Whisper Large V2 Cv11 German

🚀 用於德語自動語音識別的微調版whisper-large-v2模型

本模型是 openai/whisper-large-v2 的微調版本，在 mozilla-foundation/common_voice_11_0 德語數據集上進行了訓練。使用該模型時，請確保語音輸入的採樣率為 16Khz。該模型還可以預測大小寫和標點符號。

📄 許可證

本項目採用 Apache-2.0 許可證。

📊 模型信息

✨ 主要特性

• 基於 openai/whisper-large-v2 進行微調，針對德語自動語音識別任務進行了優化。
• 能夠預測語音中的大小寫和標點符號，提供更自然的文本輸出。

📈 性能表現

預訓練模型在 Common Voice 9.0 上的詞錯誤率（WER）

微調模型在 Common Voice 11.0 上的詞錯誤率（WER）

💻 使用示例

基礎用法

使用 🤗 Pipeline 進行推理

import torch

from datasets import load_dataset
from transformers import pipeline

device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")

# Load pipeline
pipe = pipeline("automatic-speech-recognition", model="bofenghuang/whisper-large-v2-cv11-german", device=device)

# NB: set forced_decoder_ids for generation utils
pipe.model.config.forced_decoder_ids = pipe.tokenizer.get_decoder_prompt_ids(language="de", task="transcribe")

# Load data
ds_mcv_test = load_dataset("mozilla-foundation/common_voice_11_0", "de", split="test", streaming=True)
test_segment = next(iter(ds_mcv_test))
waveform = test_segment["audio"]

# NB: decoding option
# limit the maximum number of generated tokens to 225
pipe.model.config.max_length = 225 + 1
# sampling
# pipe.model.config.do_sample = True
# beam search
# pipe.model.config.num_beams = 5
# return
# pipe.model.config.return_dict_in_generate = True
# pipe.model.config.output_scores = True
# pipe.model.config.num_return_sequences = 5

# Run
generated_sentences = pipe(waveform)["text"]

高級用法

使用 🤗 底層 API 進行推理

import torch
import torchaudio

from datasets import load_dataset
from transformers import AutoProcessor, AutoModelForSpeechSeq2Seq

device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")

# Load model
model = AutoModelForSpeechSeq2Seq.from_pretrained("bofenghuang/whisper-large-v2-cv11-german").to(device)
processor = AutoProcessor.from_pretrained("bofenghuang/whisper-large-v2-cv11-german", language="german", task="transcribe")

# NB: set forced_decoder_ids for generation utils
model.config.forced_decoder_ids = processor.get_decoder_prompt_ids(language="de", task="transcribe")

# 16_000
model_sample_rate = processor.feature_extractor.sampling_rate

# Load data
ds_mcv_test = load_dataset("mozilla-foundation/common_voice_11_0", "de", split="test", streaming=True)
test_segment = next(iter(ds_mcv_test))
waveform = torch.from_numpy(test_segment["audio"]["array"])
sample_rate = test_segment["audio"]["sampling_rate"]

# Resample
if sample_rate != model_sample_rate:
    resampler = torchaudio.transforms.Resample(sample_rate, model_sample_rate)
    waveform = resampler(waveform)

# Get feat
inputs = processor(waveform, sampling_rate=model_sample_rate, return_tensors="pt")
input_features = inputs.input_features
input_features = input_features.to(device)

# Generate
generated_ids = model.generate(inputs=input_features, max_new_tokens=225)  # greedy
# generated_ids = model.generate(inputs=input_features, max_new_tokens=225, num_beams=5)  # beam search

# Detokenize
generated_sentences = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]

# Normalise predicted sentences if necessary