speech-seq2seq
/

wav2vec2-2-bert-large

Automatic Speech Recognition

speech-encoder-decoder

generated_from_trainer

Inference Endpoints

Model card Files Files and versions Metrics Training metrics Community

sanchit-gandhi HF staff commited on Feb 10, 2022

Commit

5377a51

•

1 Parent(s): bca4b98

update model card README.md

Files changed (1) hide show

README.md +13 -13

README.md CHANGED Viewed

@@ -15,8 +15,8 @@ should probably proofread and complete it, then remove this comment. -->
 This model was trained from scratch on the librispeech_asr dataset.
 It achieves the following results on the evaluation set:
-- Loss: nan
-- Wer: 2.0097
 ## Model description
@@ -35,7 +35,7 @@ More information needed
 ### Training hyperparameters
 The following hyperparameters were used during training:
-- learning_rate: 0.0003
 - train_batch_size: 8
 - eval_batch_size: 8
 - seed: 42
@@ -51,16 +51,16 @@ The following hyperparameters were used during training:
 | Training Loss | Epoch | Step | Validation Loss | Wer    |
 |:-------------:|:-----:|:----:|:---------------:|:------:|
-| 6.9015        | 0.28  | 500  | 7.0241          | 1.0    |
-| 6.7728        | 0.56  | 1000 | 6.9759          | 1.0    |
-| 6.5915        | 0.84  | 1500 | 6.9615          | 2.0097 |
-| 0.0           | 1.12  | 2000 | nan             | 2.0097 |
-| 0.0           | 1.4   | 2500 | nan             | 2.0097 |
-| 0.0           | 1.68  | 3000 | nan             | 2.0097 |
-| 0.0           | 1.96  | 3500 | nan             | 2.0097 |
-| 0.0           | 2.24  | 4000 | nan             | 2.0097 |
-| 0.0           | 2.52  | 4500 | nan             | 2.0097 |
-| 0.0           | 2.8   | 5000 | nan             | 2.0097 |
 ### Framework versions

 This model was trained from scratch on the librispeech_asr dataset.
 It achieves the following results on the evaluation set:
+- Loss: 6.9670
+- Wer: 1.9878
 ## Model description
 ### Training hyperparameters
 The following hyperparameters were used during training:
+- learning_rate: 0.0001
 - train_batch_size: 8
 - eval_batch_size: 8
 - seed: 42
 | Training Loss | Epoch | Step | Validation Loss | Wer    |
 |:-------------:|:-----:|:----:|:---------------:|:------:|
+| 6.7599        | 0.28  | 500  | 6.8755          | 1.2551 |
+| 6.5943        | 0.56  | 1000 | 6.7702          | 1.5878 |
+| 6.3146        | 0.84  | 1500 | 6.6981          | 1.6627 |
+| 6.6112        | 1.12  | 2000 | 6.6760          | 1.9853 |
+| 6.6894        | 1.4   | 2500 | 6.6323          | 1.9376 |
+| 6.5525        | 1.68  | 3000 | 6.6185          | 1.9383 |
+| 6.571         | 1.96  | 3500 | 6.6126          | 1.9580 |
+| 6.3363        | 2.24  | 4000 | 6.7869          | 1.9818 |
+| 6.5832        | 2.52  | 4500 | 6.9096          | 2.0025 |
+| 6.3523        | 2.8   | 5000 | 6.9670          | 1.9878 |
 ### Framework versions