Update app.py
Browse files
app.py
CHANGED
|
@@ -13,6 +13,8 @@ asr_pipe = pipeline(
|
|
| 13 |
"automatic-speech-recognition",
|
| 14 |
model="GatinhoEducado/whisper-tiny-finetuned-minds14",
|
| 15 |
device=device,
|
|
|
|
|
|
|
| 16 |
tokenizer = tokenizer,
|
| 17 |
generate_kwargs = {"language":"<|pt|>",
|
| 18 |
"task": "transcribe",
|
|
@@ -32,7 +34,7 @@ speaker_embeddings = torch.tensor(embeddings_dataset[7306]["xvector"]).unsqueeze
|
|
| 32 |
|
| 33 |
def translate(audio):
|
| 34 |
#outputs = asr_pipe(audio, max_new_tokens=256, generate_kwargs={"task": "translate"})
|
| 35 |
-
outputs = asr_pipe(audio, max_new_tokens=
|
| 36 |
return outputs["text"]
|
| 37 |
|
| 38 |
|
|
|
|
| 13 |
"automatic-speech-recognition",
|
| 14 |
model="GatinhoEducado/whisper-tiny-finetuned-minds14",
|
| 15 |
device=device,
|
| 16 |
+
max_new_tokens=60,
|
| 17 |
+
chunk_length_s=0,
|
| 18 |
tokenizer = tokenizer,
|
| 19 |
generate_kwargs = {"language":"<|pt|>",
|
| 20 |
"task": "transcribe",
|
|
|
|
| 34 |
|
| 35 |
def translate(audio):
|
| 36 |
#outputs = asr_pipe(audio, max_new_tokens=256, generate_kwargs={"task": "translate"})
|
| 37 |
+
outputs = asr_pipe(audio, max_new_tokens=256)
|
| 38 |
return outputs["text"]
|
| 39 |
|
| 40 |
|