speechbrain · mravanelli · Jul 19, 2023 · Jul 4, 2023 · Jul 13, 2023 · Jul 13, 2023
diff --git a/README.md b/README.md
@@ -210,7 +210,7 @@ We are currently implementing speech synthesis pipelines and real-time speech pr
 
 # Conference Tutorials
 SpeechBrain has been presented at Interspeech 2021 and 2022 as well as ASRU 2021. When possible, we will provide some ressources here:
-- [Interspeech 2022 slides.](https://drive.google.com/drive/folders/1d6GAquxw6rZBI-7JvfUQ_-upeiKstJEo?usp=sharing)
+- [Interspeech 2022 slides.](https://drive.google.com/drive/folders/1d6GAquxw6rZBI-7JvfUQ_-upeiKstJEo)
 - [Interspeech 2021 YouTube recordings.](https://www.youtube.com/results?search_query=Interspeech+speechbrain+)
 
 # Quick installation

diff --git a/docs/experiment.md b/docs/experiment.md
@@ -46,7 +46,7 @@ model: !new:speechbrain.lobes.models.CRDNN.CRDNN
   every user either by editing the yaml, or with an override (passed to
   `load_extended_yaml`).
 
-For more details on YAML and our extensions, please see our dedicated [tutorial](https://colab.research.google.com/drive/1Pg9by4b6-8QD2iC0U7Ic3Vxq4GEwEdDz?usp=sharing).
+For more details on YAML and our extensions, please see our dedicated [tutorial](https://colab.research.google.com/drive/1Pg9by4b6-8QD2iC0U7Ic3Vxq4GEwEdDz).
 
 ## Running arguments
 SpeechBrain defines a set of running arguments that can be set from the command line args (or within the YAML file).
@@ -58,7 +58,7 @@ SpeechBrain defines a set of running arguments that can be set from the command
 - `distributed_backend`: default "nccl", options: `["nccl", "gloo", "mpi"]`, this backend will be used as a DDP communication protocol. See PyTorch documentation for more details.
 - Additional runtime arguments are documented in the Brain class.
 
-Please note that we provide a dedicated [tutorial](https://colab.research.google.com/drive/13pBUacPiotw1IvyffvGZ-HrtBr9T6l15?usp=sharing) to document the different multi-gpu training strategies:
+Please note that we provide a dedicated [tutorial](https://colab.research.google.com/drive/13pBUacPiotw1IvyffvGZ-HrtBr9T6l15) to document the different multi-gpu training strategies:
 
 You can also override parameters in YAML in this way:
 

diff --git a/recipes/AISHELL-1/ASR/CTC/README.md b/recipes/AISHELL-1/ASR/CTC/README.md
@@ -24,7 +24,7 @@ Results are reported in terms of Character Error Rate (CER).
 |:--------------------------:|:-----:| :-----:| :-----:| :-----: |
 | train_with_wav2vec.yaml | No | 5.06 | 4.52 | 1xRTX 8000 Ti 48GB |
 
-You can checkout our results (models, training logs, etc,) [here](https://drive.google.com/drive/folders/1GTB5IzQPl57j-0I1IpmvKg722Ti4ahLz?usp=sharing)
+You can checkout our results (models, training logs, etc,) [here](https://www.dropbox.com/sh/e4bth1bylk7c6h8/AADFq3cWzBBKxuDv09qjvUMta?dl=0)
 
 # Training Time
 It takes about 2h on 1 RTX 8000 (48GB)

diff --git a/recipes/AISHELL-1/ASR/seq2seq/README.md b/recipes/AISHELL-1/ASR/seq2seq/README.md
@@ -30,7 +30,7 @@ Results are reported in terms of Character Error Rate (CER). It is not clear fro
 | Base (keep spaces) | 7.51 |
 
 You can checkout our results (models, training logs, etc,) here:
-https://drive.google.com/drive/folders/1zlTBib0XEwWeyhaXDXnkqtPsIBI18Uzs?usp=sharing
+https://www.dropbox.com/sh/kefuzzf6jaljqbr/AADBRWRzHz74GCMDqJY9BES4a?dl=0
 
 # Training Time
 It takes about 1h 30 minutes on a NVIDIA V100 (32GB).

diff --git a/recipes/AISHELL-1/ASR/transformer/README.md b/recipes/AISHELL-1/ASR/transformer/README.md
@@ -28,7 +28,7 @@ Results are reported in terms of Character Error Rate (CER).
 | train_ASR_transformer_with_wav2vect.yaml | No | 5.58 | 5.19 | 1xRTX 8000 Ti 48GB |
 
 You can checkout our results (models, training logs, etc,) here:
-https://drive.google.com/drive/folders/1xKo_6Pxk0saPXjGZg8um68b_l0Tgfdjy?usp=sharing
+https://www.dropbox.com/sh/tp6tjmysorgvsr4/AAD7KNqi1ot0gR4N406JbKM6a?dl=0
 
 # Training Time
 It takes about 1h 10 minutes on a NVIDIA V100 (32GB) for train_ASR_transformer.yaml,

diff --git a/recipes/AISHELL-1/Tokenizer/README.md b/recipes/AISHELL-1/Tokenizer/README.md
@@ -7,7 +7,7 @@ The tokenizer is trained on the top of the AISHELL training transcriptions.
 
 `python train.py hparams/train_transformer_tokenizer_bpe5000.yaml`
 
-The output folder with the tokenizers and logs is available [here](https://drive.google.com/drive/folders/15wOIkFMHB-wwR1OW6NupcLcHbiJZJ_CU?usp=sharing).
+The output folder with the tokenizers and logs is available [here](https://www.dropbox.com/sh/gh1qyf833t7h3op/AADG0y1bGGIL4yufsXtuBgXma?dl=0).
 
 # **About SpeechBrain**
 - Website: https://speechbrain.github.io/

diff --git a/recipes/Aishell1Mix/separation/README.md b/recipes/Aishell1Mix/separation/README.md
@@ -50,7 +50,7 @@ Here are the SI - SNRi results (in dB) on the test set of Aishell1Mix dataset wi
 | NoDynamicMixing | 8.1 |
 | DynamicMixing | 11.2 |
 
-The output folders with model checkpoints and logs is available [here](https://drive.google.com/drive/folders/1GvJiUxhdN5bfbuBdxclPzdAPd2op1PCZ?usp=sharing).
+The output folders with model checkpoints and logs is available [here](https://www.dropbox.com/sh/6x9356yuybj8lue/AABPlpS03Vcci_E3jA69oKoXa?dl=0).
 
 # Example calls for running the training scripts
 
@@ -69,7 +69,7 @@ You can run the following command to train the model using Distributed Data Para
 ```
  python -m torch.distributed.launch --nproc_per_node=2 train.py hparams/sepformer.yaml --data_folder /yourdatapath --distributed_launch --distributed_backend='nccl'
 ```
-You can add the other runtime options as appropriate. For more complete information on multi-GPU usage, take a look at this [tutorial](https://colab.research.google.com/drive/13pBUacPiotw1IvyffvGZ-HrtBr9T6l15?usp=sharing).
+You can add the other runtime options as appropriate. For more complete information on multi-GPU usage, take a look at this [tutorial](https://colab.research.google.com/drive/13pBUacPiotw1IvyffvGZ-HrtBr9T6l15).
 
 
 # **About SpeechBrain**

diff --git a/recipes/BinauralWSJ0Mix/separation/README.md b/recipes/BinauralWSJ0Mix/separation/README.md
@@ -47,7 +47,7 @@ Here are the SNRi results (in dB) as well as ITD and ILD errors as the metric fo
 * ConvTasnet-parallel-noise.yaml refers to the above Tasnet applied to 2 speakers with DEMAND noise.
 * ConvTasnet-parallel-reverb.yaml refers to the above Tasnet applied to 2 speakers with reverberance(RT60) from the [BRIR Sim Set](http://iosr.uk/software/index.php).
 
-The output folders with the checkpoints, logs, etc are available [here](https://drive.google.com/drive/folders/17FFwlIq6MQLHT9RXPgeYssti5TEeEXsx?usp=sharing)
+The output folders with the checkpoints, logs, etc are available [here](https://www.dropbox.com/sh/i7fhu7qswjb84gw/AABsX1zP-GOTmyl86PtU8GGua?dl=0)
 
 # Example calls for running the training scripts
 
@@ -64,7 +64,7 @@ You can run the following command to train the model using Distributed Data Para
 ```
  python -m torch.distributed.launch --nproc_per_node=2 train.py hparams/convtasnet-parallel.yaml --data_folder /yourdatapath --distributed_launch --distributed_backend='nccl'
 ```
-You can add the other runtime options as appropriate. For more complete information on multi-GPU usage, take a look at this [tutorial](https://colab.research.google.com/drive/13pBUacPiotw1IvyffvGZ-HrtBr9T6l15?usp=sharing).
+You can add the other runtime options as appropriate. For more complete information on multi-GPU usage, take a look at this [tutorial](https://colab.research.google.com/drive/13pBUacPiotw1IvyffvGZ-HrtBr9T6l15).
 
 
 

diff --git a/recipes/BinauralWSJ0Mix/separation/train.py b/recipes/BinauralWSJ0Mix/separation/train.py
@@ -700,7 +700,7 @@ def audio_pipeline_noise(noise_wav):
             print("Download Datasets Generation scripts")
             fetch(
                 filename="main.zip",
-                source="https://github.com/huangzj421/Binaural-WSJ0Mix/archive/refs/heads",
+                source="https://github.com/huangzj421/Binaural-WSJ0Mix/",
                 savedir=hparams["data_folder"],
                 save_filename="Binaural-WSJ0Mix-main.zip",
             )

diff --git a/recipes/CommonLanguage/common_language_prepare.py b/recipes/CommonLanguage/common_language_prepare.py
@@ -70,7 +70,7 @@
 def prepare_common_language(data_folder, save_folder, skip_prep=False):
     """
     Prepares the csv files for the CommonLanguage dataset for LID.
-    Download: https://drive.google.com/uc?id=1Vzgod6NEYO1oZoz_EcgpZkUO9ohQcO1F
+    Download: https://www.dropbox.com/s/qqpmqay3q9xb1vf/common_voice_kpd.tar.gz?dl=0
 
     Arguments
     ---------

diff --git a/recipes/CommonLanguage/lang_id/README.md b/recipes/CommonLanguage/lang_id/README.md
@@ -14,7 +14,7 @@ better performance, at the cost of longer training time and higher compute resou
 # Performance
 | Release | hyperparams file | Val. Err | Test Err | Model link | GPUs |
 |:-------------:|:---------------------------:| -----:| -----:| --------:| :-----------:|
-| 21-06-28 | train.yaml |  13. 5 | 15.1 | https://drive.google.com/drive/folders/1btxc_H27AP_f6u4X47FM0LSteUdzhfFR?usp=sharing | 1xV100 16GB |
+| 21-06-28 | train.yaml |  13. 5 | 15.1 | https://www.dropbox.com/sh/1fxpzyv67ouwd2c/AAAeMUWYP2f1ycpE1Lp1CwEla?dl=0 | 1xV100 16GB |
 
 Each epoch takes approximately 14 minutes on an NVIDIA V100.
 

diff --git a/recipes/CommonVoice/ASR/CTC/README.md b/recipes/CommonVoice/ASR/CTC/README.md
@@ -18,14 +18,14 @@ Here is a list of the different languages that we tested within the CommonVoice
 # Results
 | Language | CommonVoice Release | hyperparams file | LM | Val. CER | Val. WER | Test CER | Test WER | HuggingFace link | Model link | GPUs |
 | ------------- |:-------------:|:---------------------------:| -----:| -----:| -----:| -----:| -----:| :-----------:| :-----------:| :-----------:|
-| English | 2020-12-11 | train_en_with_wav2vec.yaml | No | 5.01 | 12.57 | 7.32 | 15.58 | Not Avail. | [model](https://drive.google.com/drive/folders/1tYO__An68xrM5pR1UIXzEkwzvKX2Tz2o?usp=sharing) | 2xV100 32GB |
-| German | 2022-08-16 | train_de_with_wav2vec.yaml | No | 1.90 | 8.02 | 2.40 | 9.54 | [model](https://huggingface.co/speechbrain/asr-wav2vec2-commonvoice-de) | [model](https://drive.google.com/drive/folders/19G2Zm8896QSVDqVfs7PS_W86-K0-5xeC?usp=sharing) | 1xRTXA6000 48GB |
-| French | 2020-12-11 | train_fr_with_wav2vec.yaml | No | 2.60 | 8.59 | 3.19 | 9.96 | [model](https://huggingface.co/speechbrain/asr-wav2vec2-commonvoice-fr) | [model](https://drive.google.com/drive/folders/1T9DfdZwcNI9CURxhLCi8GA5JVz8adiY8?usp=sharing) | 2xV100 32GB |
-| Italian | 2020-12-11 | train_it_with_wav2vec.yaml | No | 2.77 | 9.83 | 3.16 | 10.85 | Not Avail. | [model](https://drive.google.com/drive/folders/1JhlxeA04tWg_vKcNChOoXSnjBe4luRby?usp=sharing) | 2xV100 32GB |
-| Kinyarwanda | 2020-12-11 | train_rw_with_wav2vec.yaml | No | 6.20 | 20.07 | 8.25 | 23.12 | Not Avail. | [model](https://drive.google.com/drive/folders/12_BDenvOqEERDZLAN-KdiAHklvuo35tx?usp=sharing) | 2xV100 32GB |
+| English | 2020-12-11 | train_en_with_wav2vec.yaml | No | 5.01 | 12.57 | 7.32 | 15.58 | Not Avail. | [model](https://www.dropbox.com/sh/o3q43r4wdovbmnd/AADXcVomQr549NdAgCpI7OQHa?dl=0) | 2xV100 32GB |
+| German | 2022-08-16 | train_de_with_wav2vec.yaml | No | 1.90 | 8.02 | 2.40 | 9.54 | [model](https://huggingface.co/speechbrain/asr-wav2vec2-commonvoice-de) | [model](https://www.dropbox.com/sh/vdz7apt16nbq94g/AADI5o23Ll_NmjiPlg9bzPjta?dl=0) | 1xRTXA6000 48GB |
+| French | 2020-12-11 | train_fr_with_wav2vec.yaml | No | 2.60 | 8.59 | 3.19 | 9.96 | [model](https://huggingface.co/speechbrain/asr-wav2vec2-commonvoice-fr) | [model](https://www.dropbox.com/sh/wytlbeddrt8oe4n/AAAY59qMsDlWy5F017bmBeVua?dl=0) | 2xV100 32GB |
+| Italian | 2020-12-11 | train_it_with_wav2vec.yaml | No | 2.77 | 9.83 | 3.16 | 10.85 | Not Avail. | [model](https://www.dropbox.com/sh/0v2o2hmrv1j33p6/AAA3xUiqKbSKsX88fWfptPmFa?dl=0) | 2xV100 32GB |
+| Kinyarwanda | 2020-12-11 | train_rw_with_wav2vec.yaml | No | 6.20 | 20.07 | 8.25 | 23.12 | Not Avail. | [model](https://www.dropbox.com/sh/ccgirbq9r8uzubi/AAAynCvEV8EjEpMavFRPp87Ta?dl=0) | 2xV100 32GB |
 
 *For German, it takes around 5.5 hrs an epoch.* <br>
-The output folders with checkpoints and logs can be found [here](https://drive.google.com/drive/folders/11NMzY0zV-NqJmPMyZfC3RtT64bYe-G_O?usp=sharing).
+The output folders with checkpoints and logs can be found [here](https://www.dropbox.com/sh/852eq7pbt6d65ai/AACv4wAzk1pWbDo4fjVKLICYa?dl=0).
 
 ## How to simply use pretrained models to transcribe my audio file?
 

diff --git a/recipes/CommonVoice/ASR/seq2seq/README.md b/recipes/CommonVoice/ASR/seq2seq/README.md
@@ -20,17 +20,17 @@ Here is a list of the different languages that we tested within the CommonVoice
 
 | Language | CommonVoice Release | hyperparams file | LM | Val. CER | Val. WER | Test CER | Test WER | HuggingFace link | Model link | GPUs |
 | ------------- |:-------------:|:---------------------------:| -----:| -----:| -----:| -----:| -----:| :-----------:| :-----------:| :-----------:|
-| French | 2020-12-11 | train_fr.yaml | No | 5.22 | 13.92 | 6.43 | 15.99 | [model](https://huggingface.co/speechbrain/asr-crdnn-commonvoice-fr) | [model](https://drive.google.com/drive/folders/1GShpLaX9AOLklwBAOLB9B9lLTn0HsWwS?usp=sharing) | 2xV100 16GB |
+| French | 2020-12-11 | train_fr.yaml | No | 5.22 | 13.92 | 6.43 | 15.99 | [model](https://huggingface.co/speechbrain/asr-crdnn-commonvoice-fr) | [model](https://www.dropbox.com/sh/nskc700cheejyu9/AADnRP1TO1Yh92jC-WuYCFf5a?dl=0) | 2xV100 16GB |
 | French | 2020-12-11 | train_fr_with_wav2vec.yaml | No | 6.13 | 11.82 | 9.78 | 13.34 | Not Avail. | 2xV100 32GB |
-| Kinyarwanda | 2020-12-11 | train_rw.yaml | No | 7.30 | 21.36 | 9.55 | 24.27 | Not Avail. | [model](https://drive.google.com/drive/folders/122efLUMYoc1LGoK7O6LIWkSklmjKVGxM?usp=sharing) | 2xV100 32GB |
-| Kinyarwanda | 2020-12-11 | train_rw_with_wav2vec.yaml | No | 5.08 | 15.88 | 8.33 | 18.91 | [model](https://huggingface.co/speechbrain/asr-wav2vec2-commonvoice-rw) | [model](https://drive.google.com/drive/folders/1ceHxyNojY0wXmXyPoyn9xUiH_5B5qgE4?usp=sharing) | 2xV100 16GB |
-| English | 2020-12-11 | train_en.yaml | No | 8.66 | 20.16 | 12.93 | 24.89 | Not Avail. | [model](https://drive.google.com/drive/folders/1FAKRhfu_1gLnkshYGKp-6G9ZVMIUlv9n?usp=sharing) | 2xV100 16GB |
-| English | 2020-12-11 | train_en_with_wav2vec.yaml | No | 14.50 | 13.21 | 24.65 | 15.69 | [model](https://huggingface.co/speechbrain/asr-wav2vec2-commonvoice-en) | [model](https://drive.google.com/drive/folders/1EfIZiJi8ch53mil9K4tn46OrmTJq5WYj?usp=sharing) | 2xV100 32GB |
-| Italian | 2020-12-11 | train_it.yaml | No | 5.14 | 15.59 | 15.40 | 16.61 | [model](https://huggingface.co/speechbrain/asr-crdnn-commonvoice-it) | [model](https://drive.google.com/drive/folders/1asxPsY1EBGHIpIFhBtUi9oiyR6C7gC0g?usp=sharing) | 2xV100 16GB |
-| Italian | 2020-12-11 | train_it_with_wav2vec.yaml | No | 3.11 | 8.30 | 5.75 | 9.86 | [model](https://huggingface.co/speechbrain/asr-wav2vec2-commonvoice-it) | [model](https://drive.google.com/drive/folders/1LKA50Qsr1fM1E3t4PHMWUjlBMS2QGFHj?usp=sharing) | 2xV100 16GB |
+| Kinyarwanda | 2020-12-11 | train_rw.yaml | No | 7.30 | 21.36 | 9.55 | 24.27 | Not Avail. | [model](https://www.dropbox.com/sh/glzq0hrqw2khcjq/AADfl_7ra0cLWi1VOzpy74NUa?dl=0) | 2xV100 32GB |
+| Kinyarwanda | 2020-12-11 | train_rw_with_wav2vec.yaml | No | 5.08 | 15.88 | 8.33 | 18.91 | [model](https://huggingface.co/speechbrain/asr-wav2vec2-commonvoice-rw) | [model](https://www.dropbox.com/sh/x714xp9wq1a9azr/AADoxhS7JVDQ7IY1lYL7U8rJa?dl=0) | 2xV100 16GB |
+| English | 2020-12-11 | train_en.yaml | No | 8.66 | 20.16 | 12.93 | 24.89 | Not Avail. | [model](https://www.dropbox.com/sh/bdetfgii7xwscyj/AABYQL_eso8K1937QAg1GK66a?dl=0) | 2xV100 16GB |
+| English | 2020-12-11 | train_en_with_wav2vec.yaml | No | 14.50 | 13.21 | 24.65 | 15.69 | [model](https://huggingface.co/speechbrain/asr-wav2vec2-commonvoice-en) | [model](https://www.dropbox.com/sh/s4khqzp7qe5iiaa/AADkWpWPE1UAbu2isycykcAAa?dl=0) | 2xV100 32GB |
+| Italian | 2020-12-11 | train_it.yaml | No | 5.14 | 15.59 | 15.40 | 16.61 | [model](https://huggingface.co/speechbrain/asr-crdnn-commonvoice-it) | [model](https://www.dropbox.com/sh/438hgcl1wwdzmbo/AAAPZGfNXMztNYHyWwT_kk8la?dl=0) | 2xV100 16GB |
+| Italian | 2020-12-11 | train_it_with_wav2vec.yaml | No | 3.11 | 8.30 | 5.75 | 9.86 | [model](https://huggingface.co/speechbrain/asr-wav2vec2-commonvoice-it) | [model](https://www.dropbox.com/sh/46z6xbydqjaxbb2/AACyQ_NaErAfev6JVjeaIWyea?dl=0) | 2xV100 16GB |
 | German | 2021-10-28 | train_de.yaml | No | 4.32 | 13.99 | 4.93 | 15.37 | [model](https://huggingface.co/speechbrain/asr-crdnn-commonvoice-de) | -- | 1x V100 16GB |
 
-The output folders with checkpoints and logs can be found [here](https://drive.google.com/drive/folders/11NMzY0zV-NqJmPMyZfC3RtT64bYe-G_O?usp=sharing).
+The output folders with checkpoints and logs can be found [here](https://www.dropbox.com/sh/852eq7pbt6d65ai/AACv4wAzk1pWbDo4fjVKLICYa?dl=0).
 
 ## How to simply use pretrained models to transcribe my audio file?
 

diff --git a/recipes/CommonVoice/ASR/transducer/README.md b/recipes/CommonVoice/ASR/transducer/README.md
@@ -25,9 +25,9 @@ with our transducers:
 
 | Language | Release | hyperparams file | LM | Val. CER | Val. WER | Test CER | Test WER | Model link | GPUs |
 | ------------- |:-------------:|:---------------------------:| -----:| -----:| -----:| -----:| -----:| :-----------:| :-----------:|
-| French | 2020-06-22 | train_fr.yaml | No | 6.70 | 18.97 | 7.41 | 20.18 | [model](https://drive.google.com/drive/folders/1ZwY2FaRl1gfFbupodph_xRiGj4h25I08?usp=sharing) | 2xV100 16GB |
+| French | 2020-06-22 | train_fr.yaml | No | 6.70 | 18.97 | 7.41 | 20.18 | [model](https://www.dropbox.com/sh/mp5w1asmuy88vhr/AABF9fFIOh3AIBTP-Xn-7c8_a?dl=0) | 2xV100 16GB |
 
-The output folders with checkpoints and logs can be found [here](https://drive.google.com/drive/folders/11NMzY0zV-NqJmPMyZfC3RtT64bYe-G_O?usp=sharing).
+The output folders with checkpoints and logs can be found [here](https://www.dropbox.com/sh/852eq7pbt6d65ai/AACv4wAzk1pWbDo4fjVKLICYa?dl=0).
 
 # **About SpeechBrain**
 - Website: https://speechbrain.github.io/