Failed to import the NeMo framework or its dependencies! 'pip install -U \"batchalign[speaker]

[Noah-Jaffe]

Issue:

In https://talkbank.org/info/BA2-usage.pdf:

To use Batchalign, you first have to install a version of Python greater than 3.9. You can
install the current version from https://python.org/downloads. Some external
dependencies will be installed automatically when you install Batchalign. This means
that, you should not install any dependencies (like Whisper or Nvidia Nemo) manually. If
a special function requires a manual dependency, the Batchalign program will prompt
you with instructions

Attempting to get a transcript with a working diarize to run on my local machine.

batchalign -vvv transcribe --lang=eng --whisper --num_speakers=2 --diarize .\.tmp_test\in .\.tmp_test\out\

Expand for stacktrace

[03/02/25 10:45:02] DEBUG    Attempting to create               dispatch.py:134
                             BatchalignPipeline for CLI...                     
                    DEBUG    Initializing packages, got:         dispatch.py:64
                             packages='['asr', 'speaker']' and                 
                             config='{'DEFAULT': , 'asr': ,                  
                             'ud': }'                             
                    INFO     Initializing engines...             dispatch.py:85
                    INFO     -------------------------------     dispatch.py:86
                    INFO     | asr          |      whisper |    dispatch.py:100
                    INFO     -------------------------------    dispatch.py:101
                    DEBUG    Initializing whisper model...      infer_asr.py:67
Device set to use cpu
[03/02/25 10:45:03] DEBUG    Done, initalizing processor and    infer_asr.py:81
                             config...                                         
                    DEBUG    Whisper initialization done.       infer_asr.py:83
                    DEBUG    Initializing utterance model...      whisper.py:46
                    DEBUG    Done.                                whisper.py:52
                    INFO     | speaker      | nemo_speaker |    dispatch.py:100
                    INFO     -------------------------------    dispatch.py:101
+--------------------- Traceback (most recent call last) ---------------------+
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\batchalign\models\speaker\infer.py:27 in __init__                         |
|                                                                             |
|   24 class NemoSpeakerModel(object):                                        |
|   25     def __init__(self):                                                |
|   26         try:                                                           |
| > 27             from omegaconf import OmegaConf                            |
|   28             self.__base = OmegaConf.load(resolve_config())             |
|   29         except ImportError:                                            |
|   30             self.__raise()                                             |
+-----------------------------------------------------------------------------+
ModuleNotFoundError: No module named 'omegaconf'
System.Management.Automation.RemoteException
During handling of the above exception, another exception occurred:
System.Management.Automation.RemoteException
+--------------------- Traceback (most recent call last) ---------------------+
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\batchalign\cli\dispatch.py:135 in _dispatch                               |
|                                                                             |
|   132                                                                       |
|   133         # create pipeline and read files                              |
|   134         baL.debug("Attempting to create BatchalignPipeline for CLI... |
| > 135         pipeline = BatchalignPipeline.new(Cmd2Task[command],          |
|   136                                           lang=lang, num_speakers=num |
|   137         baL.debug(f"Successfully created BatchalignPipeline... {pipel |
|   138                                                                       |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\batchalign\pipelines\pipeline.py:60 in new                                |
|                                                                             |
|    57         """                                                           |
|    58                                                                       |
|    59         from batchalign.pipelines.dispatch import dispatch_pipeline   |
| >  60         return dispatch_pipeline(tasks, lang=lang, num_speakers=num_s |
|    61                                                                       |
|    62     def __call__(self, input, callback=None, **kwargs):               |
|    63         """Call the pipeline.                                         |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\batchalign\pipelines\dispatch.py:126 in dispatch_pipeline                 |
|                                                                             |
|   123         elif engine == "evaluation":                                  |
|   124             engines.append(EvaluationEngine())                        |
|   125         elif engine == "nemo_speaker":                                |
| > 126             engines.append(NemoSpeakerEngine(num_speakers=num_speaker |
|   127         elif engine == "stanza_utt":                                  |
|   128             engines.append(StanzaUtteranceEngine())                   |
|   129         elif engine == "stanza_coref":                                |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\batchalign\pipelines\speaker\nemo_speaker.py:26 in __init__               |
|                                                                             |
|   23                                                                        |
|   24         self.status_hook = None                                        |
|   25         self.num_speakers = num_speakers                               |
| > 26         self.__model = NemoSpeakerModel()                              |
|   27                                                                        |
|   28     def process(self, doc:Document, **kwargs):                         |
|   29         # check that the document has a media path to align to         |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\batchalign\models\speaker\infer.py:30 in __init__                         |
|                                                                             |
|   27             from omegaconf import OmegaConf                            |
|   28             self.__base = OmegaConf.load(resolve_config())             |
|   29         except ImportError:                                            |
| > 30             self.__raise()                                             |
|   31                                                                        |
|   32     def __raise(self):                                                 |
|   33         raise ImportError("Failed to import the NeMo framework or its  |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\batchalign\models\speaker\infer.py:33 in __raise                          |
|                                                                             |
|   30             self.__raise()                                             |
|   31                                                                        |
|   32     def __raise(self):                                                 |
| > 33         raise ImportError("Failed to import the NeMo framework or its  |
|   34                                                                        |
|   35     def __call__(self, in_file, num_speakers=2):                       |
|   36         try:                                                           |
+-----------------------------------------------------------------------------+
ImportError: Failed to import the NeMo framework or its dependencies!
Hint: run 'pip install -U "batchalign[speaker]"' to install speaker diarization
tools.
System.Management.Automation.RemoteException
During handling of the above exception, another exception occurred:
System.Management.Automation.RemoteException
+--------------------- Traceback (most recent call last) ---------------------+
| in _run_module_as_main:198                                                  |
| in _run_code:88                                                             |
|                                                                             |
| in :7                                                               |
|                                                                             |
|   4 from batchalign.cli.cli import batchalign                               |
|   5 if __name__ == '__main__':                                              |
|   6     sys.argv[0] = re.sub(r'(-script\.pyw|\.exe)?$', '', sys.argv[0])    |
| > 7     sys.exit(batchalign())                                              |
|   8                                                                         |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\rich_click\rich_command.py:367 in __call__                                |
|                                                                             |
|   364         # Include this here because I run into a false warning        |
|   365         # in the PyCharm IDE otherwise; for some reason PyCharm doesn |
|   366         # seem to think RichGroups are callable. (No issues with Mypy |
| > 367         return super().__call__(*args, **kwargs)                      |
|   368                                                                       |
|   369                                                                       |
|   370 class RichCommandCollection(CommandCollection, RichGroup):            |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\click\core.py:1161 in __call__                                            |
|                                                                             |
|   1158                                                                      |
|   1159     def __call__(self, *args: t.Any, **kwargs: t.Any) -> t.Any:      |
|   1160         """Alias for :meth:`main`."""                                |
| > 1161         return self.main(*args, **kwargs)                            |
|   1162                                                                      |
|   1163                                                                      |
|   1164 class Command(BaseCommand):                                          |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\rich_click\rich_command.py:152 in main                                    |
|                                                                             |
|   149         try:                                                          |
|   150             try:                                                      |
|   151                 with self.make_context(prog_name, args, **extra) as c |
| > 152                     rv = self.invoke(ctx)                             |
|   153                     if not standalone_mode:                           |
|   154                         return rv                                     |
|   155                     # it's not safe to `ctx.exit(rv)` here!           |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\click\core.py:1697 in invoke                                              |
|                                                                             |
|   1694                 super().invoke(ctx)                                  |
|   1695                 sub_ctx = cmd.make_context(cmd_name, args, parent=ct |
|   1696                 with sub_ctx:                                        |
| > 1697                     return _process_result(sub_ctx.command.invoke(su |
|   1698                                                                      |
|   1699         # In chain mode we create the contexts step by step, but aft |
|   1700         # base command has been invoked.  Because at that point we d |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\click\core.py:1443 in invoke                                              |
|                                                                             |
|   1440             echo(style(message, fg="red"), err=True)                 |
|   1441                                                                      |
|   1442         if self.callback is not None:                                |
| > 1443             return ctx.invoke(self.callback, **ctx.params)           |
|   1444                                                                      |
|   1445     def shell_complete(self, ctx: Context, incomplete: str) -> t.Lis |
|   1446         """Return a list of completions for the incomplete value. Lo |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\click\core.py:788 in invoke                                               |
|                                                                             |
|    785                                                                      |
|    786         with augment_usage_errors(__self):                           |
|    787             with ctx:                                                |
| >  788                 return __callback(*args, **kwargs)                   |
|    789                                                                      |
|    790     def forward(__self, __cmd: "Command", *args: t.Any, **kwargs: t. |
|    791         """Similar to :meth:`invoke` but fills in default keyword    |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\click\decorators.py:33 in new_func                                        |
|                                                                             |
|    30     """                                                               |
|    31                                                                       |
|    32     def new_func(*args: "P.args", **kwargs: "P.kwargs") -> "R":       |
| >  33         return f(get_current_context(), *args, **kwargs)              |
|    34                                                                       |
|    35     return update_wrapper(new_func, f)                                |
|    36                                                                       |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\batchalign\cli\cli.py:187 in transcribe                                   |
|                                                                             |
|   184                                 write_wor=kwargs.get("wor", False))   |
|   185                                                                       |
|   186     if kwargs.get("diarize"):                                         |
| > 187         _dispatch("transcribe_s",                                     |
|   188                   lang, num_speakers, ["mp3", "mp4", "wav"], ctx,     |
|   189                   in_dir, out_dir,                                    |
|   190                   loader, writer, C,                                  |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\batchalign\cli\dispatch.py:126 in _dispatch                               |
|                                                                             |
|   123     # cache the errors                                                |
|   124     errors = []                                                       |
|   125                                                                       |
| > 126     with prog as prog:                                                |
|   127         tasks = {}                                                    |
|   128         errors = []                                                   |
|   129         # create the spinner bars                                     |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\rich\progress.py:1191 in __exit__                                         |
|                                                                             |
|   1188         exc_val: Optional[BaseException],                            |
|   1189         exc_tb: Optional[TracebackType],                             |
|   1190     ) -> None:                                                       |
| > 1191         self.stop()                                                  |
|   1192                                                                      |
|   1193     def track(                                                       |
|   1194         self,                                                        |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\rich\progress.py:1177 in stop                                             |
|                                                                             |
|   1174                                                                      |
|   1175     def stop(self) -> None:                                          |
|   1176         """Stop the progress display."""                             |
| > 1177         self.live.stop()                                             |
|   1178         if not self.console.is_interactive and not self.console.is_j |
|   1179             self.console.print()                                     |
|   1180                                                                      |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\rich\live.py:147 in stop                                                  |
|                                                                             |
|   144                 self._refresh_thread = None                           |
|   145             # allow it to fully render on the last even if overflow   |
|   146             self.vertical_overflow = "visible"                        |
| > 147             with self.console:                                        |
|   148                 try:                                                  |
|   149                     if not self._alt_screen and not self.console.is_j |
|   150                         self.refresh()                                |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\rich\console.py:864 in __exit__                                           |
|                                                                             |
|    861                                                                      |
|    862     def __exit__(self, exc_type: Any, exc_value: Any, traceback: Any |
|    863         """Exit buffer context."""                                   |
| >  864         self._exit_buffer()                                          |
|    865                                                                      |
|    866     def begin_capture(self) -> None:                                 |
|    867         """Begin capturing console output. Call :meth:`end_capture`  |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\rich\console.py:822 in _exit_buffer                                       |
|                                                                             |
|    819     def _exit_buffer(self) -> None:                                  |
|    820         """Leave buffer context, and render content if required."""  |
|    821         self._buffer_index -= 1                                      |
| >  822         self._check_buffer()                                         |
|    823                                                                      |
|    824     def set_live(self, live: "Live") -> None:                        |
|    825         """Set Live instance. Used by Live context manager.          |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\rich\console.py:2019 in _check_buffer                                     |
|                                                                             |
|   2016             return                                                   |
|   2017                                                                      |
|   2018         try:                                                         |
| > 2019             self._write_buffer()                                     |
|   2020         except BrokenPipeError:                                      |
|   2021             self.on_broken_pipe()                                    |
|   2022                                                                      |
|                                                                             |
| c:\Users\PC\Documents\python_scripts\videotranscript\.venv\Lib\site-package |
| s\rich\console.py:2067 in _write_buffer                                     |
|                                                                             |
|   2064                             MAX_WRITE = 32 * 1024 // 4               |
|   2065                             try:                                     |
|   2066                                 if len(text) <= MAX_WRITE:           |
| > 2067                                     write(text)                      |
|   2068                                 else:                                |
|   2069                                     batch: List[str] = []            |
|   2070                                     batch_append = batch.append      |
|                                                                             |
| C:\Users\PC\AppData\Local\Programs\Python\Python312\Lib\tempfile.py:499 in  |
| func_wrapper                                                                |
|                                                                             |
|   496             func = a                                                  |
|   497             @_functools.wraps(func)                                   |
|   498             def func_wrapper(*args, **kwargs):                        |
| > 499                 return func(*args, **kwargs)                          |
|   500             # Avoid closing the file as long as the wrapper is alive, |
|   501             # see issue #18879.                                       |
|   502             func_wrapper._closer = self._closer                       |
|                                                                             |
| C:\Users\PC\AppData\Local\Programs\Python\Python312\Lib\encodings\cp1252.py |
| :19 in encode                                                               |
|                                                                             |
|    16                                                                       |
|    17 class IncrementalEncoder(codecs.IncrementalEncoder):                  |
|    18     def encode(self, input, final=False):                             |
| >  19         return codecs.charmap_encode(input,self.errors,encoding_table |
|    20                                                                       |
|    21 class IncrementalDecoder(codecs.IncrementalDecoder):                  |
|    22     def decode(self, input, final=False):                             |
+-----------------------------------------------------------------------------+
UnicodeEncodeError: 'charmap' codec can't encode character '\u2827' in position
0: character maps to 
*** You may need to add PYTHONIOENCODING=utf-8 to your environment ***

I tried this with python 3.12.9, 3.11.11, and 3.9 (all in a fresh venv).
Whats the deal?

[Noah-Jaffe]

If I run pip install omegaconf and rerun the transcript command we get the following error

Expand for stacktrace

[03/02/25 10:53:05] DEBUG    Attempting to create               dispatch.py:134

BatchalignPipeline for CLI...

DEBUG    Initializing packages, got:         dispatch.py:64

packages='['asr', 'speaker']' and

config='{'DEFAULT': <Section:

DEFAULT>, 'asr': <Section: asr>,

'ud': <Section: ud>}'

INFO     Initializing engines...             dispatch.py:85

INFO     -------------------------------     dispatch.py:86

INFO     | asr          |      whisper |    dispatch.py:100

INFO     -------------------------------    dispatch.py:101

DEBUG    Initializing whisper model...      infer_asr.py:67

Device set to use cpu

[03/02/25 10:53:06] DEBUG    Done, initalizing processor and    infer_asr.py:81

config...

DEBUG    Whisper initialization done.       infer_asr.py:83

DEBUG    Initializing utterance model...      whisper.py:46

DEBUG    Done.                                whisper.py:52

INFO     | speaker      | nemo_speaker |    dispatch.py:100

INFO     -------------------------------    dispatch.py:101

INFO     | disfluency   |  replacement |    dispatch.py:100

INFO     -------------------------------    dispatch.py:101

INFO     | retracing    |        ngram |    dispatch.py:100

INFO     -------------------------------    dispatch.py:101

DEBUG    Done initalizing packages.         dispatch.py:136

DEBUG    Successfully created               dispatch.py:137

BatchalignPipeline...

<batchalign.pipelines.pipeline.Bat

chalignPipeline object at

0x000001B09D90A750>

INFO     Pipeline called with engines:       pipeline.py:81

generator=<batchalign.pipelines.asr

.whisper.WhisperEngine object at

0x000001B09D3E1CA0>,

processors=[<batchalign.pipelines.s

peaker.nemo_speaker.NemoSpeakerEngi

ne object at 0x000001B09D428F20>,

<batchalign.pipelines.cleanup.disfl

uencies.DisfluencyReplacementEngine

object at 0x000001B09DC98EC0>,

<batchalign.pipelines.cleanup.retra

ce.NgramRetraceEngine object at

0x000001B09DC98E00>], analyzer=None

on input of type <class 'str'>

DEBUG    Transforming input of type: <class  pipeline.py:90

'str'>

DEBUG    Final input type: <class 'str'>    pipeline.py:100

DEBUG    Calling generator:                 pipeline.py:109

<batchalign.pipelines.asr.whisper.

WhisperEngine object at

0x000001B09D3E1CA0>

INFO     Whisper transcribing file...      infer_asr.py:125

DEBUG    Whisper Preprocessing...          infer_asr.py:126

DEBUG    Whisper Transcribing...           infer_asr.py:151

You have passed task=transcribe, but also have set forced_decoder_ids to [[1, None], [2, 50359]] which creates a conflict. forced_decoder_ids will be ignored in favor of task=transcribe.

The attention mask is not set and cannot be inferred from input because pad token is same as eos token. As a consequence, you may observe unexpected behavior. Please pass your input's attention_mask to obtain reliable results.

WhisperModel is using WhisperSdpaAttention, but torch.nn.functional.scaled_dot_product_attention does not support output_attentions=True or layer_head_mask not None. Falling back to the manual attention implementation, but specifying the manual implementation will be required from Transformers version v5.0.0 onwards. This warning can be removed using the argument attn_implementation="eager" when loading the model.

Whisper did not predict an ending timestamp, which can happen if audio is cut off in the middle of a word. Also make sure WhisperTimeStampLogitsProcessor was used during generation.

[03/02/25 10:53:26] DEBUG    Whisper Postprocessing...         infer_asr.py:184

DEBUG    Whisper Done.                     infer_asr.py:253

DEBUG    Calling processor: processor 1/3,  pipeline.py:122

<batchalign.pipelines.speaker.nemo

_speaker.NemoSpeakerEngine object

at 0x000001B09D428F20>

INFO     Running speaker ID on           nemo_speaker.py:40

C:\Users\PC\Documents\python_sc

ripts\videotranscript.tmp_test

\in\clip.mp3 with 2

speaker(s)...

INFO     Pipeline called with engines:       pipeline.py:81

generator=<batchalign.pipelines.asr

.whisper.WhisperEngine object at

0x000001B09D3E1CA0>,

processors=[<batchalign.pipelines.s

peaker.nemo_speaker.NemoSpeakerEngi

ne object at 0x000001B09D428F20>,

<batchalign.pipelines.cleanup.disfl

uencies.DisfluencyReplacementEngine

object at 0x000001B09DC98EC0>,

<batchalign.pipelines.cleanup.retra

ce.NgramRetraceEngine object at

0x000001B09DC98E00>], analyzer=None

on input of type <class 'str'>

DEBUG    Transforming input of type: <class  pipeline.py:90

'str'>

DEBUG    Final input type: <class 'str'>    pipeline.py:100

DEBUG    Calling generator:                 pipeline.py:109

<batchalign.pipelines.asr.whisper.

WhisperEngine object at

0x000001B09D3E1CA0>

INFO     Whisper transcribing file...      infer_asr.py:125

DEBUG    Whisper Preprocessing...          infer_asr.py:126

DEBUG    Whisper Transcribing...           infer_asr.py:151

Whisper did not predict an ending timestamp, which can happen if audio is cut off in the middle of a word. Also make sure WhisperTimeStampLogitsProcessor was used during generation.

[03/02/25 10:53:44] DEBUG    Whisper Postprocessing...         infer_asr.py:184

DEBUG    Whisper Done.                     infer_asr.py:253

DEBUG    Calling processor: processor 1/3,  pipeline.py:122

<batchalign.pipelines.speaker.nemo

_speaker.NemoSpeakerEngine object

at 0x000001B09D428F20>

INFO     Running speaker ID on           nemo_speaker.py:40

C:\Users\PC\Documents\python_sc

ripts\videotranscript.tmp_test

\in\clip2.mp3 with 2

speaker(s)...

Traceback (most recent call last):

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\models\speaker\infer.py", line 38, in call

from nemo.collections.asr.models.msdd_models import NeuralDiarizer

ModuleNotFoundError: No module named 'nemo'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\cli\dispatch.py", line 162, in _dispatch

doc = pipeline(doc,

^^^^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\pipelines\pipeline.py", line 128, in call

doc = p.process(doc, **kwargs)

^^^^^^^^^^^^^^^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\pipelines\speaker\nemo_speaker.py", line 41, in process

speaker_ids = self.__model(doc.media.url)

^^^^^^^^^^^^^^^^^^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\models\speaker\infer.py", line 43, in call

self.__raise()

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\models\speaker\infer.py", line 33, in __raise

raise ImportError("Failed to import the NeMo framework or its

dependencies!\nHint: run 'pip install -U "batchalign[speaker]"' to install

speaker diarization tools.")

ImportError: Failed to import the NeMo framework or its dependencies!

Hint: run 'pip install -U "batchalign[speaker]"' to install speaker diarization

tools.
Traceback (most recent call last):

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\models\speaker\infer.py", line 38, in call

from nemo.collections.asr.models.msdd_models import NeuralDiarizer

ModuleNotFoundError: No module named 'nemo'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\cli\dispatch.py", line 162, in _dispatch

doc = pipeline(doc,

^^^^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\pipelines\pipeline.py", line 128, in call

doc = p.process(doc, **kwargs)

^^^^^^^^^^^^^^^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\pipelines\speaker\nemo_speaker.py", line 41, in process

speaker_ids = self.__model(doc.media.url)

^^^^^^^^^^^^^^^^^^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\models\speaker\infer.py", line 43, in call

self.__raise()

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\models\speaker\infer.py", line 33, in __raise

raise ImportError("Failed to import the NeMo framework or its

dependencies!\nHint: run 'pip install -U "batchalign[speaker]"' to install

speaker diarization tools.")

ImportError: Failed to import the NeMo framework or its dependencies!

Hint: run 'pip install -U "batchalign[speaker]"' to install speaker diarization

tools.

[Noah-Jaffe]

If we attempt to run pip install -U "batchalign[speaker]

Expand for stacktrace

Requirement already satisfied: batchalign[speaker] in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (0.7.16)

WARNING: batchalign 0.7.16 does not provide the extra 'speaker'

Requirement already satisfied: pydantic>=2.4 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (2.10.6)

Requirement already satisfied: nltk>=3.8 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (3.9.1)

Requirement already satisfied: praatio<6.1.0,>=6.0.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (6.0.1)

Requirement already satisfied: torch in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (2.6.0)

Requirement already satisfied: torchaudio in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (2.6.0)

Requirement already satisfied: hmmlearn==0.3.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (0.3.0)

Requirement already satisfied: eyed3 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (0.9.7)

Requirement already satisfied: pydub in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (0.25.1)

Requirement already satisfied: imblearn in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (0.0)

Requirement already satisfied: plotly>=5.3.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (6.0.0)

Requirement already satisfied: transformers>=4.37 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (4.49.0)

Requirement already satisfied: tokenizers>=0.14.1 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (0.21.0)

Requirement already satisfied: pycountry>=22.3 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (24.6.1)

Requirement already satisfied: stanza>=1.9.1 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from stanza[transformers]>=1.9.1->batchalign[speaker]) (1.10.1)

Requirement already satisfied: scipy~=1.11 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (1.15.2)

Requirement already satisfied: rev_ai>=2.18.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (2.21.0)

Requirement already satisfied: rich~=13.6 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (13.9.4)

Requirement already satisfied: click~=8.1 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (8.1.8)

Requirement already satisfied: matplotlib<4.0.0,>=3.8.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (3.10.1)

Requirement already satisfied: pyfiglet==1.0.2 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (1.0.2)

Requirement already satisfied: setuptools in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (75.8.2)

Requirement already satisfied: soundfile~=0.12.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (0.12.1)

Requirement already satisfied: rich-click>=1.7.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (1.8.6)

Requirement already satisfied: typing-extensions in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (4.12.2)

Requirement already satisfied: num2words in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (0.5.14)

Requirement already satisfied: tiktoken in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (0.9.0)

Requirement already satisfied: blobfile in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (3.0.0)

Requirement already satisfied: sentencepiece in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from batchalign[speaker]) (0.2.0)

Requirement already satisfied: numpy>=1.10 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from hmmlearn==0.3.0->batchalign[speaker]) (2.2.3)

Requirement already satisfied: scikit-learn!=0.22.0,>=0.16 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from hmmlearn==0.3.0->batchalign[speaker]) (1.6.1)

Requirement already satisfied: colorama in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from click~=8.1->batchalign[speaker]) (0.4.6)

Requirement already satisfied: contourpy>=1.0.1 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from matplotlib<4.0.0,>=3.8.0->batchalign[speaker]) (1.3.1)

Requirement already satisfied: cycler>=0.10 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from matplotlib<4.0.0,>=3.8.0->batchalign[speaker]) (0.12.1)

Requirement already satisfied: fonttools>=4.22.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from matplotlib<4.0.0,>=3.8.0->batchalign[speaker]) (4.56.0)

Requirement already satisfied: kiwisolver>=1.3.1 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from matplotlib<4.0.0,>=3.8.0->batchalign[speaker]) (1.4.8)

Requirement already satisfied: packaging>=20.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from matplotlib<4.0.0,>=3.8.0->batchalign[speaker]) (24.2)

Requirement already satisfied: pillow>=8 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from matplotlib<4.0.0,>=3.8.0->batchalign[speaker]) (11.1.0)

Requirement already satisfied: pyparsing>=2.3.1 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from matplotlib<4.0.0,>=3.8.0->batchalign[speaker]) (3.2.1)

Requirement already satisfied: python-dateutil>=2.7 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from matplotlib<4.0.0,>=3.8.0->batchalign[speaker]) (2.9.0.post0)

Requirement already satisfied: joblib in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from nltk>=3.8->batchalign[speaker]) (1.4.2)

Requirement already satisfied: regex>=2021.8.3 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from nltk>=3.8->batchalign[speaker]) (2024.11.6)

Requirement already satisfied: tqdm in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from nltk>=3.8->batchalign[speaker]) (4.67.1)

Requirement already satisfied: narwhals>=1.15.1 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from plotly>=5.3.0->batchalign[speaker]) (1.28.0)

Requirement already satisfied: annotated-types>=0.6.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from pydantic>=2.4->batchalign[speaker]) (0.7.0)

Requirement already satisfied: pydantic-core==2.27.2 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from pydantic>=2.4->batchalign[speaker]) (2.27.2)

Requirement already satisfied: requests<3.0.0,>=2.21.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from rev_ai>=2.18.0->batchalign[speaker]) (2.32.3)

Requirement already satisfied: six<2.0.0,>=1.12.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from rev_ai>=2.18.0->batchalign[speaker]) (1.17.0)

Requirement already satisfied: websocket-client<1.0.0,>=0.56.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from rev_ai>=2.18.0->batchalign[speaker]) (0.59.0)

Requirement already satisfied: markdown-it-py>=2.2.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from rich~=13.6->batchalign[speaker]) (3.0.0)

Requirement already satisfied: pygments<3.0.0,>=2.13.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from rich~=13.6->batchalign[speaker]) (2.19.1)

Requirement already satisfied: cffi>=1.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from soundfile~=0.12.0->batchalign[speaker]) (1.17.1)

Requirement already satisfied: emoji in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from stanza>=1.9.1->stanza[transformers]>=1.9.1->batchalign[speaker]) (2.14.1)

Requirement already satisfied: protobuf>=3.15.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from stanza>=1.9.1->stanza[transformers]>=1.9.1->batchalign[speaker]) (5.29.3)

Requirement already satisfied: networkx in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from stanza>=1.9.1->stanza[transformers]>=1.9.1->batchalign[speaker]) (3.4.2)

Requirement already satisfied: peft>=0.6.1 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from stanza[transformers]>=1.9.1->batchalign[speaker]) (0.14.0)

Requirement already satisfied: huggingface-hub<1.0,>=0.16.4 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from tokenizers>=0.14.1->batchalign[speaker]) (0.29.1)

Requirement already satisfied: filelock in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from torch->batchalign[speaker]) (3.17.0)

Requirement already satisfied: jinja2 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from torch->batchalign[speaker]) (3.1.5)

Requirement already satisfied: fsspec in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from torch->batchalign[speaker]) (2025.2.0)

Requirement already satisfied: sympy==1.13.1 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from torch->batchalign[speaker]) (1.13.1)

Requirement already satisfied: mpmath<1.4,>=1.1.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from sympy==1.13.1->torch->batchalign[speaker]) (1.3.0)

Requirement already satisfied: pyyaml>=5.1 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from transformers>=4.37->batchalign[speaker]) (6.0.2)

Requirement already satisfied: safetensors>=0.4.1 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from transformers>=4.37->batchalign[speaker]) (0.5.3)

Requirement already satisfied: pycryptodomex>=3.8 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from blobfile->batchalign[speaker]) (3.21.0)

Requirement already satisfied: urllib3<3,>=1.25.3 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from blobfile->batchalign[speaker]) (2.3.0)

Requirement already satisfied: lxml>=4.9 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from blobfile->batchalign[speaker]) (5.3.1)

Requirement already satisfied: coverage<6.0.0,>=5.3.1 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from coverage[toml]<6.0.0,>=5.3.1->eyed3->batchalign[speaker]) (5.5)

Requirement already satisfied: deprecation<3.0.0,>=2.1.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from eyed3->batchalign[speaker]) (2.1.0)

Requirement already satisfied: filetype<2.0.0,>=1.0.7 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from eyed3->batchalign[speaker]) (1.2.0)

Requirement already satisfied: imbalanced-learn in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from imblearn->batchalign[speaker]) (0.13.0)

Requirement already satisfied: docopt>=0.6.2 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from num2words->batchalign[speaker]) (0.6.2)

Requirement already satisfied: pycparser in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from cffi>=1.0->soundfile~=0.12.0->batchalign[speaker]) (2.22)

Requirement already satisfied: toml in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from coverage[toml]<6.0.0,>=5.3.1->eyed3->batchalign[speaker]) (0.10.2)

Requirement already satisfied: mdurl~=0.1 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from markdown-it-py>=2.2.0->rich~=13.6->batchalign[speaker]) (0.1.2)

Requirement already satisfied: psutil in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from peft>=0.6.1->stanza[transformers]>=1.9.1->batchalign[speaker]) (7.0.0)

Requirement already satisfied: accelerate>=0.21.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from peft>=0.6.1->stanza[transformers]>=1.9.1->batchalign[speaker]) (1.4.0)

Requirement already satisfied: charset-normalizer<4,>=2 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from requests<3.0.0,>=2.21.0->rev_ai>=2.18.0->batchalign[speaker]) (3.4.1)

Requirement already satisfied: idna<4,>=2.5 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from requests<3.0.0,>=2.21.0->rev_ai>=2.18.0->batchalign[speaker]) (3.10)

Requirement already satisfied: certifi>=2017.4.17 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from requests<3.0.0,>=2.21.0->rev_ai>=2.18.0->batchalign[speaker]) (2025.1.31)

Requirement already satisfied: threadpoolctl>=3.1.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from scikit-learn!=0.22.0,>=0.16->hmmlearn==0.3.0->batchalign[speaker]) (3.5.0)

Requirement already satisfied: sklearn-compat<1,>=0.1 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from imbalanced-learn->imblearn->batchalign[speaker]) (0.1.3)

Requirement already satisfied: MarkupSafe>=2.0 in c:\users\pc\documents\python_scripts\videotranscript.venv\lib\site-packages (from jinja2->torch->batchalign[speaker]) (3.0.2)

This line seems like it isnt right or something didnt make it to the latest release?

WARNING: batchalign 0.7.16 does not provide the extra 'speaker'

[Noah-Jaffe]

So lets break some rules and install nemo manually.

Run pip install nemo-toolkit[asr] --no-cache

Then running the transcript command gets

Expand for stacktrace

The cache for model files in Transformers v4.22.0 has been updated. Migrating your old cache. This is a one-time only operation. You can interrupt this and resume the migration later on by calling transformers.utils.move_cache().

System.Management.Automation.RemoteException

0it [00:00, ?it/s]

0it [00:00, ?it/s]

[03/02/25 11:07:28] DEBUG    Attempting to create               dispatch.py:134

BatchalignPipeline for CLI...

DEBUG    Initializing packages, got:         dispatch.py:64

packages='['asr', 'speaker']' and

config='{'DEFAULT': <Section:

DEFAULT>, 'asr': <Section: asr>,

'ud': <Section: ud>}'

INFO     Initializing engines...             dispatch.py:85

INFO     -------------------------------     dispatch.py:86

INFO     | asr          |      whisper |    dispatch.py:100

INFO     -------------------------------    dispatch.py:101

DEBUG    Initializing whisper model...      infer_asr.py:67

Device set to use cpu

[03/02/25 11:07:29] DEBUG    Done, initalizing processor and    infer_asr.py:81

config...

DEBUG    Whisper initialization done.       infer_asr.py:83

DEBUG    Initializing utterance model...      whisper.py:46

DEBUG    Done.                                whisper.py:52

INFO     | speaker      | nemo_speaker |    dispatch.py:100

INFO     -------------------------------    dispatch.py:101

INFO     | disfluency   |  replacement |    dispatch.py:100

INFO     -------------------------------    dispatch.py:101

INFO     | retracing    |        ngram |    dispatch.py:100

INFO     -------------------------------    dispatch.py:101

DEBUG    Done initalizing packages.         dispatch.py:136

DEBUG    Successfully created               dispatch.py:137

BatchalignPipeline...

<batchalign.pipelines.pipeline.Bat

chalignPipeline object at

0x000002A67CE64A70>

INFO     Pipeline called with engines:       pipeline.py:81

generator=<batchalign.pipelines.asr

.whisper.WhisperEngine object at

0x000002A67CEFF8C0>,

processors=[<batchalign.pipelines.s

peaker.nemo_speaker.NemoSpeakerEngi

ne object at 0x000002A67E27F500>,

<batchalign.pipelines.cleanup.disfl

uencies.DisfluencyReplacementEngine

object at 0x000002A67F307FB0>,

<batchalign.pipelines.cleanup.retra

ce.NgramRetraceEngine object at

0x000002A67D2D6EA0>], analyzer=None

on input of type <class 'str'>

DEBUG    Transforming input of type: <class  pipeline.py:90

'str'>

DEBUG    Final input type: <class 'str'>    pipeline.py:100

DEBUG    Calling generator:                 pipeline.py:109

<batchalign.pipelines.asr.whisper.

WhisperEngine object at

0x000002A67CEFF8C0>

INFO     Whisper transcribing file...      infer_asr.py:125

DEBUG    Whisper Preprocessing...          infer_asr.py:126

DEBUG    Whisper Transcribing...           infer_asr.py:151

You have passed task=transcribe, but also have set forced_decoder_ids to [[1, None], [2, 50359]] which creates a conflict. forced_decoder_ids will be ignored in favor of task=transcribe.

The attention mask is not set and cannot be inferred from input because pad token is same as eos token. As a consequence, you may observe unexpected behavior. Please pass your input's attention_mask to obtain reliable results.

WhisperModel is using WhisperSdpaAttention, but torch.nn.functional.scaled_dot_product_attention does not support output_attentions=True or layer_head_mask not None. Falling back to the manual attention implementation, but specifying the manual implementation will be required from Transformers version v5.0.0 onwards. This warning can be removed using the argument attn_implementation="eager" when loading the model.

From v4.47 onwards, when a model cache is to be returned, generate will return a Cache instance instead by default (as opposed to the legacy tuple of tuples format). If you want to keep returning the legacy format, please set return_legacy_cache=True.

Whisper did not predict an ending timestamp, which can happen if audio is cut off in the middle of a word. Also make sure WhisperTimeStampLogitsProcessor was used during generation.

[03/02/25 11:07:48] DEBUG    Whisper Postprocessing...         infer_asr.py:184

DEBUG    Whisper Done.                     infer_asr.py:253

DEBUG    Calling processor: processor 1/3,  pipeline.py:122

<batchalign.pipelines.speaker.nemo

speaker.NemoSpeakerEngine object

at 0x000002A67E27F500>

INFO     Running speaker ID on           nemo_speaker.py:40

C:\Users\PC\Documents\python_sc

ripts\videotranscript.tmp_test

\in\clip.mp3 with 2

speaker(s)...

[03/02/25 11:07:49] INFO     Pipeline called with engines:       pipeline.py:81

generator=<batchalign.pipelines.asr

.whisper.WhisperEngine object at

0x000002A67CEFF8C0>,

processors=[<batchalign.pipelines.s

peaker.nemo_speaker.NemoSpeakerEngi

ne object at 0x000002A67E27F500>,

<batchalign.pipelines.cleanup.disfl

uencies.DisfluencyReplacementEngine

object at 0x000002A67F307FB0>,

<batchalign.pipelines.cleanup.retra

ce.NgramRetraceEngine object at

0x000002A67D2D6EA0>], analyzer=None

on input of type <class 'str'>

DEBUG    Transforming input of type: <class  pipeline.py:90

'str'>

DEBUG    Final input type: <class 'str'>    pipeline.py:100

DEBUG    Calling generator:                 pipeline.py:109

<batchalign.pipelines.asr.whisper.

WhisperEngine object at

0x000002A67CEFF8C0>

INFO     Whisper transcribing file...      infer_asr.py:125

DEBUG    Whisper Preprocessing...          infer_asr.py:126

DEBUG    Whisper Transcribing...           infer_asr.py:151

Whisper did not predict an ending timestamp, which can happen if audio is cut off in the middle of a word. Also make sure WhisperTimeStampLogitsProcessor was used during generation.

[03/02/25 11:08:06] DEBUG    Whisper Postprocessing...         infer_asr.py:184

DEBUG    Whisper Done.                     infer_asr.py:253

DEBUG    Calling processor: processor 1/3,  pipeline.py:122

<batchalign.pipelines.speaker.nemo

speaker.NemoSpeakerEngine object

at 0x000002A67E27F500>

INFO     Running speaker ID on           nemo_speaker.py:40

C:\Users\PC\Documents\python_sc

ripts\videotranscript.tmp_test

\in\clip2.mp3 with 2

speaker(s)...

Traceback (most recent call last):

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\cli\dispatch.py", line 162, in dispatch

doc = pipeline(doc,

^^^^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\pipelines\pipeline.py", line 128, in call

doc = p.process(doc, **kwargs)

^^^^^^^^^^^^^^^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\pipelines\speaker\nemo_speaker.py", line 41, in process

speaker_ids = self.model(doc.media.url)

^^^^^^^^^^^^^^^^^^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\models\speaker\infer.py", line 38, in call

from nemo.collections.asr.models.msdd_models import NeuralDiarizer

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr_init.py", line 15, in 

from nemo.collections.asr import data, losses, models, modules

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\models_init.py", line 15, in 

from nemo.collections.asr.models.aed_multitask_models import

EncDecMultiTaskModel

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\models\aed_multitask_models.py", line 32, in 

from nemo.collections.asr.metrics import BLEU, WER

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\metrics_init.py", line 15, in 

from nemo.collections.asr.metrics.bleu import BLEU

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\metrics\bleu.py", line 22, in 

from nemo.collections.asr.parts.submodules.multitask_decoding import

AbstractMultiTaskDecoding

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\parts\submodules\multitask_decoding.py", line 22, in



from nemo.collections.asr.parts.submodules.multitask_beam_decoding import (

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\parts\submodules\multitask_beam_decoding.py", line 21, in



from nemo.collections.asr.modules.transformer import

BeamSearchSequenceGenerator

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\modules_init.py", line 23, in 

from nemo.collections.asr.modules.conformer_encoder import

ConformerEncoder, ConformerEncoderAdapter

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\modules\conformer_encoder.py", line 26, in 

from nemo.collections.asr.models.configs import CacheAwareStreamingConfig

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\models\configs_init.py", line 15, in 

from nemo.collections.asr.models.configs.asr_models_config import (

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\models\configs\asr_models_config.py", line 27, in 

from nemo.core.config import modelPT as model_cfg

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\core\config\modelPT.py", line 22, in 

from nemo.utils import exp_manager

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\utils\exp_manager.py", line 167, in 

class FaultToleranceParams:

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\utils\exp_manager.py", line 176, in FaultToleranceParams

rank_termination_signal: signal.Signals = signal.SIGKILL

^^^^^^^^^^^^^^

AttributeError: module 'signal' has no attribute 'SIGKILL'. Did you mean:

'SIGILL'?
<removed duplicated error trace that was dumped because we are trying to run against 2 files>

Looking into this one, its not a batchalign issue but an issue with windows or something. went in and manually changed the line to be signal.SIGILL.

.venv\Lib\site-packages\nemo\utils\exp_manager.py::L176

rank_termination_signal: signal.Signals = signal.SIGILL #signal.SIGKILL

Then reran again and we are finally getting somewhere

Expand for stacktrace

[03/02/25 11:15:15] DEBUG    Attempting to create               dispatch.py:134

BatchalignPipeline for CLI...

DEBUG    Initializing packages, got:         dispatch.py:64

packages='['asr', 'speaker']' and

config='{'DEFAULT': <Section:

DEFAULT>, 'asr': <Section: asr>,

'ud': <Section: ud>}'

INFO     Initializing engines...             dispatch.py:85

INFO     -------------------------------     dispatch.py:86

INFO     | asr          |      whisper |    dispatch.py:100

INFO     -------------------------------    dispatch.py:101

DEBUG    Initializing whisper model...      infer_asr.py:67

Device set to use cpu

[03/02/25 11:15:16] DEBUG    Done, initalizing processor and    infer_asr.py:81

config...

DEBUG    Whisper initialization done.       infer_asr.py:83

[03/02/25 11:15:17] DEBUG    Initializing utterance model...      whisper.py:46

[03/02/25 11:15:23] DEBUG    Done.                                whisper.py:52

INFO     | speaker      | nemo_speaker |    dispatch.py:100

INFO     -------------------------------    dispatch.py:101

INFO     | disfluency   |  replacement |    dispatch.py:100

INFO     -------------------------------    dispatch.py:101

INFO     | retracing    |        ngram |    dispatch.py:100

INFO     -------------------------------    dispatch.py:101

DEBUG    Done initalizing packages.         dispatch.py:136

DEBUG    Successfully created               dispatch.py:137

BatchalignPipeline...

<batchalign.pipelines.pipeline.Bat

chalignPipeline object at

0x000001A837987D70>

INFO     Pipeline called with engines:       pipeline.py:81

generator=<batchalign.pipelines.asr

.whisper.WhisperEngine object at

0x000001A83597FFB0>,

processors=[<batchalign.pipelines.s

peaker.nemo_speaker.NemoSpeakerEngi

ne object at 0x000001A835FA3B00>,

<batchalign.pipelines.cleanup.disfl

uencies.DisfluencyReplacementEngine

object at 0x000001A836CEF9E0>,

<batchalign.pipelines.cleanup.retra

ce.NgramRetraceEngine object at

0x000001A8385F6BD0>], analyzer=None

on input of type <class 'str'>

DEBUG    Transforming input of type: <class  pipeline.py:90

'str'>

DEBUG    Final input type: <class 'str'>    pipeline.py:100

DEBUG    Calling generator:                 pipeline.py:109

<batchalign.pipelines.asr.whisper.

WhisperEngine object at

0x000001A83597FFB0>

INFO     Whisper transcribing file...      infer_asr.py:125

DEBUG    Whisper Preprocessing...          infer_asr.py:126

DEBUG    Whisper Transcribing...           infer_asr.py:151

You have passed task=transcribe, but also have set forced_decoder_ids to [[1, None], [2, 50359]] which creates a conflict. forced_decoder_ids will be ignored in favor of task=transcribe.

The attention mask is not set and cannot be inferred from input because pad token is same as eos token. As a consequence, you may observe unexpected behavior. Please pass your input's attention_mask to obtain reliable results.

WhisperModel is using WhisperSdpaAttention, but torch.nn.functional.scaled_dot_product_attention does not support output_attentions=True or layer_head_mask not None. Falling back to the manual attention implementation, but specifying the manual implementation will be required from Transformers version v5.0.0 onwards. This warning can be removed using the argument attn_implementation="eager" when loading the model.

From v4.47 onwards, when a model cache is to be returned, generate will return a Cache instance instead by default (as opposed to the legacy tuple of tuples format). If you want to keep returning the legacy format, please set return_legacy_cache=True.

Whisper did not predict an ending timestamp, which can happen if audio is cut off in the middle of a word. Also make sure WhisperTimeStampLogitsProcessor was used during generation.

[03/02/25 11:15:51] DEBUG    Whisper Postprocessing...         infer_asr.py:184

DEBUG    Whisper Done.                     infer_asr.py:253

[03/02/25 11:15:52] DEBUG    Calling processor: processor 1/3,  pipeline.py:122

<batchalign.pipelines.speaker.nemo

_speaker.NemoSpeakerEngine object

at 0x000001A835FA3B00>

INFO     Running speaker ID on           nemo_speaker.py:40

C:\Users\PC\Documents\python_sc

ripts\videotranscript.tmp_test

\in\clip.mp3 with 2

speaker(s)...

[NeMo I 2025-03-02 11:15:53 nemo_logging:393] Loading pretrained diar_msdd_telephonic model from NGC

[NeMo I 2025-03-02 11:15:53 nemo_logging:393] Downloading from: https://api.ngc.nvidia.com/v2/models/nvidia/nemo/diar_msdd_telephonic/versions/1.0.1/files/diar_msdd_telephonic.nemo to C:\Users\PC.cache\torch\NeMo\NeMo_2.3.0rc0\diar_msdd_telephonic\3c3697a0a46f945574fa407149975a13\diar_msdd_telephonic.nemo

[NeMo I 2025-03-02 11:15:55 nemo_logging:393] Instantiating model from pre-trained checkpoint

[NeMo W 2025-03-02 11:15:56 nemo_logging:405] If you intend to do training or fine-tuning, please call the ModelPT.setup_training_data() method and provide a valid configuration file to setup the train data loader.

Train config :

manifest_filepath: null

emb_dir: null

sample_rate: 16000

num_spks: 2

soft_label_thres: 0.5

labels: null

batch_size: 15

emb_batch_size: 0

shuffle: true
[NeMo W 2025-03-02 11:15:56 nemo_logging:405] If you intend to do validation, please call the ModelPT.setup_validation_data() or ModelPT.setup_multiple_validation_data() method and provide a valid configuration file to setup the validation data loader(s).

Validation config :

manifest_filepath: null

emb_dir: null

sample_rate: 16000

num_spks: 2

soft_label_thres: 0.5

labels: null

batch_size: 15

emb_batch_size: 0

shuffle: false
[NeMo W 2025-03-02 11:15:56 nemo_logging:405] Please call the ModelPT.setup_test_data() or ModelPT.setup_multiple_test_data() method and provide a valid configuration file to setup the test data loader(s).

Test config :

manifest_filepath: null

emb_dir: null

sample_rate: 16000

num_spks: 2

soft_label_thres: 0.5

labels: null

batch_size: 15

emb_batch_size: 0

shuffle: false

seq_eval_mode: false
[NeMo I 2025-03-02 11:15:56 nemo_logging:393] PADDING: 16

[NeMo I 2025-03-02 11:15:56 nemo_logging:393] PADDING: 16

[NeMo I 2025-03-02 11:15:56 nemo_logging:393] Model EncDecDiarLabelModel was successfully restored from C:\Users\PC.cache\torch\NeMo\NeMo_2.3.0rc0\diar_msdd_telephonic\3c3697a0a46f945574fa407149975a13\diar_msdd_telephonic.nemo.

[NeMo I 2025-03-02 11:15:56 nemo_logging:393] PADDING: 16

[NeMo I 2025-03-02 11:15:56 nemo_logging:393] Loading pretrained vad_multilingual_marblenet model from NGC

[NeMo I 2025-03-02 11:15:56 nemo_logging:393] Downloading from: https://api.ngc.nvidia.com/v2/models/nvidia/nemo/vad_multilingual_marblenet/versions/1.10.0/files/vad_multilingual_marblenet.nemo to C:\Users\PC.cache\torch\NeMo\NeMo_2.3.0rc0\vad_multilingual_marblenet\670f425c7f186060b7a7268ba6dfacb2\vad_multilingual_marblenet.nemo

[NeMo I 2025-03-02 11:15:57 nemo_logging:393] Instantiating model from pre-trained checkpoint

[NeMo W 2025-03-02 11:15:57 nemo_logging:405] If you intend to do training or fine-tuning, please call the ModelPT.setup_training_data() method and provide a valid configuration file to setup the train data loader.

Train config :

manifest_filepath: /manifests/ami_train_0.63.json,/manifests/freesound_background_train.json,/manifests/freesound_laughter_train.json,/manifests/fisher_2004_background.json,/manifests/fisher_2004_speech_sampled.json,/manifests/google_train_manifest.json,/manifests/icsi_all_0.63.json,/manifests/musan_freesound_train.json,/manifests/musan_music_train.json,/manifests/musan_soundbible_train.json,/manifests/mandarin_train_sample.json,/manifests/german_train_sample.json,/manifests/spanish_train_sample.json,/manifests/french_train_sample.json,/manifests/russian_train_sample.json

sample_rate: 16000

labels:

- background

- speech

batch_size: 256

shuffle: true

is_tarred: false

tarred_audio_filepaths: null

tarred_shard_strategy: scatter

augmentor:

shift:

prob: 0.5

min_shift_ms: -10.0

max_shift_ms: 10.0

white_noise:

prob: 0.5

min_level: -90

max_level: -46

norm: true

noise:

prob: 0.5

manifest_path: /manifests/noise_0_1_musan_fs.json

min_snr_db: 0

max_snr_db: 30

max_gain_db: 300.0

norm: true

gain:

prob: 0.5

min_gain_dbfs: -10.0

max_gain_dbfs: 10.0

norm: true

num_workers: 16

pin_memory: true
[NeMo W 2025-03-02 11:15:57 nemo_logging:405] If you intend to do validation, please call the ModelPT.setup_validation_data() or ModelPT.setup_multiple_validation_data() method and provide a valid configuration file to setup the validation data loader(s).

Validation config :

manifest_filepath: /manifests/ami_dev_0.63.json,/manifests/freesound_background_dev.json,/manifests/freesound_laughter_dev.json,/manifests/ch120_moved_0.63.json,/manifests/fisher_2005_500_speech_sampled.json,/manifests/google_dev_manifest.json,/manifests/musan_music_dev.json,/manifests/mandarin_dev.json,/manifests/german_dev.json,/manifests/spanish_dev.json,/manifests/french_dev.json,/manifests/russian_dev.json

sample_rate: 16000

labels:

- background

- speech

batch_size: 256

shuffle: false

val_loss_idx: 0

num_workers: 16

pin_memory: true
[NeMo W 2025-03-02 11:15:57 nemo_logging:405] Please call the ModelPT.setup_test_data() or ModelPT.setup_multiple_test_data() method and provide a valid configuration file to setup the test data loader(s).

Test config :

manifest_filepath: null

sample_rate: 16000

labels:

- background

- speech

batch_size: 128

shuffle: false

test_loss_idx: 0
[NeMo I 2025-03-02 11:15:57 nemo_logging:393] PADDING: 16

[NeMo I 2025-03-02 11:15:57 nemo_logging:393] Model EncDecClassificationModel was successfully restored from C:\Users\PC.cache\torch\NeMo\NeMo_2.3.0rc0\vad_multilingual_marblenet\670f425c7f186060b7a7268ba6dfacb2\vad_multilingual_marblenet.nemo.

[NeMo I 2025-03-02 11:15:57 nemo_logging:393] Multiscale Weights: [1, 1, 1, 1, 1]

[NeMo I 2025-03-02 11:15:57 nemo_logging:393] Clustering Parameters: {

"oracle_num_speakers": true,

"max_num_speakers": 8,

"enhanced_count_thres": 80,

"max_rp_threshold": 0.25,

"sparse_search_volume": 30,

"maj_vote_spk_count": false,

"chunk_cluster_count": 50,

"embeddings_per_chunk": 10000

}

[NeMo I 2025-03-02 11:15:57 nemo_logging:393] Number of files to diarize: 1

[NeMo I 2025-03-02 11:15:57 nemo_logging:393] Split long audio file to avoid CUDA memory issue

System.Management.Automation.RemoteException

splitting manifest:   0%|          | 0/1 [00:00<?, ?it/s]

splitting manifest: 100%|##########| 1/1 [00:08<00:00,  8.37s/it]

splitting manifest: 100%|##########| 1/1 [00:08<00:00,  8.37s/it]

[NeMo I 2025-03-02 11:16:05 nemo_logging:393] Perform streaming frame-level VAD

[NeMo I 2025-03-02 11:16:05 nemo_logging:393] Filtered duration for loading collection is  0.00 hours.

[NeMo I 2025-03-02 11:16:05 nemo_logging:393] Dataset successfully loaded with 1 items and total duration provided from manifest is  0.01 hours.

[NeMo I 2025-03-02 11:16:05 nemo_logging:393] # 1 files loaded accounting to # 1 labels

[03/02/25 11:16:05] INFO     Pipeline called with engines:       pipeline.py:81

generator=<batchalign.pipelines.asr

.whisper.WhisperEngine object at

0x000001A83597FFB0>,

processors=[<batchalign.pipelines.s

peaker.nemo_speaker.NemoSpeakerEngi

ne object at 0x000001A835FA3B00>,

<batchalign.pipelines.cleanup.disfl

uencies.DisfluencyReplacementEngine

object at 0x000001A836CEF9E0>,

<batchalign.pipelines.cleanup.retra

ce.NgramRetraceEngine object at

0x000001A8385F6BD0>], analyzer=None

on input of type <class 'str'>

DEBUG    Transforming input of type: <class  pipeline.py:90

'str'>

DEBUG    Final input type: <class 'str'>    pipeline.py:100

DEBUG    Calling generator:                 pipeline.py:109

<batchalign.pipelines.asr.whisper.

WhisperEngine object at

0x000001A83597FFB0>

INFO     Whisper transcribing file...      infer_asr.py:125

DEBUG    Whisper Preprocessing...          infer_asr.py:126

DEBUG    Whisper Transcribing...           infer_asr.py:151

Whisper did not predict an ending timestamp, which can happen if audio is cut off in the middle of a word. Also make sure WhisperTimeStampLogitsProcessor was used during generation.

[03/02/25 11:16:22] DEBUG    Whisper Postprocessing...         infer_asr.py:184

DEBUG    Whisper Done.                     infer_asr.py:253

DEBUG    Calling processor: processor 1/3,  pipeline.py:122

<batchalign.pipelines.speaker.nemo

_speaker.NemoSpeakerEngine object

at 0x000001A835FA3B00>

INFO     Running speaker ID on           nemo_speaker.py:40

C:\Users\PC\Documents\python_sc

ripts\videotranscript.tmp_test

\in\clip2.mp3 with 2

speaker(s)...

[NeMo I 2025-03-02 11:16:22 nemo_logging:393] Loading pretrained diar_msdd_telephonic model from NGC

[NeMo I 2025-03-02 11:16:22 nemo_logging:393] Found existing object C:\Users\PC.cache\torch\NeMo\NeMo_2.3.0rc0\diar_msdd_telephonic\3c3697a0a46f945574fa407149975a13\diar_msdd_telephonic.nemo.

[NeMo I 2025-03-02 11:16:22 nemo_logging:393] Re-using file from: C:\Users\PC.cache\torch\NeMo\NeMo_2.3.0rc0\diar_msdd_telephonic\3c3697a0a46f945574fa407149975a13\diar_msdd_telephonic.nemo

[NeMo I 2025-03-02 11:16:22 nemo_logging:393] Instantiating model from pre-trained checkpoint

[NeMo W 2025-03-02 11:16:23 nemo_logging:405] If you intend to do training or fine-tuning, please call the ModelPT.setup_training_data() method and provide a valid configuration file to setup the train data loader.

Train config :

manifest_filepath: null

emb_dir: null

sample_rate: 16000

num_spks: 2

soft_label_thres: 0.5

labels: null

batch_size: 15

emb_batch_size: 0

shuffle: true
[NeMo W 2025-03-02 11:16:23 nemo_logging:405] If you intend to do validation, please call the ModelPT.setup_validation_data() or ModelPT.setup_multiple_validation_data() method and provide a valid configuration file to setup the validation data loader(s).

Validation config :

manifest_filepath: null

emb_dir: null

sample_rate: 16000

num_spks: 2

soft_label_thres: 0.5

labels: null

batch_size: 15

emb_batch_size: 0

shuffle: false
[NeMo W 2025-03-02 11:16:23 nemo_logging:405] Please call the ModelPT.setup_test_data() or ModelPT.setup_multiple_test_data() method and provide a valid configuration file to setup the test data loader(s).

Test config :

manifest_filepath: null

emb_dir: null

sample_rate: 16000

num_spks: 2

soft_label_thres: 0.5

labels: null

batch_size: 15

emb_batch_size: 0

shuffle: false

seq_eval_mode: false
[NeMo I 2025-03-02 11:16:23 nemo_logging:393] PADDING: 16

[NeMo I 2025-03-02 11:16:23 nemo_logging:393] PADDING: 16

[NeMo I 2025-03-02 11:16:24 nemo_logging:393] Model EncDecDiarLabelModel was successfully restored from C:\Users\PC.cache\torch\NeMo\NeMo_2.3.0rc0\diar_msdd_telephonic\3c3697a0a46f945574fa407149975a13\diar_msdd_telephonic.nemo.

[NeMo I 2025-03-02 11:16:24 nemo_logging:393] PADDING: 16

[NeMo I 2025-03-02 11:16:24 nemo_logging:393] Loading pretrained vad_multilingual_marblenet model from NGC

[NeMo I 2025-03-02 11:16:24 nemo_logging:393] Found existing object C:\Users\PC.cache\torch\NeMo\NeMo_2.3.0rc0\vad_multilingual_marblenet\670f425c7f186060b7a7268ba6dfacb2\vad_multilingual_marblenet.nemo.

[NeMo I 2025-03-02 11:16:24 nemo_logging:393] Re-using file from: C:\Users\PC.cache\torch\NeMo\NeMo_2.3.0rc0\vad_multilingual_marblenet\670f425c7f186060b7a7268ba6dfacb2\vad_multilingual_marblenet.nemo

[NeMo I 2025-03-02 11:16:24 nemo_logging:393] Instantiating model from pre-trained checkpoint

[NeMo W 2025-03-02 11:16:24 nemo_logging:405] If you intend to do training or fine-tuning, please call the ModelPT.setup_training_data() method and provide a valid configuration file to setup the train data loader.

Train config :

manifest_filepath: /manifests/ami_train_0.63.json,/manifests/freesound_background_train.json,/manifests/freesound_laughter_train.json,/manifests/fisher_2004_background.json,/manifests/fisher_2004_speech_sampled.json,/manifests/google_train_manifest.json,/manifests/icsi_all_0.63.json,/manifests/musan_freesound_train.json,/manifests/musan_music_train.json,/manifests/musan_soundbible_train.json,/manifests/mandarin_train_sample.json,/manifests/german_train_sample.json,/manifests/spanish_train_sample.json,/manifests/french_train_sample.json,/manifests/russian_train_sample.json

sample_rate: 16000

labels:

- background

- speech

batch_size: 256

shuffle: true

is_tarred: false

tarred_audio_filepaths: null

tarred_shard_strategy: scatter

augmentor:

shift:

prob: 0.5

min_shift_ms: -10.0

max_shift_ms: 10.0

white_noise:

prob: 0.5

min_level: -90

max_level: -46

norm: true

noise:

prob: 0.5

manifest_path: /manifests/noise_0_1_musan_fs.json

min_snr_db: 0

max_snr_db: 30

max_gain_db: 300.0

norm: true

gain:

prob: 0.5

min_gain_dbfs: -10.0

max_gain_dbfs: 10.0

norm: true

num_workers: 16

pin_memory: true
[NeMo W 2025-03-02 11:16:24 nemo_logging:405] If you intend to do validation, please call the ModelPT.setup_validation_data() or ModelPT.setup_multiple_validation_data() method and provide a valid configuration file to setup the validation data loader(s).

Validation config :

manifest_filepath: /manifests/ami_dev_0.63.json,/manifests/freesound_background_dev.json,/manifests/freesound_laughter_dev.json,/manifests/ch120_moved_0.63.json,/manifests/fisher_2005_500_speech_sampled.json,/manifests/google_dev_manifest.json,/manifests/musan_music_dev.json,/manifests/mandarin_dev.json,/manifests/german_dev.json,/manifests/spanish_dev.json,/manifests/french_dev.json,/manifests/russian_dev.json

sample_rate: 16000

labels:

- background

- speech

batch_size: 256

shuffle: false

val_loss_idx: 0

num_workers: 16

pin_memory: true
[NeMo W 2025-03-02 11:16:24 nemo_logging:405] Please call the ModelPT.setup_test_data() or ModelPT.setup_multiple_test_data() method and provide a valid configuration file to setup the test data loader(s).

Test config :

manifest_filepath: null

sample_rate: 16000

labels:

- background

- speech

batch_size: 128

shuffle: false

test_loss_idx: 0
[NeMo I 2025-03-02 11:16:24 nemo_logging:393] PADDING: 16

[NeMo I 2025-03-02 11:16:24 nemo_logging:393] Model EncDecClassificationModel was successfully restored from C:\Users\PC.cache\torch\NeMo\NeMo_2.3.0rc0\vad_multilingual_marblenet\670f425c7f186060b7a7268ba6dfacb2\vad_multilingual_marblenet.nemo.

[NeMo I 2025-03-02 11:16:24 nemo_logging:393] Multiscale Weights: [1, 1, 1, 1, 1]

[NeMo I 2025-03-02 11:16:24 nemo_logging:393] Clustering Parameters: {

"oracle_num_speakers": true,

"max_num_speakers": 8,

"enhanced_count_thres": 80,

"max_rp_threshold": 0.25,

"sparse_search_volume": 30,

"maj_vote_spk_count": false,

"chunk_cluster_count": 50,

"embeddings_per_chunk": 10000

}

[NeMo I 2025-03-02 11:16:24 nemo_logging:393] Number of files to diarize: 1

[NeMo I 2025-03-02 11:16:24 nemo_logging:393] Split long audio file to avoid CUDA memory issue

System.Management.Automation.RemoteException

splitting manifest:   0%|          | 0/1 [00:00<?, ?it/s]

splitting manifest: 100%|##########| 1/1 [00:00<00:00, 198.06it/s]

[NeMo I 2025-03-02 11:16:24 nemo_logging:393] Perform streaming frame-level VAD

[NeMo I 2025-03-02 11:16:24 nemo_logging:393] Filtered duration for loading collection is  0.00 hours.

[NeMo I 2025-03-02 11:16:24 nemo_logging:393] Dataset successfully loaded with 1 items and total duration provided from manifest is  0.01 hours.

[NeMo I 2025-03-02 11:16:24 nemo_logging:393] # 1 files loaded accounting to # 1 labels

Traceback (most recent call last):

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\cli\dispatch.py", line 162, in _dispatch

doc = pipeline(doc,

^^^^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\pipelines\pipeline.py", line 128, in call

doc = p.process(doc, **kwargs)

^^^^^^^^^^^^^^^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\pipelines\speaker\nemo_speaker.py", line 41, in process

speaker_ids = self.__model(doc.media.url)

^^^^^^^^^^^^^^^^^^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\b

atchalign\models\speaker\infer.py", line 74, in call

msdd_model.diarize()

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\t

orch\utils_contextlib.py", line 116, in decorate_context

return func(*args, **kwargs)

^^^^^^^^^^^^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\models\msdd_models.py", line 1212, in diarize

self.clustering_embedding.prepare_cluster_embs_infer()

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\models\msdd_models.py", line 722, in

prepare_cluster_embs_infer

self.emb_sess_test_dict, self.emb_seq_test, self.clus_test_label_dict, _ =

self.run_clustering_diarizer(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\models\msdd_models.py", line 894, in

run_clustering_diarizer

scores =

self.clus_diar_model.diarize(batch_size=self.cfg_diar_infer.batch_size)

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\models\clustering_diarizer.py", line 424, in diarize

self._perform_speech_activity_detection()

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\models\clustering_diarizer.py", line 319, in

_perform_speech_activity_detection

self._run_vad(manifest_vad_input)

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\models\clustering_diarizer.py", line 212, in _run_vad

for i, test_batch in enumerate(

^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\t

qdm\std.py", line 1169, in iter

for obj in iterable:

^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\t

orch\utils\data\dataloader.py", line 708, in next

data = self._next_data()

^^^^^^^^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\t

orch\utils\data\dataloader.py", line 764, in _next_data

data = self._dataset_fetcher.fetch(index)  # may raise StopIteration

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\t

orch\utils\data_utils\fetch.py", line 52, in fetch

data = [self.dataset[idx] for idx in possibly_batched_index]

~~~~~~~~~~~~^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\data\audio_to_label.py", line 337, in getitem

features = self.featurizer.process(

^^^^^^^^^^^^^^^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\parts\preprocessing\features.py", line 202, in process

audio = AudioSegment.from_file(

^^^^^^^^^^^^^^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\parts\preprocessing\segment.py", line 375, in from_file

return cls(

^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\parts\preprocessing\segment.py", line 181, in init

samples = self._convert_samples_to_float32(samples)

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

emo\collections\asr\parts\preprocessing\segment.py", line 263, in

convert_samples_to_float32

if samples.dtype in np.sctypes['int']:

^^^^^^^^^^

File

"c:\Users\PC\Documents\python_scripts\videotranscript.venv\Lib\site-packages\n

umpy_init.py", line 413, in getattr

raise AttributeError(

AttributeError: np.sctypes was removed in the NumPy 2.0 release. Access

dtypes explicitly instead.
<removed duplicated error trace that was dumped because we are trying to run against 2 files>

fix for this one can be found here NVIDIA-NeMo/NeMo#11508

Run pip install numpy==1.26.4 and rerun transcript command:

rerunning transcript command gives us output!

[Noah-Jaffe]

Summary of fixes for windows device:

• manually install omegaconf
  • pip install omegaconf
• manually install nemo toolkit from github git+https://github.com/NVIDIA/NeMo then install asr stuff nemo-toolkit[asr]
  • pip install git+https://github.com/NVIDIA/NeMo
  • pip install nemo-toolkit[asr]
  • for windows machine, fix the signal.SIGKILL issue:
    • in file .venv\Lib\site-packages\nemo\utils\exp_manager.py, in class FaultToleranceParams, in class field rank_termination_signal change signal.SIGKILL to signal.SIGILL. (probably around line 176)
    • This step should have been fixed since ### fix: correct signal for Windows NVIDIA-NeMo/NeMo#11898 (comment) was merged on Mar 4th, 2025.
• manually change version of numpy to before 2.0
  • pip install numpy==1.26.4

running the command:
batchalign transcribe --lang=eng --whisper --num_speakers=2 --diarize .\inp .\out\ now works with these changes!

[Jemoka]

Yes, thank you! We will update documentation accordingly since Nemo is no longer supported as an additional package for automatic install.