bofenghuang's picture
up
cee265c
metadata
license: apache-2.0
language: de
library_name: transformers
thumbnail: null
tags:
  - automatic-speech-recognition
  - whisper-event
datasets:
  - mozilla-foundation/common_voice_11_0
metrics:
  - wer
model-index:
  - name: Fine-tuned whisper-large-v2 model for ASR in German
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Common Voice 11.0
          type: mozilla-foundation/common_voice_11_0
          config: de
          split: test
          args: de
        metrics:
          - name: WER (Greedy)
            type: wer
            value: 5.76

Model architecture Model size Language

Fine-tuned whisper-large-v2 model for ASR in German

This model is a fine-tuned version of openai/whisper-large-v2, trained on the mozilla-foundation/common_voice_11_0 de dataset. When using the model make sure that your speech input is also sampled at 16Khz. This model also predicts casing and punctuation.

Performance

Below are the WERs of the pre-trained models on the Common Voice 9.0. These results are reported in the original paper.

Below are the WERs of the fine-tuned models on the Common Voice 11.0.

Usage

Inference with 🤗 Pipeline

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"]

Inference with 🤗 low-level APIs

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