vall-e/vall_e/emb/transcribe.py
2024-12-11 20:06:55 -06:00

372 lines
9.4 KiB
Python

"""
# Handles transcribing audio provided through --input-audio
"""
import os
import json
import argparse
import torch
import torchaudio
"""
try:
import whisperx
except Exception as e:
whisperx = None
print(f"Error while querying for whisperx: {str(e)}")
pass
"""
from transformers import pipeline
from functools import cache
from tqdm.auto import tqdm
from pathlib import Path
from ..utils import coerce_dtype
def pad(num, zeroes):
return str(num).zfill(zeroes+1)
def process_items( items, stride=0, stride_offset=0 ):
items = sorted( items )
return items if stride == 0 else [ item for i, item in enumerate( items ) if (i+stride_offset) % stride == 0 ]
# major cringe but should automatically unload models when loading a different one
_cached_models = {
"model": (None, None),
"diarization": (None, None),
"align": (None, None),
}
# yes I can write a decorator to do this
def _load_model(model_name="openai/whisper-large-v3", device="cuda", dtype="float16", language="auto", backend="auto", attention="sdpa"):
cache_key = f'{model_name}:{device}:{dtype}:{language}'
if _cached_models["model"][0] == cache_key:
return _cached_models["model"][1]
del _cached_models["model"]
if not isinstance( dtype, str ):
if dtype == torch.float32:
dtype = "float32"
elif dtype == torch.float16:
dtype = "float16"
elif dtype == torch.bfloat16:
dtype = "bfloat16"
# doesnt support it for some reason
if dtype == "bfloat16":
dtype = "float16"
kwargs = {}
kwargs["compute_type"] = dtype
kwargs["task"] = "transcribe"
kwargs["device"] = device
if language != "auto":
kwargs["language"] = language
if backend == "auto" and whisperx is not None:
backend = "whisperx"
if backend == "whisperx":
model_name = model_name.replace("openai/whisper-", "")
model = whisperx.load_model(model_name, **kwargs)
else:
model = pipeline(
"automatic-speech-recognition",
model=model_name,
torch_dtype=coerce_dtype(dtype),
device=device,
model_kwargs={"attn_implementation": attention},
)
_cached_models["model"] = (cache_key, model)
return model
def _load_diarization_model(device="cuda", backend="auto"):
cache_key = f'{device}'
if _cached_models["diarization"][0] == cache_key:
return _cached_models["diarization"][1]
del _cached_models["diarization"]
if backend == "auto" and whisperx is not None:
backend = "whisperx"
if backend == "whisperx":
model = whisperx.DiarizationPipeline(device=device)
else:
model = None # to do later
_cached_models["diarization"] = (cache_key, model)
return model
def _load_align_model(language, device="cuda", backend="auto"):
cache_key = f'{language}:{device}'
if _cached_models["align"][0] == cache_key:
return _cached_models["align"][1]
del _cached_models["align"]
if backend == "auto" and whisperx is not None:
backend = "whisperx"
if backend == "whisperx":
model = whisperx.load_align_model(language_code=language, device=device)
else:
model = None # to do later
_cached_models["align"] = (cache_key, model)
return model
# yes I can just do a for-loop
def unload_model():
del _cached_models["model"]
del _cached_models["diarization"]
del _cached_models["align"]
_cached_models["model"] = (None, None)
_cached_models["diarization"] = (None, None)
_cached_models["align"] = (None, None)
def transcribe(
audio,
language = "auto",
diarize = False,
batch_size = 16,
verbose=False,
align=True,
**model_kwargs,
):
metadata = {
"segments": [],
"language": "",
"text": "",
"start": 0,
"end": 0,
}
# load requested models
model_kwargs["backend"] = "automatic-speech-recognition"
device = model_kwargs.get("device", "cuda")
model = _load_model(language=language, **model_kwargs)
result = model(
str(audio),
chunk_length_s=30,
batch_size=batch_size,
generate_kwargs={"task": "transcribe", "language": None if language == "auto" else language},
return_timestamps="word" if align else False,
return_language=True,
)
start = 0
end = 0
segments = []
for segment in result["chunks"]:
text = segment["text"]
if "timestamp" in segment:
s, e = segment["timestamp"]
start = min( start, s )
end = max( end, e )
else:
s, e = None, None
if language == "auto":
language = segment["language"]
segments.append({
"start": s,
"end": e,
"text": text,
})
if language != "auto":
metadata["language"] = language
metadata["segments"] = segments
metadata["text"] = result["text"].strip()
metadata["start"] = start
metadata["end"] = end
return metadata
# for backwards compat since it also handles some other things for me
"""
def transcribe_whisperx(
audio,
language = "auto",
diarize = False,
batch_size = 16,
verbose=False,
align=True,
**model_kwargs,
):
metadata = {
"segments": [],
"language": "",
"text": "",
"start": 0,
"end": 0,
}
# load requested models
device = model_kwargs.get("device", "cuda")
model = _load_model(language=language, **model_kwargs)
diarize_model = _load_diarization_model(device=device) if diarize else None
# audio is a path, load it
if isinstance(audio, str) or isinstance(audio, Path):
#audio = load_audio(audio)
audio = whisperx.load_audio(audio)
result = model.transcribe(audio, batch_size=batch_size)
if language == "auto":
language = result["language"]
if align:
align_model, align_model_metadata = _load_align_model(language=language, device=device)
result = whisperx.align(result["segments"], align_model, align_model_metadata, audio, device, return_char_alignments=False)
if diarize_model is not None:
diarize_segments = diarize_model(audio)
result = whisperx.assign_word_speakers(diarize_segments, result)
text = []
start = 0
end = 0
for segment in result["segments"]:
text.append( segment["text"] )
start = min( start, segment["start"] )
end = max( end, segment["end"] )
metadata["language"] = language
metadata["segments"] = result["segments"]
metadata["text"] = " ".join(text).strip()
metadata["start"] = start
metadata["end"] = end
return metadata
"""
def transcribe_batch(
input_audio = "voices",
input_voice = None,
output_metadata = "training/metadata",
model_name = "openai/whisper-large-v3",
skip_existing = True,
diarize = False,
stride = 0,
stride_offset = 0,
batch_size = 16,
device = "cuda",
dtype = "float16",
):
# to-do: make this also prepared from args
language_map = {} # k = group, v = language
ignore_groups = [] # skip these groups
ignore_speakers = [] # skip these speakers
only_groups = [] # only process these groups
only_speakers = [] # only process these speakers
if input_voice is not None:
only_speakers = [input_voice]
for dataset_name in os.listdir(f'./{input_audio}/'):
if not os.path.isdir(f'./{input_audio}/{dataset_name}/'):
continue
if group_name in ignore_groups:
continue
if only_groups and group_name not in only_groups:
continue
for speaker_id in tqdm(process_items(os.listdir(f'./{input_audio}/{dataset_name}/')), desc="Processing speaker"):
if not os.path.isdir(f'./{input_audio}/{dataset_name}/{speaker_id}'):
continue
if speaker_id in ignore_speakers:
continue
if only_speakers and speaker_id not in only_speakers:
continue
outpath = Path(f'./{output_metadata}/{dataset_name}/{speaker_id}/whisper.json')
if outpath.exists():
metadata = json.loads(open(outpath, 'r', encoding='utf-8').read())
else:
os.makedirs(f'./{output_metadata}/{dataset_name}/{speaker_id}/', exist_ok=True)
metadata = {}
for filename in tqdm(os.listdir(f'./{input_audio}/{dataset_name}/{speaker_id}/'), desc=f"Processing speaker: {speaker_id}"):
if skip_existing and filename in metadata:
continue
if ".json" in filename:
continue
inpath = f'./{input_audio}/{dataset_name}/{speaker_id}/{filename}'
if os.path.isdir(inpath):
continue
metadata[filename] = transcribe( inpath, model_name=model_name, diarize=diarize, device=device, dtype=dtype )
open(outpath, 'w', encoding='utf-8').write(json.dumps(metadata))
def main():
parser = argparse.ArgumentParser()
parser.add_argument("--input-audio", type=str, default="voices")
parser.add_argument("--input-voice", type=str, default=None)
parser.add_argument("--output-metadata", type=str, default="training/metadata")
parser.add_argument("--model-name", type=str, default="openai/whisper-large-v3")
parser.add_argument("--skip-existing", action="store_true")
parser.add_argument("--diarize", action="store_true")
parser.add_argument("--batch-size", type=int, default=16)
parser.add_argument("--stride", type=int, default=0)
parser.add_argument("--stride-offset", type=int, default=0)
parser.add_argument("--device", type=str, default="cuda")
parser.add_argument("--dtype", type=str, default="bfloat16")
parser.add_argument("--amp", action="store_true")
# parser.add_argument("--raise-exceptions", action="store_true")
args = parser.parse_args()
# do some assumption magic
# to-do: find a nice way to spawn multiple processes where tqdm plays nicely
if args.device.isnumeric():
args.stride = torch.cuda.device_count()
args.stride_offset = int(args.device)
args.device = f'cuda:{args.device}'
transcribe_batch(
input_audio = args.input_audio,
input_voice = args.input_voice,
output_metadata = args.output_metadata,
model_name = args.model_name,
skip_existing = args.skip_existing,
diarize = args.diarize,
stride = args.stride,
stride_offset = args.stride_offset,
batch_size = args.batch_size,
device = args.device,
dtype = args.dtype,
)
if __name__ == "__main__":
main()