Maniac commited on
Commit
d147c6c
1 Parent(s): b4e891f

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +0 -22
README.md CHANGED
@@ -79,25 +79,3 @@ python eval.py --model_id Maniac/wav2vec2-xls-r-urdu --dataset mozilla-foundatio
79
  ```bash
80
  python eval.py --model_id Maniac/wav2vec2-xls-r-urdu --dataset speech-recognition-community-v2/dev_data --config ur --split validation --chunk_length_s 5.0 --stride_length_s 1.0
81
  ```
82
-
83
- ### Inference With LM
84
-
85
- ```python
86
- import torch
87
- from datasets import load_dataset
88
- from transformers import AutoModelForCTC, AutoProcessor
89
- import torchaudio.functional as F
90
-
91
- model_id = "Maniac/wav2vec2-xls-r-urdu"
92
-
93
- sample_iter = iter(load_dataset("mozilla-foundation/common_voice_7_0", "ur", split="test", streaming=True, use_auth_token=True))
94
- sample = next(sample_iter)
95
- resampled_audio = F.resample(torch.tensor(sample["audio"]["array"]), 48_000, 16_000).numpy()
96
- model = AutoModelForCTC.from_pretrained(model_id)
97
- processor = AutoProcessor.from_pretrained(model_id)
98
- input_values = processor(resampled_audio, return_tensors="pt").input_values
99
-
100
- with torch.no_grad():
101
- logits = model(input_values).logits
102
- transcription = processor.batch_decode(logits.numpy()).text
103
- ```
 
79
  ```bash
80
  python eval.py --model_id Maniac/wav2vec2-xls-r-urdu --dataset speech-recognition-community-v2/dev_data --config ur --split validation --chunk_length_s 5.0 --stride_length_s 1.0
81
  ```