gpu 8
5
+ 2025-11-11 07:47:14,280 INFO [offline_tokenization_codec.py:47] Using device: cuda:6
6
+ /root/miniconda3/envs/uniaudio2/lib/python3.10/site-packages/torch/nn/utils/weight_norm.py:134: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.
7
+ WeightNorm.apply(module, name, dim)
8
+ 2025-11-11 07:47:24,047 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
9
+ 2025-11-11 07:47:24,049 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
10
+ 2025-11-11 07:47:24,352 DEBUG [auth.py:50] Trying paths: ['/root/.docker/config.json', '/root/.dockercfg']
11
+ 2025-11-11 07:47:24,352 DEBUG [auth.py:57] No config file found
12
+ 2025-11-11 07:47:24,838 DEBUG [__init__.py:44] Skipping import of cpp extensions
13
+ [2025-11-11 07:47:25,399] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
14
+ 2025-11-11 07:47:25,480 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpl15nc76z/test.c -o /tmp/tmpl15nc76z/test.o
15
+ 2025-11-11 07:47:25,496 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpl15nc76z/test.o -laio -o /tmp/tmpl15nc76z/a.out
16
+ 2025-11-11 07:47:26,099 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmp73sxwklc/test.c -o /tmp/tmp73sxwklc/test.o
17
+ 2025-11-11 07:47:26,113 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmp73sxwklc/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp73sxwklc/a.out
18
+ /turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single/tools/tokenizer/higgCodec/audio_processing/higgs_audio_tokenizer.py:325: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
19
+ parameter_dict = torch.load(model_path, map_location=device)
20
+ 2025-11-11 07:47:30,308 INFO [offline_tokenization_codec.py:73] tokenizer built
21
+ 2025-11-11 07:47:30,309 INFO [offline_tokenization_codec.py:85] [Rank 6] Checkpoint path is: /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.15.pt.checkpoint_rec.pth
22
+ 2025-11-11 07:47:30,309 INFO [offline_tokenization_codec.py:102] [Rank 6] No checkpoint found. Starting from scratch.
23
+
24
+ audio_signal = torch.tensor(raw_audio)
25
+
26
+ 2025-11-11 07:47:34,784 INFO [offline_tokenization_codec.py:170] [Rank 6] Processing complete. Saving final data...
27
+ 2025-11-11 07:47:34,789 INFO [offline_tokenization_codec.py:188] [Rank 6] Final GPU cache cleanup completed
28
+ 2025-11-11 07:47:34,798 DEBUG [_api.py:331] Attempting to acquire lock 140596330485008 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
29
+ 2025-11-11 07:47:34,799 DEBUG [_api.py:334] Lock 140596330485008 acquired on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
30
+ 2025-11-11 07:47:34,799 DEBUG [_api.py:364] Attempting to release lock 140596330485008 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
31
+ 2025-11-11 07:47:34,799 DEBUG [_api.py:367] Lock 140596330485008 released on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
32
+ 2025-11-11 07:47:34,802 DEBUG [_api.py:331] Attempting to acquire lock 140596330492256 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
33
+ 2025-11-11 07:47:34,802 DEBUG [_api.py:334] Lock 140596330492256 acquired on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
34
+ 2025-11-11 07:47:34,802 DEBUG [_api.py:364] Attempting to release lock 140596330492256 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
35
+ 2025-11-11 07:47:34,802 DEBUG [_api.py:367] Lock 140596330492256 released on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
36
+ # Accounting: time=25 threads=1
37
+ # Ended (code 0) at Tue Nov 11 07:47:37 UTC 2025, elapsed time 25 seconds
music/test/16splits/log/audio_tokenizer_higg.16.log ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # python egs/pretraining/data_scripts/offline_tokenization_codec.py --rank 16 --input-file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/wav.16.scp --output_file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.16.pt --model_path /turing_music_fs/music_data/ydc/exp2/reason_ckpt/ReasoningCodec.checkpoint --train_config /turing_music_fs/music_data/ydc/exp2/reason_ckpt/infer_config.yaml --tokenizer hugg
2
+ # Started at Tue Nov 11 07:47:12 UTC 2025
3
+ #
4
+ 2025-11-11 07:47:14,299 INFO [offline_tokenization_codec.py:44] max gpu 8
5
+ 2025-11-11 07:47:14,299 INFO [offline_tokenization_codec.py:47] Using device: cuda:7
6
+ /root/miniconda3/envs/uniaudio2/lib/python3.10/site-packages/torch/nn/utils/weight_norm.py:134: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.
7
+ WeightNorm.apply(module, name, dim)
8
+ 2025-11-11 07:47:24,296 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
9
+ 2025-11-11 07:47:24,298 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
10
+ 2025-11-11 07:47:24,546 DEBUG [auth.py:50] Trying paths: ['/root/.docker/config.json', '/root/.dockercfg']
11
+ 2025-11-11 07:47:24,546 DEBUG [auth.py:57] No config file found
12
+ 2025-11-11 07:47:25,029 DEBUG [__init__.py:44] Skipping import of cpp extensions
13
+ [2025-11-11 07:47:25,478] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
14
+ 2025-11-11 07:47:25,563 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpe5mhbxit/test.c -o /tmp/tmpe5mhbxit/test.o
15
+ 2025-11-11 07:47:25,579 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpe5mhbxit/test.o -laio -o /tmp/tmpe5mhbxit/a.out
16
+ 2025-11-11 07:47:26,213 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpialvtwea/test.c -o /tmp/tmpialvtwea/test.o
17
+ 2025-11-11 07:47:26,229 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpialvtwea/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpialvtwea/a.out
18
+ /turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single/tools/tokenizer/higgCodec/audio_processing/higgs_audio_tokenizer.py:325: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
19
+ parameter_dict = torch.load(model_path, map_location=device)
20
+ 2025-11-11 07:47:30,339 INFO [offline_tokenization_codec.py:73] tokenizer built
21
+ 2025-11-11 07:47:30,339 INFO [offline_tokenization_codec.py:85] [Rank 7] Checkpoint path is: /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.16.pt.checkpoint_rec.pth
22
+ 2025-11-11 07:47:30,340 INFO [offline_tokenization_codec.py:102] [Rank 7] No checkpoint found. Starting from scratch.
23
+
24
+ audio_signal = torch.tensor(raw_audio)
25
+
26
+ 2025-11-11 07:47:34,754 INFO [offline_tokenization_codec.py:170] [Rank 7] Processing complete. Saving final data...
27
+ 2025-11-11 07:47:34,760 INFO [offline_tokenization_codec.py:188] [Rank 7] Final GPU cache cleanup completed
28
+ 2025-11-11 07:47:34,770 DEBUG [_api.py:331] Attempting to acquire lock 140201250493712 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
29
+ 2025-11-11 07:47:34,770 DEBUG [_api.py:334] Lock 140201250493712 acquired on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
30
+ 2025-11-11 07:47:34,771 DEBUG [_api.py:364] Attempting to release lock 140201250493712 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
31
+ 2025-11-11 07:47:34,771 DEBUG [_api.py:367] Lock 140201250493712 released on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
32
+ 2025-11-11 07:47:34,774 DEBUG [_api.py:331] Attempting to acquire lock 140201250500960 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
33
+ 2025-11-11 07:47:34,774 DEBUG [_api.py:334] Lock 140201250500960 acquired on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
34
+ 2025-11-11 07:47:34,774 DEBUG [_api.py:364] Attempting to release lock 140201250500960 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
35
+ 2025-11-11 07:47:34,774 DEBUG [_api.py:367] Lock 140201250500960 released on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
36
+ # Accounting: time=26 threads=1
37
+ # Ended (code 0) at Tue Nov 11 07:47:38 UTC 2025, elapsed time 26 seconds
music/test/16splits/log/audio_tokenizer_higg.2.log ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # python egs/pretraining/data_scripts/offline_tokenization_codec.py --rank 2 --input-file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/wav.2.scp --output_file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.2.pt --model_path /turing_music_fs/music_data/ydc/exp2/reason_ckpt/ReasoningCodec.checkpoint --train_config /turing_music_fs/music_data/ydc/exp2/reason_ckpt/infer_config.yaml --tokenizer hugg
2
+ # Started at Tue Nov 11 07:47:12 UTC 2025
3
+ #
4
+ 2025-11-11 07:47:14,304 INFO [offline_tokenization_codec.py:44] max gpu 8
5
+ 2025-11-11 07:47:14,304 INFO [offline_tokenization_codec.py:47] Using device: cuda:1
6
+ /root/miniconda3/envs/uniaudio2/lib/python3.10/site-packages/torch/nn/utils/weight_norm.py:134: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.
7
+ WeightNorm.apply(module, name, dim)
8
+ 2025-11-11 07:47:24,401 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
9
+ 2025-11-11 07:47:24,403 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
10
+ 2025-11-11 07:47:24,664 DEBUG [auth.py:50] Trying paths: ['/root/.docker/config.json', '/root/.dockercfg']
11
+ 2025-11-11 07:47:24,664 DEBUG [auth.py:57] No config file found
12
+ 2025-11-11 07:47:25,147 DEBUG [__init__.py:44] Skipping import of cpp extensions
13
+ [2025-11-11 07:47:25,632] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
14
+ 2025-11-11 07:47:25,713 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpjr2qejqo/test.c -o /tmp/tmpjr2qejqo/test.o
15
+ 2025-11-11 07:47:25,729 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpjr2qejqo/test.o -laio -o /tmp/tmpjr2qejqo/a.out
16
+ 2025-11-11 07:47:26,289 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmp3h0bng7a/test.c -o /tmp/tmp3h0bng7a/test.o
17
+ 2025-11-11 07:47:26,304 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmp3h0bng7a/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp3h0bng7a/a.out
18
+ /turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single/tools/tokenizer/higgCodec/audio_processing/higgs_audio_tokenizer.py:325: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
19
+ parameter_dict = torch.load(model_path, map_location=device)
20
+ 2025-11-11 07:47:30,348 INFO [offline_tokenization_codec.py:73] tokenizer built
21
+ 2025-11-11 07:47:30,348 INFO [offline_tokenization_codec.py:85] [Rank 1] Checkpoint path is: /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.2.pt.checkpoint_rec.pth
22
+ 2025-11-11 07:47:30,349 INFO [offline_tokenization_codec.py:102] [Rank 1] No checkpoint found. Starting from scratch.
23
+
24
+ audio_signal = torch.tensor(raw_audio)
25
+
26
+ 2025-11-11 07:47:34,834 INFO [offline_tokenization_codec.py:170] [Rank 1] Processing complete. Saving final data...
27
+ 2025-11-11 07:47:34,839 INFO [offline_tokenization_codec.py:188] [Rank 1] Final GPU cache cleanup completed
28
+ 2025-11-11 07:47:34,851 DEBUG [_api.py:331] Attempting to acquire lock 140331473425680 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
29
+ 2025-11-11 07:47:34,851 DEBUG [_api.py:334] Lock 140331473425680 acquired on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
30
+ 2025-11-11 07:47:34,852 DEBUG [_api.py:364] Attempting to release lock 140331473425680 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
31
+ 2025-11-11 07:47:34,852 DEBUG [_api.py:367] Lock 140331473425680 released on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
32
+ 2025-11-11 07:47:34,855 DEBUG [_api.py:331] Attempting to acquire lock 140331473432928 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
33
+ 2025-11-11 07:47:34,855 DEBUG [_api.py:334] Lock 140331473432928 acquired on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
34
+ 2025-11-11 07:47:34,855 DEBUG [_api.py:364] Attempting to release lock 140331473432928 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
35
+ 2025-11-11 07:47:34,855 DEBUG [_api.py:367] Lock 140331473432928 released on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
36
+ # Accounting: time=25 threads=1
37
+ # Ended (code 0) at Tue Nov 11 07:47:37 UTC 2025, elapsed time 25 seconds
music/test/16splits/log/audio_tokenizer_higg.3.log ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # python egs/pretraining/data_scripts/offline_tokenization_codec.py --rank 3 --input-file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/wav.3.scp --output_file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.3.pt --model_path /turing_music_fs/music_data/ydc/exp2/reason_ckpt/ReasoningCodec.checkpoint --train_config /turing_music_fs/music_data/ydc/exp2/reason_ckpt/infer_config.yaml --tokenizer hugg
2
+ # Started at Tue Nov 11 07:47:12 UTC 2025
3
+ #
4
+ 2025-11-11 07:47:14,250 INFO [offline_tokenization_codec.py:44] max gpu 8
5
+ 2025-11-11 07:47:14,250 INFO [offline_tokenization_codec.py:47] Using device: cuda:2
6
+ /root/miniconda3/envs/uniaudio2/lib/python3.10/site-packages/torch/nn/utils/weight_norm.py:134: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.
7
+ WeightNorm.apply(module, name, dim)
8
+ 2025-11-11 07:47:24,270 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
9
+ 2025-11-11 07:47:24,272 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
10
+ 2025-11-11 07:47:24,524 DEBUG [auth.py:50] Trying paths: ['/root/.docker/config.json', '/root/.dockercfg']
11
+ 2025-11-11 07:47:24,524 DEBUG [auth.py:57] No config file found
12
+ 2025-11-11 07:47:25,000 DEBUG [__init__.py:44] Skipping import of cpp extensions
13
+ [2025-11-11 07:47:25,471] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
14
+ 2025-11-11 07:47:25,552 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpdd8fw2qb/test.c -o /tmp/tmpdd8fw2qb/test.o
15
+ 2025-11-11 07:47:25,567 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpdd8fw2qb/test.o -laio -o /tmp/tmpdd8fw2qb/a.out
16
+ 2025-11-11 07:47:26,101 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmp2141fwof/test.c -o /tmp/tmp2141fwof/test.o
17
+ 2025-11-11 07:47:26,115 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmp2141fwof/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp2141fwof/a.out
18
+ /turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single/tools/tokenizer/higgCodec/audio_processing/higgs_audio_tokenizer.py:325: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
19
+ parameter_dict = torch.load(model_path, map_location=device)
20
+ 2025-11-11 07:47:30,351 INFO [offline_tokenization_codec.py:73] tokenizer built
21
+ 2025-11-11 07:47:30,351 INFO [offline_tokenization_codec.py:85] [Rank 2] Checkpoint path is: /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.3.pt.checkpoint_rec.pth
22
+ 2025-11-11 07:47:30,351 INFO [offline_tokenization_codec.py:102] [Rank 2] No checkpoint found. Starting from scratch.
23
+
24
+ audio_signal = torch.tensor(raw_audio)
25
+
26
+ 2025-11-11 07:47:34,767 INFO [offline_tokenization_codec.py:170] [Rank 2] Processing complete. Saving final data...
27
+ 2025-11-11 07:47:34,772 INFO [offline_tokenization_codec.py:188] [Rank 2] Final GPU cache cleanup completed
28
+ 2025-11-11 07:47:34,783 DEBUG [_api.py:331] Attempting to acquire lock 140135361254672 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
29
+ 2025-11-11 07:47:34,783 DEBUG [_api.py:334] Lock 140135361254672 acquired on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
30
+ 2025-11-11 07:47:34,783 DEBUG [_api.py:364] Attempting to release lock 140135361254672 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
31
+ 2025-11-11 07:47:34,783 DEBUG [_api.py:367] Lock 140135361254672 released on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
32
+ 2025-11-11 07:47:34,786 DEBUG [_api.py:331] Attempting to acquire lock 140135361261920 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
33
+ 2025-11-11 07:47:34,786 DEBUG [_api.py:334] Lock 140135361261920 acquired on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
34
+ 2025-11-11 07:47:34,786 DEBUG [_api.py:364] Attempting to release lock 140135361261920 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
35
+ 2025-11-11 07:47:34,786 DEBUG [_api.py:367] Lock 140135361261920 released on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
36
+ # Accounting: time=26 threads=1
37
+ # Ended (code 0) at Tue Nov 11 07:47:38 UTC 2025, elapsed time 26 seconds
music/test/16splits/log/audio_tokenizer_higg.4.log ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # python egs/pretraining/data_scripts/offline_tokenization_codec.py --rank 4 --input-file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/wav.4.scp --output_file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.4.pt --model_path /turing_music_fs/music_data/ydc/exp2/reason_ckpt/ReasoningCodec.checkpoint --train_config /turing_music_fs/music_data/ydc/exp2/reason_ckpt/infer_config.yaml --tokenizer hugg
2
+ # Started at Tue Nov 11 07:47:12 UTC 2025
3
+ #
4
+ 2025-11-11 07:47:14,311 INFO [offline_tokenization_codec.py:44] max gpu 8
5
+ 2025-11-11 07:47:14,311 INFO [offline_tokenization_codec.py:47] Using device: cuda:3
6
+ /root/miniconda3/envs/uniaudio2/lib/python3.10/site-packages/torch/nn/utils/weight_norm.py:134: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.
7
+ WeightNorm.apply(module, name, dim)
8
+ 2025-11-11 07:47:24,395 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
9
+ 2025-11-11 07:47:24,397 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
10
+ 2025-11-11 07:47:24,660 DEBUG [auth.py:50] Trying paths: ['/root/.docker/config.json', '/root/.dockercfg']
11
+ 2025-11-11 07:47:24,660 DEBUG [auth.py:57] No config file found
12
+ 2025-11-11 07:47:25,144 DEBUG [__init__.py:44] Skipping import of cpp extensions
13
+ [2025-11-11 07:47:25,638] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
14
+ 2025-11-11 07:47:25,717 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpfu44fg8l/test.c -o /tmp/tmpfu44fg8l/test.o
15
+ 2025-11-11 07:47:25,732 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpfu44fg8l/test.o -laio -o /tmp/tmpfu44fg8l/a.out
16
+ 2025-11-11 07:47:26,303 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpxkt80q7_/test.c -o /tmp/tmpxkt80q7_/test.o
17
+ 2025-11-11 07:47:26,318 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpxkt80q7_/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpxkt80q7_/a.out
18
+ /turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single/tools/tokenizer/higgCodec/audio_processing/higgs_audio_tokenizer.py:325: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
19
+ parameter_dict = torch.load(model_path, map_location=device)
20
+ 2025-11-11 07:47:30,301 INFO [offline_tokenization_codec.py:73] tokenizer built
21
+ 2025-11-11 07:47:30,301 INFO [offline_tokenization_codec.py:85] [Rank 3] Checkpoint path is: /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.4.pt.checkpoint_rec.pth
22
+ 2025-11-11 07:47:30,302 INFO [offline_tokenization_codec.py:102] [Rank 3] No checkpoint found. Starting from scratch.
23
+
24
+ audio_signal = torch.tensor(raw_audio)
25
+
26
+ 2025-11-11 07:47:34,729 INFO [offline_tokenization_codec.py:170] [Rank 3] Processing complete. Saving final data...
27
+ 2025-11-11 07:47:34,734 INFO [offline_tokenization_codec.py:188] [Rank 3] Final GPU cache cleanup completed
28
+ 2025-11-11 07:47:34,743 DEBUG [_api.py:331] Attempting to acquire lock 139764354340112 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
29
+ 2025-11-11 07:47:34,743 DEBUG [_api.py:334] Lock 139764354340112 acquired on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
30
+ 2025-11-11 07:47:34,744 DEBUG [_api.py:364] Attempting to release lock 139764354340112 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
31
+ 2025-11-11 07:47:34,744 DEBUG [_api.py:367] Lock 139764354340112 released on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
32
+ 2025-11-11 07:47:34,747 DEBUG [_api.py:331] Attempting to acquire lock 139764354347360 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
33
+ 2025-11-11 07:47:34,747 DEBUG [_api.py:334] Lock 139764354347360 acquired on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
34
+ 2025-11-11 07:47:34,747 DEBUG [_api.py:364] Attempting to release lock 139764354347360 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
35
+ 2025-11-11 07:47:34,747 DEBUG [_api.py:367] Lock 139764354347360 released on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
36
+ # Accounting: time=25 threads=1
37
+ # Ended (code 0) at Tue Nov 11 07:47:37 UTC 2025, elapsed time 25 seconds
music/test/16splits/log/audio_tokenizer_higg.5.log ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # python egs/pretraining/data_scripts/offline_tokenization_codec.py --rank 5 --input-file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/wav.5.scp --output_file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.5.pt --model_path /turing_music_fs/music_data/ydc/exp2/reason_ckpt/ReasoningCodec.checkpoint --train_config /turing_music_fs/music_data/ydc/exp2/reason_ckpt/infer_config.yaml --tokenizer hugg
2
+ # Started at Tue Nov 11 07:47:12 UTC 2025
3
+ #
4
+ 2025-11-11 07:47:14,250 INFO [offline_tokenization_codec.py:44] max gpu 8
5
+ 2025-11-11 07:47:14,250 INFO [offline_tokenization_codec.py:47] Using device: cuda:4
6
+ /root/miniconda3/envs/uniaudio2/lib/python3.10/site-packages/torch/nn/utils/weight_norm.py:134: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.
7
+ WeightNorm.apply(module, name, dim)
8
+ 2025-11-11 07:47:24,265 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
9
+ 2025-11-11 07:47:24,267 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
10
+ 2025-11-11 07:47:24,519 DEBUG [auth.py:50] Trying paths: ['/root/.docker/config.json', '/root/.dockercfg']
11
+ 2025-11-11 07:47:24,519 DEBUG [auth.py:57] No config file found
12
+ 2025-11-11 07:47:24,999 DEBUG [__init__.py:44] Skipping import of cpp extensions
13
+ [2025-11-11 07:47:25,466] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
14
+ 2025-11-11 07:47:25,545 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmprh0l7qq4/test.c -o /tmp/tmprh0l7qq4/test.o
15
+ 2025-11-11 07:47:25,560 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmprh0l7qq4/test.o -laio -o /tmp/tmprh0l7qq4/a.out
16
+ 2025-11-11 07:47:26,092 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmp1fdljbk6/test.c -o /tmp/tmp1fdljbk6/test.o
17
+ 2025-11-11 07:47:26,108 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmp1fdljbk6/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp1fdljbk6/a.out
18
+ /turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single/tools/tokenizer/higgCodec/audio_processing/higgs_audio_tokenizer.py:325: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
19
+ parameter_dict = torch.load(model_path, map_location=device)
20
+ 2025-11-11 07:47:30,315 INFO [offline_tokenization_codec.py:73] tokenizer built
21
+ 2025-11-11 07:47:30,315 INFO [offline_tokenization_codec.py:85] [Rank 4] Checkpoint path is: /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.5.pt.checkpoint_rec.pth
22
+ 2025-11-11 07:47:30,315 INFO [offline_tokenization_codec.py:102] [Rank 4] No checkpoint found. Starting from scratch.
23
+
24
+ audio_signal = torch.tensor(raw_audio)
25
+
26
+ 2025-11-11 07:47:34,811 INFO [offline_tokenization_codec.py:170] [Rank 4] Processing complete. Saving final data...
27
+ 2025-11-11 07:47:34,816 INFO [offline_tokenization_codec.py:188] [Rank 4] Final GPU cache cleanup completed
28
+ 2025-11-11 07:47:34,825 DEBUG [_api.py:331] Attempting to acquire lock 140281799234832 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
29
+ 2025-11-11 07:47:34,826 DEBUG [_api.py:334] Lock 140281799234832 acquired on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
30
+ 2025-11-11 07:47:34,826 DEBUG [_api.py:364] Attempting to release lock 140281799234832 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
31
+ 2025-11-11 07:47:34,826 DEBUG [_api.py:367] Lock 140281799234832 released on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
32
+ 2025-11-11 07:47:34,829 DEBUG [_api.py:331] Attempting to acquire lock 140281799242080 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
33
+ 2025-11-11 07:47:34,829 DEBUG [_api.py:334] Lock 140281799242080 acquired on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
34
+ 2025-11-11 07:47:34,829 DEBUG [_api.py:364] Attempting to release lock 140281799242080 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
35
+ 2025-11-11 07:47:34,829 DEBUG [_api.py:367] Lock 140281799242080 released on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
36
+ # Accounting: time=24 threads=1
37
+ # Ended (code 0) at Tue Nov 11 07:47:36 UTC 2025, elapsed time 24 seconds
music/test/16splits/log/audio_tokenizer_higg.6.log ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # python egs/pretraining/data_scripts/offline_tokenization_codec.py --rank 6 --input-file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/wav.6.scp --output_file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.6.pt --model_path /turing_music_fs/music_data/ydc/exp2/reason_ckpt/ReasoningCodec.checkpoint --train_config /turing_music_fs/music_data/ydc/exp2/reason_ckpt/infer_config.yaml --tokenizer hugg
2
+ # Started at Tue Nov 11 07:47:12 UTC 2025
3
+ #
4
+ 2025-11-11 07:47:14,259 INFO [offline_tokenization_codec.py:44] max gpu 8
5
+ 2025-11-11 07:47:14,259 INFO [offline_tokenization_codec.py:47] Using device: cuda:5
6
+ /root/miniconda3/envs/uniaudio2/lib/python3.10/site-packages/torch/nn/utils/weight_norm.py:134: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.
7
+ WeightNorm.apply(module, name, dim)
8
+ 2025-11-11 07:47:24,330 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
9
+ 2025-11-11 07:47:24,331 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
10
+ 2025-11-11 07:47:24,582 DEBUG [auth.py:50] Trying paths: ['/root/.docker/config.json', '/root/.dockercfg']
11
+ 2025-11-11 07:47:24,582 DEBUG [auth.py:57] No config file found
12
+ 2025-11-11 07:47:25,057 DEBUG [__init__.py:44] Skipping import of cpp extensions
13
+ [2025-11-11 07:47:25,495] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
14
+ 2025-11-11 07:47:25,574 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpghzjhtuz/test.c -o /tmp/tmpghzjhtuz/test.o
15
+ 2025-11-11 07:47:25,589 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpghzjhtuz/test.o -laio -o /tmp/tmpghzjhtuz/a.out
16
+ 2025-11-11 07:47:26,117 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpzz15lcgl/test.c -o /tmp/tmpzz15lcgl/test.o
17
+ 2025-11-11 07:47:26,132 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpzz15lcgl/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpzz15lcgl/a.out
18
+ /turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single/tools/tokenizer/higgCodec/audio_processing/higgs_audio_tokenizer.py:325: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
19
+ parameter_dict = torch.load(model_path, map_location=device)
20
+ 2025-11-11 07:47:30,324 INFO [offline_tokenization_codec.py:73] tokenizer built
21
+ 2025-11-11 07:47:30,325 INFO [offline_tokenization_codec.py:85] [Rank 5] Checkpoint path is: /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.6.pt.checkpoint_rec.pth
22
+ 2025-11-11 07:47:30,325 INFO [offline_tokenization_codec.py:102] [Rank 5] No checkpoint found. Starting from scratch.
23
+
24
+ audio_signal = torch.tensor(raw_audio)
25
+
26
+ 2025-11-11 07:47:34,788 INFO [offline_tokenization_codec.py:170] [Rank 5] Processing complete. Saving final data...
27
+ 2025-11-11 07:47:34,793 INFO [offline_tokenization_codec.py:188] [Rank 5] Final GPU cache cleanup completed
28
+ 2025-11-11 07:47:34,802 DEBUG [_api.py:331] Attempting to acquire lock 140528161019152 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
29
+ 2025-11-11 07:47:34,802 DEBUG [_api.py:334] Lock 140528161019152 acquired on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
30
+ 2025-11-11 07:47:34,802 DEBUG [_api.py:364] Attempting to release lock 140528161019152 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
31
+ 2025-11-11 07:47:34,802 DEBUG [_api.py:367] Lock 140528161019152 released on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
32
+ 2025-11-11 07:47:34,805 DEBUG [_api.py:331] Attempting to acquire lock 140528161026400 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
33
+ 2025-11-11 07:47:34,805 DEBUG [_api.py:334] Lock 140528161026400 acquired on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
34
+ 2025-11-11 07:47:34,805 DEBUG [_api.py:364] Attempting to release lock 140528161026400 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
35
+ 2025-11-11 07:47:34,805 DEBUG [_api.py:367] Lock 140528161026400 released on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
36
+ # Accounting: time=25 threads=1
37
+ # Ended (code 0) at Tue Nov 11 07:47:37 UTC 2025, elapsed time 25 seconds
music/test/16splits/log/audio_tokenizer_higg.7.log ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # python egs/pretraining/data_scripts/offline_tokenization_codec.py --rank 7 --input-file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/wav.7.scp --output_file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.7.pt --model_path /turing_music_fs/music_data/ydc/exp2/reason_ckpt/ReasoningCodec.checkpoint --train_config /turing_music_fs/music_data/ydc/exp2/reason_ckpt/infer_config.yaml --tokenizer hugg
2
+ # Started at Tue Nov 11 07:47:12 UTC 2025
3
+ #
4
+ 2025-11-11 07:47:14,290 INFO [offline_tokenization_codec.py:44] max gpu 8
5
+ 2025-11-11 07:47:14,290 INFO [offline_tokenization_codec.py:47] Using device: cuda:6
6
+ /root/miniconda3/envs/uniaudio2/lib/python3.10/site-packages/torch/nn/utils/weight_norm.py:134: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.
7
+ WeightNorm.apply(module, name, dim)
8
+ 2025-11-11 07:47:24,308 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
9
+ 2025-11-11 07:47:24,310 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
10
+ 2025-11-11 07:47:24,564 DEBUG [auth.py:50] Trying paths: ['/root/.docker/config.json', '/root/.dockercfg']
11
+ 2025-11-11 07:47:24,564 DEBUG [auth.py:57] No config file found
12
+ 2025-11-11 07:47:25,042 DEBUG [__init__.py:44] Skipping import of cpp extensions
13
+ [2025-11-11 07:47:25,507] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
14
+ 2025-11-11 07:47:25,586 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpcng4j604/test.c -o /tmp/tmpcng4j604/test.o
15
+ 2025-11-11 07:47:25,602 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpcng4j604/test.o -laio -o /tmp/tmpcng4j604/a.out
16
+ 2025-11-11 07:47:26,153 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpq2i617y5/test.c -o /tmp/tmpq2i617y5/test.o
17
+ 2025-11-11 07:47:26,168 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpq2i617y5/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpq2i617y5/a.out
18
+ /turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single/tools/tokenizer/higgCodec/audio_processing/higgs_audio_tokenizer.py:325: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
19
+ parameter_dict = torch.load(model_path, map_location=device)
20
+ 2025-11-11 07:47:30,353 INFO [offline_tokenization_codec.py:73] tokenizer built
21
+ 2025-11-11 07:47:30,353 INFO [offline_tokenization_codec.py:85] [Rank 6] Checkpoint path is: /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.7.pt.checkpoint_rec.pth
22
+ 2025-11-11 07:47:30,353 INFO [offline_tokenization_codec.py:102] [Rank 6] No checkpoint found. Starting from scratch.
23
+
24
+ audio_signal = torch.tensor(raw_audio)
25
+
26
+ 2025-11-11 07:47:34,820 INFO [offline_tokenization_codec.py:170] [Rank 6] Processing complete. Saving final data...
27
+ 2025-11-11 07:47:34,825 INFO [offline_tokenization_codec.py:188] [Rank 6] Final GPU cache cleanup completed
28
+ 2025-11-11 07:47:34,836 DEBUG [_api.py:331] Attempting to acquire lock 140463004030224 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
29
+ 2025-11-11 07:47:34,837 DEBUG [_api.py:334] Lock 140463004030224 acquired on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
30
+ 2025-11-11 07:47:34,837 DEBUG [_api.py:364] Attempting to release lock 140463004030224 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
31
+ 2025-11-11 07:47:34,837 DEBUG [_api.py:367] Lock 140463004030224 released on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
32
+ 2025-11-11 07:47:34,840 DEBUG [_api.py:331] Attempting to acquire lock 140463004037472 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
33
+ 2025-11-11 07:47:34,840 DEBUG [_api.py:334] Lock 140463004037472 acquired on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
34
+ 2025-11-11 07:47:34,840 DEBUG [_api.py:364] Attempting to release lock 140463004037472 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
35
+ 2025-11-11 07:47:34,840 DEBUG [_api.py:367] Lock 140463004037472 released on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
36
+ # Accounting: time=25 threads=1
37
+ # Ended (code 0) at Tue Nov 11 07:47:37 UTC 2025, elapsed time 25 seconds
music/test/16splits/log/audio_tokenizer_higg.8.log ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # python egs/pretraining/data_scripts/offline_tokenization_codec.py --rank 8 --input-file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/wav.8.scp --output_file /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.8.pt --model_path /turing_music_fs/music_data/ydc/exp2/reason_ckpt/ReasoningCodec.checkpoint --train_config /turing_music_fs/music_data/ydc/exp2/reason_ckpt/infer_config.yaml --tokenizer hugg
2
+ # Started at Tue Nov 11 07:47:12 UTC 2025
3
+ #
4
+ 2025-11-11 07:47:14,191 INFO [offline_tokenization_codec.py:44] max gpu 8
5
+ 2025-11-11 07:47:14,191 INFO [offline_tokenization_codec.py:47] Using device: cuda:7
6
+ /root/miniconda3/envs/uniaudio2/lib/python3.10/site-packages/torch/nn/utils/weight_norm.py:134: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.
7
+ WeightNorm.apply(module, name, dim)
8
+ 2025-11-11 07:47:24,034 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
9
+ 2025-11-11 07:47:24,036 DEBUG [cmd.py:1253] Popen(['git', 'version'], cwd=/turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single, stdin=None, shell=False, universal_newlines=False)
10
+ 2025-11-11 07:47:24,304 DEBUG [auth.py:50] Trying paths: ['/root/.docker/config.json', '/root/.dockercfg']
11
+ 2025-11-11 07:47:24,304 DEBUG [auth.py:57] No config file found
12
+ 2025-11-11 07:47:24,775 DEBUG [__init__.py:44] Skipping import of cpp extensions
13
+ [2025-11-11 07:47:25,261] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
14
+ 2025-11-11 07:47:25,339 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpfvq8p6gz/test.c -o /tmp/tmpfvq8p6gz/test.o
15
+ 2025-11-11 07:47:25,355 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpfvq8p6gz/test.o -laio -o /tmp/tmpfvq8p6gz/a.out
16
+ 2025-11-11 07:47:25,882 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -O2 -isystem /root/miniconda3/envs/uniaudio2/include -fPIC -c /tmp/tmpom0qpfgb/test.c -o /tmp/tmpom0qpfgb/test.o
17
+ 2025-11-11 07:47:25,897 INFO [spawn.py:60] gcc -pthread -B /root/miniconda3/envs/uniaudio2/compiler_compat /tmp/tmpom0qpfgb/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpom0qpfgb/a.out
18
+ /turing_music_fs/music_data/ydc/code2/TokenPPL/Token_LM_single/tools/tokenizer/higgCodec/audio_processing/higgs_audio_tokenizer.py:325: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
19
+ parameter_dict = torch.load(model_path, map_location=device)
20
+ 2025-11-11 07:47:30,301 INFO [offline_tokenization_codec.py:73] tokenizer built
21
+ 2025-11-11 07:47:30,302 INFO [offline_tokenization_codec.py:85] [Rank 7] Checkpoint path is: /turing_music_fs/music_data/ydc/code2/TokenPPL/data/music/test/16splits/hugg_tokens.8.pt.checkpoint_rec.pth
22
+ 2025-11-11 07:47:30,302 INFO [offline_tokenization_codec.py:102] [Rank 7] No checkpoint found. Starting from scratch.
23
+
24
+ audio_signal = torch.tensor(raw_audio)
25
+
26
+ 2025-11-11 07:47:34,683 INFO [offline_tokenization_codec.py:170] [Rank 7] Processing complete. Saving final data...
27
+ 2025-11-11 07:47:34,709 INFO [offline_tokenization_codec.py:188] [Rank 7] Final GPU cache cleanup completed
28
+ 2025-11-11 07:47:34,718 DEBUG [_api.py:331] Attempting to acquire lock 140627373233424 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
29
+ 2025-11-11 07:47:34,719 DEBUG [_api.py:334] Lock 140627373233424 acquired on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
30
+ 2025-11-11 07:47:34,719 DEBUG [_api.py:364] Attempting to release lock 140627373233424 on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
31
+ 2025-11-11 07:47:34,719 DEBUG [_api.py:367] Lock 140627373233424 released on /root/.triton/autotune/Fp16Matmul_2d_kernel.pickle.lock
32
+ 2025-11-11 07:47:34,722 DEBUG [_api.py:331] Attempting to acquire lock 140627373240672 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
33
+ 2025-11-11 07:47:34,722 DEBUG [_api.py:334] Lock 140627373240672 acquired on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
34
+ 2025-11-11 07:47:34,722 DEBUG [_api.py:364] Attempting to release lock 140627373240672 on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
35
+ 2025-11-11 07:47:34,722 DEBUG [_api.py:367] Lock 140627373240672 released on /root/.triton/autotune/Fp16Matmul_4d_kernel.pickle.lock
36
+ # Accounting: time=24 threads=1
37
+ # Ended (code 0) at Tue Nov 11 07:47:36 UTC 2025, elapsed time 24 seconds