update model

README.md

@@ -32,7 +32,7 @@ model-index:
     metrics:
     - name: Test WER
       type: wer
-      value: 2.26
+      value: 2.27
   - task:
       name: Automatic Speech Recognition
       type: automatic-speech-recognition
@@ -46,7 +46,7 @@ model-index:
     metrics:
     - name: Test WER
       type: wer
-      value: 5.52
+      value: 5.53
 ---
 
 <iframe src="https://ghbtns.com/github-btn.html?user=speechbrain&repo=speechbrain&type=star&count=true&size=large&v=2" frameborder="0" scrolling="0" width="170" height="30" title="GitHub"></iframe>
@@ -62,7 +62,7 @@ The performance of the model is the following:
 
 | Release | Test clean WER | Test other WER | GPUs |
 |:-------------:|:--------------:|:--------------:|:--------:|
-| 24-03-22 | 2.26 | 5.52 | 1xA100 40GB |
+| 24-03-22 | 2.27 | 5.53 | 4xV100 32GB |
 
 ## Pipeline description
 
@@ -122,7 +122,7 @@ cd recipes/LibriSpeech/ASR/transformer
 python train.py hparams/transformer.yaml --data_folder=your_data_folder
 ```
 
-You can find our training results (models, logs, etc) [here](https://drive.google.com/drive/folders/1ZudxqMWb8VNCJKvY2Ws5oNY3WI1To0I7?usp=sharing).
+You can find our training results (models, logs, etc) [here](https://drive.google.com/drive/folders/1Nv1OLbHLqVeShyZ8LY9gjhYGE1DBFzFf?usp=sharing).
 
 ### Limitations
 The SpeechBrain team does not provide any warranty on the performance achieved by this model when used on other datasets.

asr.ckpt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2db0253dbd1cc0116907d789252d0e6154522cdf686138c1c0a38137668d7fa7
+oid sha256:5a04a1b364fe37769ed47904e0237c412e874321efeda55986f08a06822fe6ea
 size 291335121

hyperparams.yaml

@@ -20,18 +20,17 @@ nhead: 4
 num_encoder_layers: 12
 num_decoder_layers: 6
 d_ffn: 2048
-transformer_dropout: 0.0
+transformer_dropout: 0.1
 activation: !name:torch.nn.GELU
 output_neurons: 5000
 vocab_size: 5000
 
 # Outputs
 blank_index: 0
-label_smoothing: 0.1
+label_smoothing: 0.0
 pad_index: 0
 bos_index: 1
 eos_index: 2
-unk_index: 0
 
 # Decoding parameters
 min_decode_ratio: 0.0

normalizer.ckpt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d5650a789fc12867073ec4d44e9b5343f3db2b043ebfbf5843a2c02db5ca0c69
+oid sha256:956e9c1751beac2f75bfff8253f9811b76a1d57bb52f28587d3b6d45bbce6315
 size 1703

pretrained_models/asr-transformer-transformerlm-librispeech/asr.ckpt

@@ -0,0 +1 @@
+/Users/titlet/CloudStation/work/speechbrain/workspace/huggingface_hub/asr-transformer-transformerlm-librispeech/asr.ckpt

pretrained_models/asr-transformer-transformerlm-librispeech/custom.py

@@ -0,0 +1 @@
+/Users/titlet/CloudStation/work/speechbrain/workspace/huggingface_hub/asr-transformer-transformerlm-librispeech/custom.py

pretrained_models/asr-transformer-transformerlm-librispeech/hyperparams.yaml

@@ -0,0 +1 @@
+/Users/titlet/CloudStation/work/speechbrain/workspace/huggingface_hub/asr-transformer-transformerlm-librispeech/hyperparams.yaml

pretrained_models/asr-transformer-transformerlm-librispeech/lm.ckpt

@@ -0,0 +1 @@
+/Users/titlet/CloudStation/work/speechbrain/workspace/huggingface_hub/asr-transformer-transformerlm-librispeech/lm.ckpt

pretrained_models/asr-transformer-transformerlm-librispeech/normalizer.ckpt

@@ -0,0 +1 @@
+/Users/titlet/CloudStation/work/speechbrain/workspace/huggingface_hub/asr-transformer-transformerlm-librispeech/normalizer.ckpt

pretrained_models/asr-transformer-transformerlm-librispeech/tokenizer.ckpt

@@ -0,0 +1 @@
+/Users/titlet/CloudStation/work/speechbrain/workspace/huggingface_hub/asr-transformer-transformerlm-librispeech/tokenizer.ckpt