ceyda
/
wav2vec2-base-760-turkish
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fairseq tuned_wer

This commit is contained in:
ceyda 2021-04-02 19:00:44 +00:00
parent 3d0bc0ac66
commit ab8048eec0
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
README.md
View File

@ -23,7 +23,7 @@ model-index:
metrics: metrics:
- name: Test WER - name: Test WER
type: wer type: wer
value: 27.08 value: 24.91
--- ---
# Wav2Vec2-Base-760-Turkish # Wav2Vec2-Base-760-Turkish
@ -109,7 +109,7 @@ def evaluate(batch):
logits = model(inputs.input_values.to("cuda"), attention_mask=inputs.attention_mask.to("cuda")).logits logits = model(inputs.input_values.to("cuda"), attention_mask=inputs.attention_mask.to("cuda")).logits
pred_ids = torch.argmax(logits, dim=-1) pred_ids = torch.argmax(logits, dim=-1)
batch["pred_strings"] = processor.batch_decode(pred_ids) batch["pred_strings"] = processor.batch_decode(pred_ids,skip_special_tokens=True)
return batch return batch
result = test_dataset.map(evaluate, batched=True, batch_size=8) result = test_dataset.map(evaluate, batched=True, batch_size=8)
@ -117,7 +117,7 @@ result = test_dataset.map(evaluate, batched=True, batch_size=8)
print("WER: {:2f}".format(100 * wer.compute(predictions=result["pred_strings"], references=result["sentence"]))) print("WER: {:2f}".format(100 * wer.compute(predictions=result["pred_strings"], references=result["sentence"])))
``` ```
**Test Result**: 27.08 % (in progress) **Test Result**: 24.91 % (in progress)
## Training ## Training