speechbrain
diff --git a/‎docs/README.md‎
Lines changed: 2 additions & 0 deletions b/‎docs/README.md‎
Lines changed: 2 additions & 0 deletions
diff --git a/‎docs/tutorials/assets/attention-chunking-dep.png‎
149 KB b/‎docs/tutorials/assets/attention-chunking-dep.png‎
149 KB
diff --git a/‎docs/tutorials/assets/attention-chunking-no-lc.png‎
128 KB b/‎docs/tutorials/assets/attention-chunking-no-lc.png‎
128 KB
diff --git a/‎docs/tutorials/assets/attention-chunking.png‎
54 KB b/‎docs/tutorials/assets/attention-chunking.png‎
54 KB
diff --git a/‎docs/tutorials/assets/attn-restrict.png‎
189 KB b/‎docs/tutorials/assets/attn-restrict.png‎
189 KB
diff --git a/‎docs/tutorials/assets/conformer-simple.png‎
109 KB b/‎docs/tutorials/assets/conformer-simple.png‎
109 KB
diff --git a/‎docs/tutorials/assets/dcc-causal.png‎
15.9 KB b/‎docs/tutorials/assets/dcc-causal.png‎
15.9 KB
diff --git a/‎docs/tutorials/assets/dcc-dcc.png‎
18.9 KB b/‎docs/tutorials/assets/dcc-dcc.png‎
18.9 KB
diff --git a/‎docs/tutorials/assets/dcc-regular.png‎
28.1 KB b/‎docs/tutorials/assets/dcc-regular.png‎
28.1 KB
diff --git a/‎docs/tutorials/nn.rst‎
Lines changed: 21 additions & 0 deletions b/‎docs/tutorials/nn.rst‎
Lines changed: 21 additions & 0 deletions
@@ -36,6 +36,8 @@ The `docs/tutorials` directory exclusively contains tutorials in Jupyter Noteboo
   - It's OK if the user has to run the notebook to get some of the heavier outputs.
 - Preferably use Jupyter Notebook for final editing of your notebook.
   - Jupyter Notebook tends to have somewhat sane `.ipynb` output. This avoids Git diffs from being excessively large.
+- **Images can be put in the `docs/tutorials/assets` directory,** rather than embedded as base64. You can then refer to them in Markdown like `![alt text](../assets/myimage.png)`. These will work correctly when imported on Colab.
+  - Pick descriptive names.
 
 #### Integration in documentation
 
 
@@ -11,6 +11,7 @@ Neural Architectures
    nn/using-wav2vec-2.0-hubert-wavlm-and-whisper-from-huggingface-with-speechbrain.ipynb
    nn/complex-and-quaternion-neural-networks.ipynb
    nn/recurrent-neural-networks-and-speechbrain.ipynb
+   nn/conformer-streaming-asr.ipynb
 
 
 .. rubric:: `🔗 Fine-tuning or using Whisper, wav2vec2, HuBERT and others with SpeechBrain and HuggingFace <nn/using-wav2vec-2.0-hubert-wavlm-and-whisper-from-huggingface-with-speechbrain.html>`_
@@ -67,3 +68,23 @@ Linear, Convolution, Recurrent and Normalisation.
 Recurrent Neural Networks (RNNs) offer a natural way to process sequences.
 This tutorial demonstrates how to use the SpeechBrain implementations of RNNs including LSTMs, GRU, RNN and LiGRU a specific recurrent cell designed
 for speech-related tasks. RNNs are at the core of many sequence to sequence models.
+
+
+.. rubric:: `🔗 Streaming Speech Recognition with Conformers <nn/conformer-streaming-asr.html>`_
+   :heading-level: 2
+
+.. list-table::
+   :widths: 20 20 20 20 20
+   :header-rows: 0
+
+   * - de Langen S.
+     - Sep. 2024
+     - Difficulty: medium
+     - Time: 60min+
+     - `🔗 Google Colab <https://colab.research.google.com/github/speechbrain/speechbrain/blob/develop/docs/tutorials/nn/conformer-streaming-asr.ipynb>`__
+
+
+Automatic Speech Recognition (ASR) models are often only designed to transcribe an entire large chunk of audio and are unsuitable for usecases like live stream transcription, which requires low-latency, long-form transcription.
+
+This tutorial introduces the Dynamic Chunk Training approach and architectural changes you can apply to make the Conformer model streamable. It introduces the tooling for training and inference that SpeechBrain can provide for you.
+This might be a good starting point if you're interested in training and understanding your own streaming models, or even if you want to explore improved streaming architectures.