lorenlugosch
commited on
Commit
•
c0375f0
1
Parent(s):
d15cb46
Update README.md
Browse files
README.md
CHANGED
@@ -13,7 +13,18 @@ metrics:
|
|
13 |
|
14 |
# End-to-end SLU model for Timers and Such
|
15 |
|
16 |
-
Attention-based RNN sequence-to-sequence model for Timers and Such trained on the `train-real` subset.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
17 |
|
18 |
#### Referencing SpeechBrain
|
19 |
|
@@ -24,7 +35,7 @@ title = {SpeechBrain},
|
|
24 |
year = {2021},
|
25 |
publisher = {GitHub},
|
26 |
journal = {GitHub repository},
|
27 |
-
howpublished = {
|
28 |
}
|
29 |
```
|
30 |
|
|
|
13 |
|
14 |
# End-to-end SLU model for Timers and Such
|
15 |
|
16 |
+
Attention-based RNN sequence-to-sequence model for [Timers and Such](https://zenodo.org/record/4623772) trained on the `train-real` subset. This model checkpoint achieves 86.7% accuracy on `test-real`.
|
17 |
+
|
18 |
+
The model uses an ASR model trained on LibriSpeech (`speechbrain/asr-crdnn-rnnlm-librispeech`) to extract features from the input audio, then maps these features to an intent and slot labels using a beam search.
|
19 |
+
|
20 |
+
The dataset has four intents: `SetTimer`, `SetAlarm`, `SimpleMath`, and `UnitConversion`. Try testing the model by saying something like "set a timer for 5 minutes" or "what's 32 degrees Celsius in Fahrenheit?"
|
21 |
+
|
22 |
+
You can try the model on the `math.wav` file included here as follows:
|
23 |
+
```
|
24 |
+
from speechbrain.pretrained import EndToEndSLU
|
25 |
+
slu = EndToEndSLU.from_hparams("speechbrain/slu-timers-and-such-direct-librispeech-asr")
|
26 |
+
slu.decode_file("math.wav")
|
27 |
+
```
|
28 |
|
29 |
#### Referencing SpeechBrain
|
30 |
|
|
|
35 |
year = {2021},
|
36 |
publisher = {GitHub},
|
37 |
journal = {GitHub repository},
|
38 |
+
howpublished = {\\\\\\\\\\\\\\\\url{https://github.com/speechbrain/speechbrain}},
|
39 |
}
|
40 |
```
|
41 |
|