|
--- |
|
language: [mr,hi] |
|
datasets: |
|
- openslr |
|
- interspeech_2021_asr |
|
metrics: |
|
- wer |
|
tags: |
|
- audio |
|
- automatic-speech-recognition |
|
- speech |
|
- xlsr-fine-tuning-week |
|
- hindi |
|
- marathi |
|
license: apache-2.0 |
|
model-index: |
|
- name: XLSR Wav2Vec2 Large 53 Hindi-Marathi by Tanmay Laud |
|
results: |
|
- task: |
|
name: Speech Recognition |
|
type: automatic-speech-recognition |
|
dataset: |
|
name: OpenSLR hi, OpenSLR mr |
|
type: openslr, interspeech_2021_asr |
|
metrics: |
|
- name: Test WER |
|
type: wer |
|
value: 23.736641 |
|
--- |
|
|
|
# Wav2Vec2-Large-XLSR-53-Hindi-Marathi |
|
Fine-tuned facebook/wav2vec2-large-xlsr-53 on Hindi and Marathi using the OpenSLR SLR64 datasets. When using this model, make sure that your speech input is sampled at 16kHz. |
|
|
|
## Installation |
|
```bash |
|
pip install git+https://github.com/huggingface/transformers.git datasets librosa torch==1.7.0 torchaudio==0.7.0 jiwer |
|
``` |
|
|
|
## Eval dataset: |
|
```bash |
|
wget https://www.openslr.org/resources/103/Marathi_test.zip -P data/marathi |
|
unzip -P "K3[2?do9" data/marathi/Marathi_test.zip -d data/marathi/. |
|
tar -xzf data/marathi/Marathi_test.tar.gz -C data/marathi/. |
|
wget https://www.openslr.org/resources/103/Hindi_test.zip -P data/hindi |
|
unzip -P "w9I2{3B*" data/hindi/Hindi_test.zip -d data/hindi/. |
|
tar -xzf data/hindi/Hindi_test.tar.gz -C data/hindi/. |
|
wget -O test.csv 'https://filebin.net/snrz6bt13usv8w2e/test_large.csv?t=ps3n99ho' |
|
#If download does not work, paste this link in browser: https://filebin.net/snrz6bt13usv8w2e/test_large.csv |
|
``` |
|
## Usage |
|
The model can be used directly (without a language model) as follows, assuming you have a dataset with Marathi text and path fields: |
|
|
|
|
|
|
|
```python |
|
import torch |
|
import torchaudio |
|
import librosa |
|
from datasets import load_dataset |
|
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor |
|
|
|
from datasets import load_metric, Dataset |
|
from transformers import Wav2Vec2Processor, Wav2Vec2ForCTC |
|
|
|
wer = load_metric("wer") |
|
processor = Wav2Vec2Processor.from_pretrained('tanmaylaud/wav2vec2-large-xlsr-hindi-marathi') |
|
model = Wav2Vec2ForCTC.from_pretrained('tanmaylaud/wav2vec2-large-xlsr-hindi-marathi').to("cuda") |
|
|
|
# Preprocessing the datasets. |
|
# We need to read the audio files as arrays |
|
def speech_file_to_array_fn(batch): |
|
batch["sentence"] = re.sub(chars_to_ignore_regex, '', batch["sentence"]) |
|
speech_array, sampling_rate = torchaudio.load(batch["path"]) |
|
batch["speech"] = speech_array[0].numpy() |
|
batch["sampling_rate"] = sampling_rate |
|
batch["target_text"] = batch["sentence"] |
|
batch["speech"] = librosa.resample(np.asarray(batch["speech"]), sampling_rate, 16_000) |
|
batch["sampling_rate"] = 16_000 |
|
return batch |
|
|
|
test_data= test_data.map(speech_file_to_array_fn) |
|
inputs = processor(test_data["speech"][:2], sampling_rate=16_000, return_tensors="pt", padding=True) |
|
|
|
with torch.no_grad(): |
|
logits = model(inputs.input_values, attention_mask=inputs.attention_mask).logits |
|
|
|
predicted_ids = torch.argmax(logits, dim=-1) |
|
|
|
print("Prediction:", processor.batch_decode(predicted_ids)) |
|
print("Reference:", test_data["text"][:2]) |
|
``` |
|
|
|
|
|
|
|
# Code For Evaluation on OpenSLR (Hindi + Marathi : https://filebin.net/snrz6bt13usv8w2e/test_large.csv) |
|
```python |
|
import torchaudio |
|
import torch |
|
import librosa |
|
import numpy as np |
|
import re |
|
|
|
test = Dataset.from_csv('test.csv') |
|
|
|
|
|
chars_to_ignore_regex = '[\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\,\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\?\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\.\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\!\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\-\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\;\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\:\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\"\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\“\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\%\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\‘\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\”\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\�\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\।]' |
|
|
|
# Preprocessing the datasets. |
|
# We need to read the audio files as arrays |
|
def speech_file_to_array_fn(batch): |
|
batch["sentence"] = re.sub(chars_to_ignore_regex, '', batch["sentence"]) |
|
speech_array, sampling_rate = torchaudio.load(batch["path"]) |
|
batch["speech"] = speech_array[0].numpy() |
|
batch["sampling_rate"] = sampling_rate |
|
batch["target_text"] = batch["sentence"] |
|
batch["speech"] = librosa.resample(np.asarray(batch["speech"]), sampling_rate, 16_000) |
|
batch["sampling_rate"] = 16_000 |
|
return batch |
|
|
|
test= test.map(speech_file_to_array_fn) |
|
|
|
# Preprocessing the datasets. |
|
# We need to read the audio files as arrays |
|
def evaluate(batch): |
|
inputs = processor(batch["speech"], sampling_rate=16_000, return_tensors="pt", padding=True) |
|
with torch.no_grad(): |
|
logits = model(inputs.input_values.to("cuda"), attention_mask=inputs.attention_mask.to("cuda")).logits |
|
pred_ids = torch.argmax(logits, dim=-1) |
|
# we do not want to group tokens when computing the metrics |
|
batch["pred_strings"] = processor.batch_decode(pred_ids) |
|
return batch |
|
|
|
test = test.map(evaluate, batched=True, batch_size=32) |
|
print("WER: {:2f}".format(100 * wer.compute(predictions=test["pred_strings"], references=test["sentence"]))) |
|
``` |
|
|
|
|
|
|
|
#### Code for Evaluation on Common Voice Hindi (Common voice does not have Marathi yet) |
|
```python |
|
import torchaudio |
|
import torch |
|
import librosa |
|
import numpy as np |
|
import re |
|
from datasets import load_metric, load_dataset, Dataset |
|
from transformers import Wav2Vec2Processor, Wav2Vec2ForCTC |
|
|
|
wer = load_metric("wer") |
|
processor = Wav2Vec2Processor.from_pretrained('tanmaylaud/wav2vec2-large-xlsr-hindi-marathi') |
|
model = Wav2Vec2ForCTC.from_pretrained('tanmaylaud/wav2vec2-large-xlsr-hindi-marathi').to("cuda") |
|
|
|
chars_to_ignore_regex = '[\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\,\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\?\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\.\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\!\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\-\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\;\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\:\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\"\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\“\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\%\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\‘\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\”\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\�\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\।]' |
|
|
|
# Preprocessing the datasets. |
|
# We need to read the audio files as arrays |
|
def speech_file_to_array_fn(batch): |
|
batch["sentence"] = re.sub(chars_to_ignore_regex, '', batch["sentence"]) |
|
speech_array, sampling_rate = torchaudio.load(batch["path"]) |
|
batch["speech"] = speech_array[0].numpy() |
|
batch["sampling_rate"] = sampling_rate |
|
batch["target_text"] = batch["sentence"] |
|
batch["speech"] = librosa.resample(np.asarray(batch["speech"]), sampling_rate, 16_000) |
|
batch["sampling_rate"] = 16_000 |
|
return batch |
|
|
|
#Run prediction on batch |
|
def evaluate(batch): |
|
inputs = processor(batch["speech"], sampling_rate=16_000, return_tensors="pt", padding=True) |
|
with torch.no_grad(): |
|
logits = model(inputs.input_values.to("cuda"), attention_mask=inputs.attention_mask.to("cuda")).logits |
|
pred_ids = torch.argmax(logits, dim=-1) |
|
# we do not want to group tokens when computing the metrics |
|
batch["pred_strings"] = processor.batch_decode(pred_ids) |
|
return batch |
|
|
|
|
|
test_data = load_dataset("common_voice", "hi", split="test") |
|
test_data = test_data.map(speech_file_to_array_fn) |
|
test_data = test_data.map(evaluate, batched=True, batch_size=32) |
|
print("WER: {:2f}".format(100 * wer.compute(predictions=test_data["pred_strings"], |
|
references=test_data["sentence"]))) |
|
``` |
|
|
|
Link to eval notebook : https://colab.research.google.com/drive/1nZRTgKfxCD9cvy90wikTHkg2il3zgcqW#scrollTo=cXWFbhb0d7DT |
|
|
|
WER : 23.736641% (OpenSLR Hindi+Marathi Test set : https://filebin.net/snrz6bt13usv8w2e/test_large.csv) |
|
|
|
|
|
WER: 44.083527% (Common Voice Hindi Test Split) |
