Skip to content
This repository has been archived by the owner on Oct 9, 2023. It is now read-only.

Update Speech Recognition example (with learning_rate) #1272

Closed
wants to merge 4 commits into from
Closed

Update Speech Recognition example (with learning_rate) #1272

Changes from 1 commit
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
6c3b5ff
Minor fix with the example, add lr
krshrimali Apr 4, 2022
f1a55e0
Merge branch 'master' into example_fix/speech-recog
mergify[bot] Apr 4, 2022
4b1f72b
Update flash_examples/speech_recognition.py
krshrimali Apr 5, 2022
6ee6570
Merge branch 'master' into example_fix/speech-recog
mergify[bot] Apr 6, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 5 additions & 4 deletions flash_examples/speech_recognition.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,16 +29,17 @@
)

# 2. Build the task
model = SpeechRecognition(backbone="facebook/wav2vec2-base-960h")
model = SpeechRecognition(backbone="facebook/wav2vec2-base-960h", learning_rate=1e-5)

# 3. Create the trainer and finetune the model
trainer = flash.Trainer(max_epochs=1, gpus=torch.cuda.device_count())
trainer.finetune(model, datamodule=datamodule, strategy="no_freeze")

# 4. Predict on audio files!
datamodule = SpeechRecognitionData.from_files(predict_files=["data/timit/example.wav"], batch_size=4)
predictions = trainer.predict(model, datamodule=datamodule)
print(predictions)
predict_datamodule = SpeechRecognitionData.from_files(predict_files=["data/timit/example.wav"], batch_size=4)
predictions = trainer.predict(model, datamodule=predict_datamodule)
krshrimali marked this conversation as resolved.
Show resolved Hide resolved

print("Predictions: ", predictions)
ethanwharris marked this conversation as resolved.
Show resolved Hide resolved

# 5. Save the model!
trainer.save_checkpoint("speech_recognition_model.pt")