Instructions to use TriLinSL/w2v-bert-2.0-sinhala-30h with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use TriLinSL/w2v-bert-2.0-sinhala-30h with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("automatic-speech-recognition", model="TriLinSL/w2v-bert-2.0-sinhala-30h")# pip install -U transformers accelerate # Load model directly from transformers import AutoProcessor, AutoModelForCTC processor = AutoProcessor.from_pretrained("TriLinSL/w2v-bert-2.0-sinhala-30h") model = AutoModelForCTC.from_pretrained("TriLinSL/w2v-bert-2.0-sinhala-30h", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Commit History
-
final model | val (no punct) WER 28.65 CER 5.96
6211787 -
Model save
ec34167 -
Training in progress, step 4638, checkpoint
b84bcf8 -
Training in progress, step 4638
43bcf28 -
Training in progress, step 4500, checkpoint
0f81767 -
Training in progress, step 4500
cdacb67 -
best model: step 4500, eval_wer_np 29.00
5632742 -
Training in progress, step 4000, checkpoint
e117e59 -
best model: step 4000, eval_wer_np 29.49
5fa5687 -
Training in progress, step 3500, checkpoint
054bfe8 -
best model: step 3500, eval_wer_np 31.62
34f5c4e -
Training in progress, step 3000, checkpoint
5e2307f -
best model: step 3000, eval_wer_np 32.94
9528aa1 -
Training in progress, step 2500, checkpoint
2e6a0c6 -
best model: step 2500, eval_wer_np 35.85
f344c5a -
Training in progress, step 2000, checkpoint
4b6a9c7 -
best model: step 2000, eval_wer_np 37.08
379320b -
best model: step 1500, eval_wer_np 40.13
942d8c0 -
Training in progress, step 1500, checkpoint
efa38e9 -
best model: step 1500, eval_wer_np 40.13
651e942 -
Training in progress, step 1000, checkpoint
2f53d3d -
Training in progress, step 1000
2304adf -
best model: step 1000, eval_wer_np 43.47
efa5fe1 -
Training in progress, step 1000, checkpoint
b8b8d95 -
best model: step 1000, eval_wer_np 43.47
e5323c2 -
Training in progress, step 500, checkpoint
28a0159 -
Training in progress, step 500
cd1caeb -
best model: step 500, eval_wer_np 53.02
15cc4c7 -
Upload subset_fingerprint.txt with huggingface_hub
4500fb9 -
remove smoke test
ededa10 -
smoke test (temporary)
834439b -
remove smoke test
bedaca8 -
smoke test (temporary)
a9eca4e -
remove smoke test
aac3113 -
smoke test (temporary)
01601ae
-
remove smoke test
156a257 -
smoke test (temporary)
6e65bbb -
initial commit
41e9bfa