kotoba-tech

/

kotoba-whisper-v2.1

like 19

Follow

Kotoba Technologies 141

issue

when i use chunk_length_s=15. it always return error. how to solove.tks

code

`import torch
from transformers import pipeline
from datasets import load_dataset
import numpy as np

model_id = "kotoba-tech/kotoba-whisper-v2.1"
torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
device = "cuda:0" if torch.cuda.is_available() else "cpu"
model_kwargs = {"attn_implementation": "sdpa"} if torch.cuda.is_available() else {}
generate_kwargs = {"language": "ja", "task": "transcribe"}

pipe = pipeline(
model=model_id,
torch_dtype=torch_dtype,
device=device,
model_kwargs=model_kwargs,
batch_size=16,
trust_remote_code=True,
punctuator=True,
)

dataset = load_dataset("japanese-asr/ja_asr.reazonspeech_test", split="test")
sample = dataset[0]["audio"]
sample = {"array": np.concatenate([i["array"] for i in dataset[:20]["audio"]]), "sampling_rate": dataset[0]['audio']['sampling_rate']}

result = pipe(
r"./myoutput.wav",
chunk_length_s=15,
return_timestamps=True,
generate_kwargs=generate_kwargs
)`

Error

Traceback (most recent call last):
File "f:\Pydata\DL_LLM\Hgface\transcribe.py", line 43, in
result = pipe(
^^^^^
File "D:\Pydata\anacondai\envs\hgface\Lib\site-packages\transformers\pipelines\automatic_speech_recognition.py", line 295, in call
return super().call(inputs, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\Pydata\anacondai\envs\hgface\Lib\site-packages\transformers\pipelines\base.py", line 1423, in call
return next(
^^^^^
File "D:\Pydata\anacondai\envs\hgface\Lib\site-packages\transformers\pipelines\pt_utils.py", line 124, in next
item = next(self.iterator)
^^^^^^^^^^^^^^^^^^^
File "D:\Pydata\anacondai\envs\hgface\Lib\site-packages\transformers\pipelines\pt_utils.py", line 269, in next
processed = self.infer(next(self.iterator), **self.params)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\Pydata\anacondai\envs\hgface\Lib\site-packages\transformers\pipelines\base.py", line 1338, in forward
model_outputs = self._forward(model_inputs, **forward_params)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "F:\models\huggingface\modules\transformers_modules\kotoba-tech\kotoba-whisper-v2.1\57a9d8ab771a0124706b67d22509bedd07c36187\kotoba_whisper.py", line 271, in _forward
tokens = self.model.generate(attention_mask=attention_mask, **generate_kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\Pydata\anacondai\envs\hgface\Lib\site-packages\transformers\models\whisper\generation_whisper.py", line 712, in generate
input_features, cur_bsz, batch_idx_map = self._maybe_reduce_batch(
^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\Pydata\anacondai\envs\hgface\Lib\site-packages\transformers\models\whisper\generation_whisper.py", line 1727, in _maybe_reduce_batch
input_features = torch.cat([input_features[:cut_index], input_features[cut_index + 1 :]], dim=0)
~~~~~~~~~~~~~~^^^^^^^^^^^^
TypeError: 'NoneType' object is not subscriptable