support exporting to ncnn format via PNNX (#571)

.github/scripts/run-librispeech-lstm-transducer-stateless2-2022-09-03.yml

@@ -0,0 +1,160 @@
+#!/usr/bin/env bash
+
+log() {
+  # This function is from espnet
+  local fname=${BASH_SOURCE[1]##*/}
+  echo -e "$(date '+%Y-%m-%d %H:%M:%S') (${fname}:${BASH_LINENO[0]}:${FUNCNAME[1]}) $*"
+}
+
+cd egs/librispeech/ASR
+
+repo_url=https://huggingface.co/csukuangfj/icefall-asr-librispeech-lstm-transducer-stateless2-2022-09-03
+
+log "Downloading pre-trained model from $repo_url"
+git lfs install
+git clone $repo_url
+repo=$(basename $repo_url)
+
+log "Display test files"
+tree $repo/
+soxi $repo/test_wavs/*.wav
+ls -lh $repo/test_wavs/*.wav
+
+pushd $repo/exp
+ln -s pretrained-iter-468000-avg-16.pt pretrained.pt
+ln -s pretrained-iter-468000-avg-16.pt epoch-99.pt
+popd
+
+log  "Install ncnn and pnnx"
+
+# We are using a modified ncnn here. Will try to merge it to the official repo
+# of ncnn
+git clone https://github.com/csukuangfj/ncnn
+pushd ncnn
+git submodule init
+git submodule update python/pybind11
+python3 setup.py bdist_wheel
+ls -lh dist/
+pip install dist/*.whl
+cd tools/pnnx
+mkdir build
+cd build
+cmake ..
+make -j4 pnnx
+
+./src/pnnx || echo "pass"
+
+popd
+
+log "Test exporting to pnnx format"
+
+./lstm_transducer_stateless2/export.py \
+  --exp-dir $repo/exp \
+  --bpe-model $repo/data/lang_bpe_500/bpe.model \
+  --epoch 99 \
+  --avg 1 \
+  --use-averaged-model 0 \
+  --pnnx 1
+
+./ncnn/tools/pnnx/build/src/pnnx $repo/exp/encoder_jit_trace-pnnx.pt
+./ncnn/tools/pnnx/build/src/pnnx $repo/exp/decoder_jit_trace-pnnx.pt
+./ncnn/tools/pnnx/build/src/pnnx $repo/exp/joiner_jit_trace-pnnx.pt
+
+./lstm_transducer_stateless2/ncnn-decode.py \
+ --bpe-model-filename $repo/data/lang_bpe_500/bpe.model \
+ --encoder-param-filename $repo/exp/encoder_jit_trace-pnnx.ncnn.param \
+ --encoder-bin-filename $repo/exp/encoder_jit_trace-pnnx.ncnn.bin \
+ --decoder-param-filename $repo/exp/decoder_jit_trace-pnnx.ncnn.param \
+ --decoder-bin-filename $repo/exp/decoder_jit_trace-pnnx.ncnn.bin \
+ --joiner-param-filename $repo/exp/joiner_jit_trace-pnnx.ncnn.param \
+ --joiner-bin-filename $repo/exp/joiner_jit_trace-pnnx.ncnn.bin \
+ $repo/test_wavs/1089-134686-0001.wav
+
+./lstm_transducer_stateless2/streaming-ncnn-decode.py \
+ --bpe-model-filename $repo/data/lang_bpe_500/bpe.model \
+ --encoder-param-filename $repo/exp/encoder_jit_trace-pnnx.ncnn.param \
+ --encoder-bin-filename $repo/exp/encoder_jit_trace-pnnx.ncnn.bin \
+ --decoder-param-filename $repo/exp/decoder_jit_trace-pnnx.ncnn.param \
+ --decoder-bin-filename $repo/exp/decoder_jit_trace-pnnx.ncnn.bin \
+ --joiner-param-filename $repo/exp/joiner_jit_trace-pnnx.ncnn.param \
+ --joiner-bin-filename $repo/exp/joiner_jit_trace-pnnx.ncnn.bin \
+ $repo/test_wavs/1089-134686-0001.wav
+
+
+
+log "Test exporting with torch.jit.trace()"
+
+./lstm_transducer_stateless2/export.py \
+  --exp-dir $repo/exp \
+  --bpe-model $repo/data/lang_bpe_500/bpe.model \
+  --epoch 99 \
+  --avg 1 \
+  --use-averaged-model 0 \
+  --jit-trace 1
+
+log "Decode with models exported by torch.jit.trace()"
+
+./lstm_transducer_stateless2/jit_pretrained.py \
+  --bpe-model $repo/data/lang_bpe_500/bpe.model \
+  --encoder-model-filename $repo/exp/encoder_jit_trace.pt \
+  --decoder-model-filename $repo/exp/decoder_jit_trace.pt \
+  --joiner-model-filename $repo/exp/joiner_jit_trace.pt \
+  $repo/test_wavs/1089-134686-0001.wav \
+  $repo/test_wavs/1221-135766-0001.wav \
+  $repo/test_wavs/1221-135766-0002.wav
+
+for sym in 1 2 3; do
+  log "Greedy search with --max-sym-per-frame $sym"
+
+  ./lstm_transducer_stateless2/pretrained.py \
+    --method greedy_search \
+    --max-sym-per-frame $sym \
+    --checkpoint $repo/exp/pretrained.pt \
+    --bpe-model $repo/data/lang_bpe_500/bpe.model \
+    $repo/test_wavs/1089-134686-0001.wav \
+    $repo/test_wavs/1221-135766-0001.wav \
+    $repo/test_wavs/1221-135766-0002.wav
+done
+
+for method in modified_beam_search beam_search fast_beam_search; do
+  log "$method"
+
+  ./lstm_transducer_stateless2/pretrained.py \
+    --method $method \
+    --beam-size 4 \
+    --checkpoint $repo/exp/pretrained.pt \
+    --bpe-model $repo/data/lang_bpe_500/bpe.model \
+    $repo/test_wavs/1089-134686-0001.wav \
+    $repo/test_wavs/1221-135766-0001.wav \
+    $repo/test_wavs/1221-135766-0002.wav
+done
+
+echo "GITHUB_EVENT_NAME: ${GITHUB_EVENT_NAME}"
+echo "GITHUB_EVENT_LABEL_NAME: ${GITHUB_EVENT_LABEL_NAME}"
+if [[ x"${GITHUB_EVENT_NAME}" == x"schedule" || x"${GITHUB_EVENT_LABEL_NAME}" == x"ncnn"  ]]; then
+  mkdir -p lstm_transducer_stateless2/exp
+  ln -s $PWD/$repo/exp/pretrained.pt lstm_transducer_stateless2/exp/epoch-999.pt
+  ln -s $PWD/$repo/data/lang_bpe_500 data/
+
+  ls -lh data
+  ls -lh lstm_transducer_stateless2/exp
+
+  log "Decoding test-clean and test-other"
+
+  # use a small value for decoding with CPU
+  max_duration=100
+
+  for method in greedy_search fast_beam_search modified_beam_search; do
+    log "Decoding with $method"
+
+    ./lstm_transducer_stateless2/decode.py \
+      --decoding-method $method \
+      --epoch 999 \
+      --avg 1 \
+      --use-averaged-model 0 \
+      --max-duration $max_duration \
+      --exp-dir lstm_transducer_stateless2/exp
+  done
+
+  rm lstm_transducer_stateless2/exp/*.pt
+fi

.github/workflows/run-librispeech-lstm-transducer-stateless2-2022-09-03.yml

@@ -0,0 +1,136 @@
+name: run-librispeech-lstm-transducer-2022-09-03
+
+on:
+  push:
+    branches:
+      - master
+  pull_request:
+    types: [labeled]
+
+  schedule:
+    # minute (0-59)
+    # hour (0-23)
+    # day of the month (1-31)
+    # month (1-12)
+    # day of the week (0-6)
+    # nightly build at 15:50 UTC time every day
+    - cron: "50 15 * * *"
+
+jobs:
+  run_librispeech_pruned_transducer_stateless3_2022_05_13:
+    if: github.event.label.name == 'ncnn' || github.event_name == 'push' || github.event_name == 'schedule'
+    runs-on: ${{ matrix.os }}
+    strategy:
+      matrix:
+        os: [ubuntu-18.04]
+        python-version: [3.8]
+
+      fail-fast: false
+
+    steps:
+      - uses: actions/checkout@v2
+        with:
+          fetch-depth: 0
+
+      - name: Setup Python ${{ matrix.python-version }}
+        uses: actions/setup-python@v2
+        with:
+          python-version: ${{ matrix.python-version }}
+          cache: 'pip'
+          cache-dependency-path: '**/requirements-ci.txt'
+
+      - name: Install Python dependencies
+        run: |
+          grep -v '^#' ./requirements-ci.txt  | xargs -n 1 -L 1 pip install
+          pip uninstall -y protobuf
+          pip install --no-binary protobuf protobuf
+
+      - name: Cache kaldifeat
+        id: my-cache
+        uses: actions/cache@v2
+        with:
+          path: |
+            ~/tmp/kaldifeat
+          key: cache-tmp-${{ matrix.python-version }}
+
+      - name: Install kaldifeat
+        if: steps.my-cache.outputs.cache-hit != 'true'
+        shell: bash
+        run: |
+          .github/scripts/install-kaldifeat.sh
+
+      - name: Cache LibriSpeech test-clean and test-other datasets
+        id: libri-test-clean-and-test-other-data
+        uses: actions/cache@v2
+        with:
+          path: |
+            ~/tmp/download
+          key: cache-libri-test-clean-and-test-other
+
+      - name: Download LibriSpeech test-clean and test-other
+        if: steps.libri-test-clean-and-test-other-data.outputs.cache-hit != 'true'
+        shell: bash
+        run: |
+          .github/scripts/download-librispeech-test-clean-and-test-other-dataset.sh
+
+      - name: Prepare manifests for LibriSpeech test-clean and test-other
+        shell: bash
+        run: |
+          .github/scripts/prepare-librispeech-test-clean-and-test-other-manifests.sh
+
+      - name: Cache LibriSpeech test-clean and test-other fbank features
+        id: libri-test-clean-and-test-other-fbank
+        uses: actions/cache@v2
+        with:
+          path: |
+            ~/tmp/fbank-libri
+          key: cache-libri-fbank-test-clean-and-test-other-v2
+
+      - name: Compute fbank for LibriSpeech test-clean and test-other
+        if: steps.libri-test-clean-and-test-other-fbank.outputs.cache-hit != 'true'
+        shell: bash
+        run: |
+          .github/scripts/compute-fbank-librispeech-test-clean-and-test-other.sh
+
+      - name: Inference with pre-trained model
+        shell: bash
+        env:
+          GITHUB_EVENT_NAME: ${{ github.event_name }}
+          GITHUB_EVENT_LABEL_NAME: ${{ github.event.label.name }}
+        run: |
+          mkdir -p egs/librispeech/ASR/data
+          ln -sfv ~/tmp/fbank-libri egs/librispeech/ASR/data/fbank
+          ls -lh egs/librispeech/ASR/data/*
+
+          sudo apt-get -qq install git-lfs tree sox
+          export PYTHONPATH=$PWD:$PYTHONPATH
+          export PYTHONPATH=~/tmp/kaldifeat/kaldifeat/python:$PYTHONPATH
+          export PYTHONPATH=~/tmp/kaldifeat/build/lib:$PYTHONPATH
+
+          .github/scripts/run-librispeech-lstm-transducer-stateless2-2022-09-03.yml
+
+      - name: Display decoding results for lstm_transducer_stateless2
+        if: github.event_name == 'schedule' || github.event.label.name == 'ncnn'
+        shell: bash
+        run: |
+          cd egs/librispeech/ASR
+          tree lstm_transducer_stateless2/exp
+          cd lstm_transducer_stateless2/exp
+          echo "===greedy search==="
+          find greedy_search -name "log-*" -exec grep -n --color "best for test-clean" {} + | sort -n -k2
+          find greedy_search -name "log-*" -exec grep -n --color "best for test-other" {} + | sort -n -k2
+
+          echo "===fast_beam_search==="
+          find fast_beam_search -name "log-*" -exec grep -n --color "best for test-clean" {} + | sort -n -k2
+          find fast_beam_search -name "log-*" -exec grep -n --color "best for test-other" {} + | sort -n -k2
+
+          echo "===modified beam search==="
+          find modified_beam_search -name "log-*" -exec grep -n --color "best for test-clean" {} + | sort -n -k2
+          find modified_beam_search -name "log-*" -exec grep -n --color "best for test-other" {} + | sort -n -k2
+
+      - name: Upload decoding results for lstm_transducer_stateless2
+        uses: actions/upload-artifact@v2
+        if: github.event_name == 'schedule' || github.event.label.name == 'ncnn'
+        with:
+          name: torch-${{ matrix.torch }}-python-${{ matrix.python-version }}-ubuntu-18.04-cpu-lstm_transducer_stateless2-2022-09-03
+          path: egs/librispeech/ASR/lstm_transducer_stateless2/exp/

.gitignore

@@ -11,3 +11,5 @@ log
 *.bak
 *-bak
 *bak.py
+*.param
+*.bin

docs/source/recipes/librispeech/index.rst

@@ -6,3 +6,4 @@ LibriSpeech
 
    tdnn_lstm_ctc
    conformer_ctc
+   lstm_pruned_stateless_transducer

docs/source/recipes/librispeech/lstm_pruned_stateless_transducer.rst

@@ -0,0 +1,625 @@
+Transducer
+==========
+
+.. hint::
+
+   Please scroll down to the bottom of this page to find download links
+   for pretrained models if you don't want to train a model from scratch.
+
+
+This tutorial shows you how to train a transducer model
+with the `LibriSpeech <https://www.openslr.org/12>`_ dataset.
+
+We use pruned RNN-T to compute the loss.
+
+.. note::
+
+   You can find the paper about pruned RNN-T at the following address:
+
+   `<https://arxiv.org/abs/2206.13236>`_
+
+The transducer model consists of 3 parts:
+
+  - Encoder, a.k.a, transcriber. We use an LSTM model
+  - Decoder, a.k.a, predictor. We use a model consisting of ``nn.Embedding``
+    and ``nn.Conv1d``
+  - Joiner, a.k.a, the joint network.
+
+.. caution::
+
+   Contrary to the conventional RNN-T models, we use a stateless decoder.
+   That is, it has no recurrent connections.
+
+.. hint::
+
+   Since the encoder model is an LSTM, not Transformer/Conformer, the
+   resulting model is suitable for streaming/online ASR.
+
+
+Which model to use
+------------------
+
+Currently, there are two folders about LSTM stateless transducer training:
+
+  - ``(1)`` `<https://github.com/k2-fsa/icefall/tree/master/egs/librispeech/ASR/lstm_transducer_stateless>`_
+
+    This recipe uses only LibriSpeech during training.
+
+  - ``(2)`` `<https://github.com/k2-fsa/icefall/tree/master/egs/librispeech/ASR/lstm_transducer_stateless2>`_
+
+    This recipe uses GigaSpeech + LibriSpeech during training.
+
+``(1)`` and ``(2)`` use the same model architecture. The only difference is that ``(2)`` supports
+multi-dataset. Since ``(2)`` uses more data, it has a lower WER than ``(1)`` but it needs
+more training time.
+
+We use ``lstm_transducer_stateless2`` as an example below.
+
+.. note::
+
+   You need to download the `GigaSpeech <https://github.com/SpeechColab/GigaSpeech>`_ dataset
+   to run ``(2)``. If you have only ``LibriSpeech`` dataset available, feel free to use ``(1)``.
+
+Data preparation
+----------------
+
+.. code-block:: bash
+
+  $ cd egs/librispeech/ASR
+  $ ./prepare.sh
+
+  # If you use (1), you can **skip** the following command
+  $ ./prepare_giga_speech.sh
+
+The script ``./prepare.sh`` handles the data preparation for you, **automagically**.
+All you need to do is to run it.
+
+The data preparation contains several stages, you can use the following two
+options:
+
+  - ``--stage``
+  - ``--stop-stage``
+
+to control which stage(s) should be run. By default, all stages are executed.
+
+
+For example,
+
+.. code-block:: bash
+
+  $ cd egs/librispeech/ASR
+  $ ./prepare.sh --stage 0 --stop-stage 0
+
+means to run only stage 0.
+
+To run stage 2 to stage 5, use:
+
+.. code-block:: bash
+
+  $ ./prepare.sh --stage 2 --stop-stage 5
+
+.. hint::
+
+  If you have pre-downloaded the `LibriSpeech <https://www.openslr.org/12>`_
+  dataset and the `musan <http://www.openslr.org/17/>`_ dataset, say,
+  they are saved in ``/tmp/LibriSpeech`` and ``/tmp/musan``, you can modify
+  the ``dl_dir`` variable in ``./prepare.sh`` to point to ``/tmp`` so that
+  ``./prepare.sh`` won't re-download them.
+
+.. note::
+
+  All generated files by ``./prepare.sh``, e.g., features, lexicon, etc,
+  are saved in ``./data`` directory.
+
+We provide the following YouTube video showing how to run ``./prepare.sh``.
+
+.. note::
+
+   To get the latest news of `next-gen Kaldi <https://github.com/k2-fsa>`_, please subscribe
+   the following YouTube channel by `Nadira Povey <https://www.youtube.com/channel/UC_VaumpkmINz1pNkFXAN9mw>`_:
+
+      `<https://www.youtube.com/channel/UC_VaumpkmINz1pNkFXAN9mw>`_
+
+..  youtube:: ofEIoJL-mGM
+
+Training
+--------
+
+Configurable options
+~~~~~~~~~~~~~~~~~~~~
+
+.. code-block:: bash
+
+  $ cd egs/librispeech/ASR
+  $ ./lstm_transducer_stateless2/train.py --help
+
+shows you the training options that can be passed from the commandline.
+The following options are used quite often:
+
+  - ``--full-libri``
+
+    If it's True, the training part uses all the training data, i.e.,
+    960 hours. Otherwise, the training part uses only the subset
+    ``train-clean-100``, which has 100 hours of training data.
+
+    .. CAUTION::
+
+      The training set is perturbed by speed with two factors: 0.9 and 1.1.
+      If ``--full-libri`` is True, each epoch actually processes
+      ``3x960 == 2880`` hours of data.
+
+  - ``--num-epochs``
+
+    It is the number of epochs to train. For instance,
+    ``./lstm_transducer_stateless2/train.py --num-epochs 30`` trains for 30 epochs
+    and generates ``epoch-1.pt``, ``epoch-2.pt``, ..., ``epoch-30.pt``
+    in the folder ``./lstm_transducer_stateless2/exp``.
+
+  - ``--start-epoch``
+
+    It's used to resume training.
+    ``./lstm_transducer_stateless2/train.py --start-epoch 10`` loads the
+    checkpoint ``./lstm_transducer_stateless2/exp/epoch-9.pt`` and starts
+    training from epoch 10, based on the state from epoch 9.
+
+  - ``--world-size``
+
+    It is used for multi-GPU single-machine DDP training.
+
+      - (a) If it is 1, then no DDP training is used.
+
+      - (b) If it is 2, then GPU 0 and GPU 1 are used for DDP training.
+
+    The following shows some use cases with it.
+
+      **Use case 1**: You have 4 GPUs, but you only want to use GPU 0 and
+      GPU 2 for training. You can do the following:
+
+        .. code-block:: bash
+
+          $ cd egs/librispeech/ASR
+          $ export CUDA_VISIBLE_DEVICES="0,2"
+          $ ./lstm_transducer_stateless2/train.py --world-size 2
+
+      **Use case 2**: You have 4 GPUs and you want to use all of them
+      for training. You can do the following:
+
+        .. code-block:: bash
+
+          $ cd egs/librispeech/ASR
+          $ ./lstm_transducer_stateless2/train.py --world-size 4
+
+      **Use case 3**: You have 4 GPUs but you only want to use GPU 3
+      for training. You can do the following:
+
+        .. code-block:: bash
+
+          $ cd egs/librispeech/ASR
+          $ export CUDA_VISIBLE_DEVICES="3"
+          $ ./lstm_transducer_stateless2/train.py --world-size 1
+
+    .. caution::
+
+      Only multi-GPU single-machine DDP training is implemented at present.
+      Multi-GPU multi-machine DDP training will be added later.
+
+  - ``--max-duration``
+
+    It specifies the number of seconds over all utterances in a
+    batch, before **padding**.
+    If you encounter CUDA OOM, please reduce it.
+
+    .. HINT::
+
+      Due to padding, the number of seconds of all utterances in a
+      batch will usually be larger than ``--max-duration``.
+
+      A larger value for ``--max-duration`` may cause OOM during training,
+      while a smaller value may increase the training time. You have to
+      tune it.
+
+  - ``--giga-prob``
+
+    The probability to select a batch from the ``GigaSpeech`` dataset.
+    Note: It is available only for ``(2)``.
+
+Pre-configured options
+~~~~~~~~~~~~~~~~~~~~~~
+
+There are some training options, e.g., weight decay,
+number of warmup steps, results dir, etc,
+that are not passed from the commandline.
+They are pre-configured by the function ``get_params()`` in
+`lstm_transducer_stateless2/train.py <https://github.com/k2-fsa/icefall/blob/master/egs/librispeech/ASR/lstm_transducer_stateless2/train.py>`_
+
+You don't need to change these pre-configured parameters. If you really need to change
+them, please modify ``./lstm_transducer_stateless2/train.py`` directly.
+
+Training logs
+~~~~~~~~~~~~~
+
+Training logs and checkpoints are saved in ``lstm_transducer_stateless2/exp``.
+You will find the following files in that directory:
+
+  - ``epoch-1.pt``, ``epoch-2.pt``, ...
+
+    These are checkpoint files saved at the end of each epoch, containing model
+    ``state_dict`` and optimizer ``state_dict``.
+    To resume training from some checkpoint, say ``epoch-10.pt``, you can use:
+
+      .. code-block:: bash
+
+        $ ./lstm_transducer_stateless2/train.py --start-epoch 11
+
+  - ``checkpoint-436000.pt``, ``checkpoint-438000.pt``, ...
+
+    These are checkpoint files saved every ``--save-every-n`` batches,
+    containing model ``state_dict`` and optimizer ``state_dict``.
+    To resume training from some checkpoint, say ``checkpoint-436000``, you can use:
+
+      .. code-block:: bash
+
+        $ ./lstm_transducer_stateless2/train.py --start-batch 436000
+
+  - ``tensorboard/``
+
+    This folder contains TensorBoard logs. Training loss, validation loss, learning
+    rate, etc, are recorded in these logs. You can visualize them by:
+
+      .. code-block:: bash
+
+        $ cd lstm_transducer_stateless2/exp/tensorboard
+        $ tensorboard dev upload --logdir . --description "LSTM transducer training for LibriSpeech with icefall"
+
+    It will print something like below:
+
+      .. code-block::
+
+        TensorFlow installation not found - running with reduced feature set.
+        Upload started and will continue reading any new data as it's added to the logdir.
+
+        To stop uploading, press Ctrl-C.
+
+        New experiment created. View your TensorBoard at: https://tensorboard.dev/experiment/cj2vtPiwQHKN9Q1tx6PTpg/
+
+        [2022-09-20T15:50:50] Started scanning logdir.
+        Uploading 4468 scalars...
+        [2022-09-20T15:53:02] Total uploaded: 210171 scalars, 0 tensors, 0 binary objects
+        Listening for new data in logdir...
+
+    Note there is a URL in the above output, click it and you will see
+    the following screenshot:
+
+      .. figure:: images/librispeech-lstm-transducer-tensorboard-log.png
+         :width: 600
+         :alt: TensorBoard screenshot
+         :align: center
+         :target: https://tensorboard.dev/experiment/lzGnETjwRxC3yghNMd4kPw/
+
+         TensorBoard screenshot.
+
+  .. hint::
+
+    If you don't have access to google, you can use the following command
+    to view the tensorboard log locally:
+
+      .. code-block:: bash
+
+        cd lstm_transducer_stateless2/exp/tensorboard
+        tensorboard --logdir . --port 6008
+
+    It will print the following message:
+
+      .. code-block::
+
+        Serving TensorBoard on localhost; to expose to the network, use a proxy or pass --bind_all
+        TensorBoard 2.8.0 at http://localhost:6008/ (Press CTRL+C to quit)
+
+    Now start your browser and go to `<http://localhost:6008>`_ to view the tensorboard
+    logs.
+
+
+  - ``log/log-train-xxxx``
+
+    It is the detailed training log in text format, same as the one
+    you saw printed to the console during training.
+
+Usage example
+~~~~~~~~~~~~~
+
+You can use the following command to start the training using 8 GPUs:
+
+.. code-block:: bash
+
+  export CUDA_VISIBLE_DEVICES="0,1,2,3,4,5,6,7"
+  ./lstm_transducer_stateless2/train.py \
+    --world-size 8 \
+    --num-epochs 35 \
+    --start-epoch 1 \
+    --full-libri 1 \
+    --exp-dir lstm_transducer_stateless2/exp \
+    --max-duration 500 \
+    --use-fp16 0 \
+    --lr-epochs 10 \
+    --num-workers 2 \
+    --giga-prob 0.9
+
+Decoding
+--------
+
+The decoding part uses checkpoints saved by the training part, so you have
+to run the training part first.
+
+.. hint::
+
+   There are two kinds of checkpoints:
+
+    - (1) ``epoch-1.pt``, ``epoch-2.pt``, ..., which are saved at the end
+      of each epoch. You can pass ``--epoch`` to
+      ``lstm_transducer_stateless2/decode.py`` to use them.
+
+    - (2) ``checkpoints-436000.pt``, ``epoch-438000.pt``, ..., which are saved
+      every ``--save-every-n`` batches. You can pass ``--iter`` to
+      ``lstm_transducer_stateless2/decode.py`` to use them.
+
+    We suggest that you try both types of checkpoints and choose the one
+    that produces the lowest WERs.
+
+.. code-block:: bash
+
+  $ cd egs/librispeech/ASR
+  $ ./lstm_transducer_stateless2/decode.py --help
+
+shows the options for decoding.
+
+The following shows two examples:
+
+.. code-block:: bash
+
+  for m in greedy_search fast_beam_search modified_beam_search; do
+    for epoch in 17; do
+      for avg in 1 2; do
+        ./lstm_transducer_stateless2/decode.py \
+          --epoch $epoch \
+          --avg $avg \
+          --exp-dir lstm_transducer_stateless2/exp \
+          --max-duration 600 \
+          --num-encoder-layers 12 \
+          --rnn-hidden-size 1024 \
+          --decoding-method $m \
+          --use-averaged-model True \
+          --beam 4 \
+          --max-contexts 4 \
+          --max-states 8 \
+          --beam-size 4
+      done
+    done
+  done
+
+
+.. code-block:: bash
+
+  for m in greedy_search fast_beam_search modified_beam_search; do
+    for iter in 474000; do
+      for avg in 8 10 12 14 16 18; do
+        ./lstm_transducer_stateless2/decode.py \
+          --iter $iter \
+          --avg $avg \
+          --exp-dir lstm_transducer_stateless2/exp \
+          --max-duration 600 \
+          --num-encoder-layers 12 \
+          --rnn-hidden-size 1024 \
+          --decoding-method $m \
+          --use-averaged-model True \
+          --beam 4 \
+          --max-contexts 4 \
+          --max-states 8 \
+          --beam-size 4
+      done
+    done
+  done
+
+Export models
+-------------
+
+`lstm_transducer_stateless2/export.py <https://github.com/k2-fsa/icefall/blob/master/egs/librispeech/ASR/lstm_transducer_stateless2/export.py>`_ supports to export checkpoints from ``lstm_transducer_stateless2/exp`` in the following ways.
+
+Export ``model.state_dict()``
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Checkpoints saved by ``lstm_transducer_stateless2/train.py`` also include
+``optimizer.state_dict()``. It is useful for resuming training. But after training,
+we are interested only in ``model.state_dict()``. You can use the following
+command to extract ``model.state_dict()``.
+
+.. code-block:: bash
+
+  # Assume that --iter 468000 --avg 16 produces the smallest WER
+  # (You can get such information after running ./lstm_transducer_stateless2/decode.py)
+
+  iter=468000
+  avg=16
+
+  ./lstm_transducer_stateless2/export.py \
+    --exp-dir ./lstm_transducer_stateless2/exp \
+    --bpe-model data/lang_bpe_500/bpe.model \
+    --iter $iter \
+    --avg  $avg
+
+It will generate a file ``./lstm_transducer_stateless2/exp/pretrained.pt``.
+
+.. hint::
+
+   To use the generated ``pretrained.pt`` for ``lstm_transducer_stateless2/decode.py``,
+   you can run:
+
+   .. code-block:: bash
+
+      cd lstm_transducer_stateless2/exp
+      ln -s pretrained epoch-9999.pt
+
+   And then pass `--epoch 9999 --avg 1 --use-averaged-model 0` to
+   ``./lstm_transducer_stateless2/decode.py``.
+
+To use the exported model with ``./lstm_transducer_stateless2/pretrained.py``, you
+can run:
+
+.. code-block:: bash
+
+  ./lstm_transducer_stateless2/pretrained.py \
+    --checkpoint ./lstm_transducer_stateless2/exp/pretrained.pt \
+    --bpe-model ./data/lang_bpe_500/bpe.model \
+    --method greedy_search \
+    /path/to/foo.wav \
+    /path/to/bar.wav
+
+Export model using ``torch.jit.trace()``
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+.. code-block:: bash
+
+  iter=468000
+  avg=16
+
+  ./lstm_transducer_stateless2/export.py \
+    --exp-dir ./lstm_transducer_stateless2/exp \
+    --bpe-model data/lang_bpe_500/bpe.model \
+    --iter $iter \
+    --avg  $avg \
+    --jit-trace 1
+
+It will generate 3 files:
+
+  - ``./lstm_transducer_stateless2/exp/encoder_jit_trace.pt``
+  - ``./lstm_transducer_stateless2/exp/decoder_jit_trace.pt``
+  - ``./lstm_transducer_stateless2/exp/joiner_jit_trace.pt``
+
+To use the generated files with ``./lstm_transducer_stateless2/jit_pretrained``:
+
+.. code-block:: bash
+
+  ./lstm_transducer_stateless2/jit_pretrained.py \
+    --bpe-model ./data/lang_bpe_500/bpe.model \
+    --encoder-model-filename ./lstm_transducer_stateless2/exp/encoder_jit_trace.pt \
+    --decoder-model-filename ./lstm_transducer_stateless2/exp/decoder_jit_trace.pt \
+    --joiner-model-filename ./lstm_transducer_stateless2/exp/joiner_jit_trace.pt \
+    /path/to/foo.wav \
+    /path/to/bar.wav
+
+Export model for ncnn
+~~~~~~~~~~~~~~~~~~~~~
+
+We support exporting pretrained LSTM transducer models to
+`ncnn <https://github.com/tencent/ncnn>`_ using
+`pnnx <https://github.com/Tencent/ncnn/tree/master/tools/pnnx>`_.
+
+First, let us install a modified version of ``ncnn``:
+
+.. code-block:: bash
+
+  git clone https://github.com/csukuangfj/ncnn
+  cd ncnn
+  git submodule update --recursive --init
+  python3 setup.py bdist_wheel
+  ls -lh dist/
+  pip install ./dist/*.whl
+
+  # now build pnnx
+  cd tools/pnnx
+  mkdir build
+  cd build
+  make -j4
+  export PATH=$PWD/src:$PATH
+
+  ./src/pnnx
+
+.. note::
+
+   We assume that you have added the path to the binary ``pnnx`` to the
+   environment variable ``PATH``.
+
+Second, let us export the model using ``torch.jit.trace()`` that is suitable
+for ``pnnx``:
+
+.. code-block:: bash
+
+  iter=468000
+  avg=16
+
+  ./lstm_transducer_stateless2/export.py \
+    --exp-dir ./lstm_transducer_stateless2/exp \
+    --bpe-model data/lang_bpe_500/bpe.model \
+    --iter $iter \
+    --avg  $avg \
+    --pnnx 1
+
+It will generate 3 files:
+
+  - ``./lstm_transducer_stateless2/exp/encoder_jit_trace-pnnx.pt``
+  - ``./lstm_transducer_stateless2/exp/decoder_jit_trace-pnnx.pt``
+  - ``./lstm_transducer_stateless2/exp/joiner_jit_trace-pnnx.pt``
+
+Third, convert torchscript model to ``ncnn`` format:
+
+.. code-block::
+
+   pnnx ./lstm_transducer_stateless2/exp/encoder_jit_trace-pnnx.pt
+   pnnx ./lstm_transducer_stateless2/exp/decoder_jit_trace-pnnx.pt
+   pnnx ./lstm_transducer_stateless2/exp/joiner_jit_trace-pnnx.pt
+
+It will generate the following files:
+
+  - ``./lstm_transducer_stateless2/exp/encoder_jit_trace-pnnx.ncnn.param``
+  - ``./lstm_transducer_stateless2/exp/encoder_jit_trace-pnnx.ncnn.bin``
+  - ``./lstm_transducer_stateless2/exp/decoder_jit_trace-pnnx.ncnn.param``
+  - ``./lstm_transducer_stateless2/exp/decoder_jit_trace-pnnx.ncnn.bin``
+  - ``./lstm_transducer_stateless2/exp/joiner_jit_trace-pnnx.ncnn.param``
+  - ``./lstm_transducer_stateless2/exp/joiner_jit_trace-pnnx.ncnn.bin``
+
+To use the above generate files, run:
+
+.. code-block:: bash
+
+./lstm_transducer_stateless2/ncnn-decode.py \
+ --bpe-model-filename ./data/lang_bpe_500/bpe.model \
+ --encoder-param-filename ./lstm_transducer_stateless2/exp/encoder_jit_trace-pnnx.ncnn.param \
+ --encoder-bin-filename ./lstm_transducer_stateless2/exp/encoder_jit_trace-pnnx.ncnn.bin \
+ --decoder-param-filename ./lstm_transducer_stateless2/exp/decoder_jit_trace-pnnx.ncnn.param \
+ --decoder-bin-filename ./lstm_transducer_stateless2/exp/decoder_jit_trace-pnnx.ncnn.bin \
+ --joiner-param-filename ./lstm_transducer_stateless2/exp/joiner_jit_trace-pnnx.ncnn.param \
+ --joiner-bin-filename ./lstm_transducer_stateless2/exp/joiner_jit_trace-pnnx.ncnn.bin \
+ /path/to/foo.wav
+
+.. code-block:: bash
+
+./lstm_transducer_stateless2/streaming-ncnn-decode.py \
+ --bpe-model-filename ./data/lang_bpe_500/bpe.model \
+ --encoder-param-filename ./lstm_transducer_stateless2/exp/encoder_jit_trace-pnnx.ncnn.param \
+ --encoder-bin-filename ./lstm_transducer_stateless2/exp/encoder_jit_trace-pnnx.ncnn.bin \
+ --decoder-param-filename ./lstm_transducer_stateless2/exp/decoder_jit_trace-pnnx.ncnn.param \
+ --decoder-bin-filename ./lstm_transducer_stateless2/exp/decoder_jit_trace-pnnx.ncnn.bin \
+ --joiner-param-filename ./lstm_transducer_stateless2/exp/joiner_jit_trace-pnnx.ncnn.param \
+ --joiner-bin-filename ./lstm_transducer_stateless2/exp/joiner_jit_trace-pnnx.ncnn.bin \
+ /path/to/foo.wav
+
+To use the above generated files in C++, please see
+`<https://github.com/k2-fsa/sherpa-ncnn>`_
+
+It is able to generate a static linked library that can be run on Linux, Windows,
+macOS, Raspberry Pi, etc.
+
+Download pretrained models
+--------------------------
+
+If you don't want to train from scratch, you can download the pretrained models
+by visiting the following links:
+
+  - `<https://huggingface.co/csukuangfj/icefall-asr-librispeech-lstm-transducer-stateless2-2022-09-03>`_
+
+  - `<https://huggingface.co/Zengwei/icefall-asr-librispeech-lstm-transducer-stateless-2022-08-18>`_
+
+  See `<https://github.com/k2-fsa/icefall/blob/master/egs/librispeech/ASR/RESULTS.md>`_
+  for the details of the above pretrained models
+
+You can find more usages of the pretrained models in
+`<https://k2-fsa.github.io/sherpa/python/streaming_asr/lstm/index.html>`_

egs/librispeech/ASR/lstm_transducer_stateless/lstm.py

@@ -116,6 +116,8 @@ class RNN(EncoderInterface):
         Period of auxiliary layers used for random combiner during training.
         If set to 0, will not use the random combiner (Default).
         You can set a positive integer to use the random combiner, e.g., 3.
+      is_pnnx:
+        True to make this class exportable via PNNX.
     """
 
     def __init__(
@@ -129,6 +131,7 @@ class RNN(EncoderInterface):
         dropout: float = 0.1,
         layer_dropout: float = 0.075,
         aux_layer_period: int = 0,
+        is_pnnx: bool = False,
     ) -> None:
         super(RNN, self).__init__()
 
@@ -142,7 +145,13 @@ class RNN(EncoderInterface):
         # That is, it does two things simultaneously:
         #   (1) subsampling: T -> T//subsampling_factor
         #   (2) embedding: num_features -> d_model
-        self.encoder_embed = Conv2dSubsampling(num_features, d_model)
+        self.encoder_embed = Conv2dSubsampling(
+            num_features,
+            d_model,
+            is_pnnx=is_pnnx,
+        )
+
+        self.is_pnnx = is_pnnx
 
         self.num_encoder_layers = num_encoder_layers
         self.d_model = d_model
@@ -209,7 +218,13 @@ class RNN(EncoderInterface):
         # lengths = ((x_lens - 3) // 2 - 1) // 2 # issue an warning
         #
         # Note: rounding_mode in torch.div() is available only in torch >= 1.8.0
-        lengths = (((x_lens - 3) >> 1) - 1) >> 1
+        if not self.is_pnnx:
+            lengths = (((x_lens - 3) >> 1) - 1) >> 1
+        else:
+            lengths1 = torch.floor((x_lens - 3) / 2)
+            lengths = torch.floor((lengths1 - 1) / 2)
+            lengths = lengths.to(x_lens)
+
         if not torch.jit.is_tracing():
             assert x.size(0) == lengths.max().item()
 
@@ -359,7 +374,7 @@ class RNNEncoderLayer(nn.Module):
                 # for cell state
                 assert states[1].shape == (1, src.size(1), self.rnn_hidden_size)
             src_lstm, new_states = self.lstm(src, states)
-        src = src + self.dropout(src_lstm)
+        src = self.dropout(src_lstm) + src
 
         # feed forward module
         src = src + self.dropout(self.feed_forward(src))
@@ -505,6 +520,7 @@ class Conv2dSubsampling(nn.Module):
         layer1_channels: int = 8,
         layer2_channels: int = 32,
         layer3_channels: int = 128,
+        is_pnnx: bool = False,
     ) -> None:
         """
         Args:
@@ -517,6 +533,9 @@ class Conv2dSubsampling(nn.Module):
             Number of channels in layer1
           layer1_channels:
             Number of channels in layer2
+          is_pnnx:
+            True if we are converting the model to PNNX format.
+            False otherwise.
         """
         assert in_channels >= 9
         super().__init__()
@@ -559,6 +578,10 @@ class Conv2dSubsampling(nn.Module):
             channel_dim=-1, min_positive=0.45, max_positive=0.55
         )
 
+        # ncnn supports only batch size == 1
+        self.is_pnnx = is_pnnx
+        self.conv_out_dim = self.out.weight.shape[1]
+
     def forward(self, x: torch.Tensor) -> torch.Tensor:
         """Subsample x.
 
@@ -572,9 +595,15 @@ class Conv2dSubsampling(nn.Module):
         # On entry, x is (N, T, idim)
         x = x.unsqueeze(1)  # (N, T, idim) -> (N, 1, T, idim) i.e., (N, C, H, W)
         x = self.conv(x)
-        # Now x is of shape (N, odim, ((T-3)//2-1)//2, ((idim-3)//2-1)//2)
+
-        b, c, t, f = x.size()
+        if torch.jit.is_tracing() and self.is_pnnx:
-        x = self.out(x.transpose(1, 2).contiguous().view(b, t, c * f))
+            x = x.permute(0, 2, 1, 3).reshape(1, -1, self.conv_out_dim)
+            x = self.out(x)
+        else:
+            # Now x is of shape (N, odim, ((T-3)//2-1)//2, ((idim-3)//2-1)//2)
+            b, c, t, f = x.size()
+            x = self.out(x.transpose(1, 2).contiguous().view(b, t, c * f))
+
         # Now x is of shape (N, ((T-3)//2-1))//2, odim)
         x = self.out_norm(x)
         x = self.out_balancer(x)

egs/librispeech/ASR/lstm_transducer_stateless2/export.py

@@ -169,6 +169,18 @@ def get_parser():
         """,
     )
 
+    parser.add_argument(
+        "--pnnx",
+        type=str2bool,
+        default=False,
+        help="""True to save a model after applying torch.jit.trace for later
+        converting to PNNX. It will generate 3 files:
+         - encoder_jit_trace-pnnx.pt
+         - decoder_jit_trace-pnnx.pt
+         - joiner_jit_trace-pnnx.pt
+        """,
+    )
+
     parser.add_argument(
         "--context-size",
         type=int,
@@ -277,6 +289,10 @@ def main():
 
     logging.info(params)
 
+    if params.pnnx:
+        params.is_pnnx = params.pnnx
+        logging.info("For PNNX")
+
     logging.info("About to create model")
     model = get_transducer_model(params, enable_giga=False)
 
@@ -371,7 +387,18 @@ def main():
     model.to("cpu")
     model.eval()
 
-    if params.jit_trace is True:
+    if params.pnnx:
+        convert_scaled_to_non_scaled(model, inplace=True)
+        logging.info("Using torch.jit.trace()")
+        encoder_filename = params.exp_dir / "encoder_jit_trace-pnnx.pt"
+        export_encoder_model_jit_trace(model.encoder, encoder_filename)
+
+        decoder_filename = params.exp_dir / "decoder_jit_trace-pnnx.pt"
+        export_decoder_model_jit_trace(model.decoder, decoder_filename)
+
+        joiner_filename = params.exp_dir / "joiner_jit_trace-pnnx.pt"
+        export_joiner_model_jit_trace(model.joiner, joiner_filename)
+    elif params.jit_trace is True:
         convert_scaled_to_non_scaled(model, inplace=True)
         logging.info("Using torch.jit.trace()")
         encoder_filename = params.exp_dir / "encoder_jit_trace.pt"

egs/librispeech/ASR/lstm_transducer_stateless2/ncnn-decode.py

@@ -0,0 +1,295 @@
+#!/usr/bin/env python3
+# flake8: noqa
+#
+# Copyright      2022  Xiaomi Corp.        (authors: Fangjun Kuang, Zengwei Yao)
+#
+# See ../../../../LICENSE for clarification regarding multiple authors
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""
+Usage:
+  ./lstm_transducer_stateless2/ncnn-decode.py \
+   --bpe-model-filename ./data/lang_bpe_500/bpe.model \
+   --encoder-param-filename ./lstm_transducer_stateless2/exp/encoder_jit_trace-iter-468000-avg-16-pnnx.ncnn.param \
+   --encoder-bin-filename ./lstm_transducer_stateless2/exp/encoder_jit_trace-iter-468000-avg-16-pnnx.ncnn.bin \
+   --decoder-param-filename ./lstm_transducer_stateless2/exp/decoder_jit_trace-iter-468000-avg-16-pnnx.ncnn.param \
+   --decoder-bin-filename ./lstm_transducer_stateless2/exp/decoder_jit_trace-iter-468000-avg-16-pnnx.ncnn.bin \
+   --joiner-param-filename ./lstm_transducer_stateless2/exp/joiner_jit_trace-iter-468000-avg-16-pnnx.ncnn.param \
+   --joiner-bin-filename ./lstm_transducer_stateless2/exp/joiner_jit_trace-iter-468000-avg-16-pnnx.ncnn.bin \
+   ./test_wavs/1089-134686-0001.wav
+"""
+
+import argparse
+import logging
+from typing import List
+
+import kaldifeat
+import ncnn
+import sentencepiece as spm
+import torch
+import torchaudio
+
+
+def get_args():
+    parser = argparse.ArgumentParser()
+
+    parser.add_argument(
+        "--bpe-model-filename",
+        type=str,
+        help="Path to bpe.model",
+    )
+
+    parser.add_argument(
+        "--encoder-param-filename",
+        type=str,
+        help="Path to encoder.ncnn.param",
+    )
+
+    parser.add_argument(
+        "--encoder-bin-filename",
+        type=str,
+        help="Path to encoder.ncnn.bin",
+    )
+
+    parser.add_argument(
+        "--decoder-param-filename",
+        type=str,
+        help="Path to decoder.ncnn.param",
+    )
+
+    parser.add_argument(
+        "--decoder-bin-filename",
+        type=str,
+        help="Path to decoder.ncnn.bin",
+    )
+
+    parser.add_argument(
+        "--joiner-param-filename",
+        type=str,
+        help="Path to joiner.ncnn.param",
+    )
+
+    parser.add_argument(
+        "--joiner-bin-filename",
+        type=str,
+        help="Path to joiner.ncnn.bin",
+    )
+
+    parser.add_argument(
+        "sound_filename",
+        type=str,
+        help="Path to foo.wav",
+    )
+
+    return parser.parse_args()
+
+
+class Model:
+    def __init__(self, args):
+        self.init_encoder(args)
+        self.init_decoder(args)
+        self.init_joiner(args)
+
+    def init_encoder(self, args):
+        encoder_net = ncnn.Net()
+        encoder_net.opt.use_packing_layout = False
+        encoder_net.opt.use_fp16_storage = False
+        encoder_param = args.encoder_param_filename
+        encoder_model = args.encoder_bin_filename
+
+        encoder_net.load_param(encoder_param)
+        encoder_net.load_model(encoder_model)
+
+        self.encoder_net = encoder_net
+
+    def init_decoder(self, args):
+        decoder_param = args.decoder_param_filename
+        decoder_model = args.decoder_bin_filename
+
+        decoder_net = ncnn.Net()
+        decoder_net.opt.use_packing_layout = False
+
+        decoder_net.load_param(decoder_param)
+        decoder_net.load_model(decoder_model)
+
+        self.decoder_net = decoder_net
+
+    def init_joiner(self, args):
+        joiner_param = args.joiner_param_filename
+        joiner_model = args.joiner_bin_filename
+        joiner_net = ncnn.Net()
+        joiner_net.opt.use_packing_layout = False
+        joiner_net.load_param(joiner_param)
+        joiner_net.load_model(joiner_model)
+
+        self.joiner_net = joiner_net
+
+    def run_encoder(self, x, states):
+        with self.encoder_net.create_extractor() as ex:
+            ex.set_num_threads(10)
+            ex.input("in0", ncnn.Mat(x.numpy()).clone())
+            x_lens = torch.tensor([x.size(0)], dtype=torch.float32)
+            ex.input("in1", ncnn.Mat(x_lens.numpy()).clone())
+            ex.input("in2", ncnn.Mat(states[0].numpy()).clone())
+            ex.input("in3", ncnn.Mat(states[1].numpy()).clone())
+
+            ret, ncnn_out0 = ex.extract("out0")
+            assert ret == 0, ret
+
+            ret, ncnn_out1 = ex.extract("out1")
+            assert ret == 0, ret
+
+            ret, ncnn_out2 = ex.extract("out2")
+            assert ret == 0, ret
+
+            ret, ncnn_out3 = ex.extract("out3")
+            assert ret == 0, ret
+
+            encoder_out = torch.from_numpy(ncnn_out0.numpy()).clone()
+            encoder_out_lens = torch.from_numpy(ncnn_out1.numpy()).to(
+                torch.int32
+            )
+            hx = torch.from_numpy(ncnn_out2.numpy()).clone()
+            cx = torch.from_numpy(ncnn_out3.numpy()).clone()
+            return encoder_out, encoder_out_lens, hx, cx
+
+    def run_decoder(self, decoder_input):
+        assert decoder_input.dtype == torch.int32
+
+        with self.decoder_net.create_extractor() as ex:
+            ex.set_num_threads(10)
+            ex.input("in0", ncnn.Mat(decoder_input.numpy()).clone())
+            ret, ncnn_out0 = ex.extract("out0")
+            assert ret == 0, ret
+            decoder_out = torch.from_numpy(ncnn_out0.numpy()).clone()
+            return decoder_out
+
+    def run_joiner(self, encoder_out, decoder_out):
+        with self.joiner_net.create_extractor() as ex:
+            ex.set_num_threads(10)
+            ex.input("in0", ncnn.Mat(encoder_out.numpy()).clone())
+            ex.input("in1", ncnn.Mat(decoder_out.numpy()).clone())
+            ret, ncnn_out0 = ex.extract("out0")
+            assert ret == 0, ret
+            joiner_out = torch.from_numpy(ncnn_out0.numpy()).clone()
+            return joiner_out
+
+
+def read_sound_files(
+    filenames: List[str], expected_sample_rate: float
+) -> List[torch.Tensor]:
+    """Read a list of sound files into a list 1-D float32 torch tensors.
+    Args:
+      filenames:
+        A list of sound filenames.
+      expected_sample_rate:
+        The expected sample rate of the sound files.
+    Returns:
+      Return a list of 1-D float32 torch tensors.
+    """
+    ans = []
+    for f in filenames:
+        wave, sample_rate = torchaudio.load(f)
+        assert sample_rate == expected_sample_rate, (
+            f"expected sample rate: {expected_sample_rate}. "
+            f"Given: {sample_rate}"
+        )
+        # We use only the first channel
+        ans.append(wave[0])
+    return ans
+
+
+def greedy_search(model: Model, encoder_out: torch.Tensor):
+    assert encoder_out.ndim == 2
+    T = encoder_out.size(0)
+
+    context_size = 2
+    blank_id = 0  # hard-code to 0
+    hyp = [blank_id] * context_size
+
+    decoder_input = torch.tensor(hyp, dtype=torch.int32)  # (1, context_size)
+
+    decoder_out = model.run_decoder(decoder_input).squeeze(0)
+    #  print(decoder_out.shape)  # (512,)
+
+    for t in range(T):
+        encoder_out_t = encoder_out[t]
+        joiner_out = model.run_joiner(encoder_out_t, decoder_out)
+        #  print(joiner_out.shape) # [500]
+        y = joiner_out.argmax(dim=0).tolist()
+        if y != blank_id:
+            hyp.append(y)
+            decoder_input = hyp[-context_size:]
+            decoder_input = torch.tensor(decoder_input, dtype=torch.int32)
+            decoder_out = model.run_decoder(decoder_input).squeeze(0)
+    return hyp[context_size:]
+
+
+def main():
+    args = get_args()
+    logging.info(vars(args))
+
+    model = Model(args)
+
+    sp = spm.SentencePieceProcessor()
+    sp.load(args.bpe_model_filename)
+
+    sound_file = args.sound_filename
+
+    sample_rate = 16000
+
+    logging.info("Constructing Fbank computer")
+    opts = kaldifeat.FbankOptions()
+    opts.device = "cpu"
+    opts.frame_opts.dither = 0
+    opts.frame_opts.snip_edges = False
+    opts.frame_opts.samp_freq = sample_rate
+    opts.mel_opts.num_bins = 80
+
+    fbank = kaldifeat.Fbank(opts)
+
+    logging.info(f"Reading sound files: {sound_file}")
+    wave_samples = read_sound_files(
+        filenames=[sound_file],
+        expected_sample_rate=sample_rate,
+    )[0]
+
+    logging.info("Decoding started")
+    features = fbank(wave_samples)
+
+    num_encoder_layers = 12
+    d_model = 512
+    rnn_hidden_size = 1024
+
+    states = (
+        torch.zeros(num_encoder_layers, d_model),
+        torch.zeros(
+            num_encoder_layers,
+            rnn_hidden_size,
+        ),
+    )
+
+    encoder_out, encoder_out_lens, hx, cx = model.run_encoder(features, states)
+    hyp = greedy_search(model, encoder_out)
+    logging.info(sound_file)
+    logging.info(sp.decode(hyp))
+
+
+if __name__ == "__main__":
+    formatter = (
+        "%(asctime)s %(levelname)s [%(filename)s:%(lineno)d] %(message)s"
+    )
+
+    logging.basicConfig(format=formatter, level=logging.INFO)
+
+    main()

egs/librispeech/ASR/lstm_transducer_stateless2/streaming-ncnn-decode.py

@@ -0,0 +1,353 @@
+#!/usr/bin/env python3
+# flake8: noqa
+#
+# Copyright      2022  Xiaomi Corp.        (authors: Fangjun Kuang, Zengwei Yao)
+#
+# See ../../../../LICENSE for clarification regarding multiple authors
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import argparse
+import logging
+from typing import List, Optional
+
+import ncnn
+import sentencepiece as spm
+import torch
+import torchaudio
+from kaldifeat import FbankOptions, OnlineFbank, OnlineFeature
+
+
+def get_args():
+    parser = argparse.ArgumentParser()
+
+    parser.add_argument(
+        "--bpe-model-filename",
+        type=str,
+        help="Path to bpe.model",
+    )
+
+    parser.add_argument(
+        "--encoder-param-filename",
+        type=str,
+        help="Path to encoder.ncnn.param",
+    )
+
+    parser.add_argument(
+        "--encoder-bin-filename",
+        type=str,
+        help="Path to encoder.ncnn.bin",
+    )
+
+    parser.add_argument(
+        "--decoder-param-filename",
+        type=str,
+        help="Path to decoder.ncnn.param",
+    )
+
+    parser.add_argument(
+        "--decoder-bin-filename",
+        type=str,
+        help="Path to decoder.ncnn.bin",
+    )
+
+    parser.add_argument(
+        "--joiner-param-filename",
+        type=str,
+        help="Path to joiner.ncnn.param",
+    )
+
+    parser.add_argument(
+        "--joiner-bin-filename",
+        type=str,
+        help="Path to joiner.ncnn.bin",
+    )
+
+    parser.add_argument(
+        "sound_filename",
+        type=str,
+        help="Path to foo.wav",
+    )
+
+    return parser.parse_args()
+
+
+class Model:
+    def __init__(self, args):
+        self.init_encoder(args)
+        self.init_decoder(args)
+        self.init_joiner(args)
+
+    def init_encoder(self, args):
+        encoder_net = ncnn.Net()
+        encoder_net.opt.use_packing_layout = False
+        encoder_net.opt.use_fp16_storage = False
+        encoder_param = args.encoder_param_filename
+        encoder_model = args.encoder_bin_filename
+
+        encoder_net.load_param(encoder_param)
+        encoder_net.load_model(encoder_model)
+
+        self.encoder_net = encoder_net
+
+    def init_decoder(self, args):
+        decoder_param = args.decoder_param_filename
+        decoder_model = args.decoder_bin_filename
+
+        decoder_net = ncnn.Net()
+        decoder_net.opt.use_packing_layout = False
+
+        decoder_net.load_param(decoder_param)
+        decoder_net.load_model(decoder_model)
+
+        self.decoder_net = decoder_net
+
+    def init_joiner(self, args):
+        joiner_param = args.joiner_param_filename
+        joiner_model = args.joiner_bin_filename
+        joiner_net = ncnn.Net()
+        joiner_net.opt.use_packing_layout = False
+        joiner_net.load_param(joiner_param)
+        joiner_net.load_model(joiner_model)
+
+        self.joiner_net = joiner_net
+
+    def run_encoder(self, x, states):
+        with self.encoder_net.create_extractor() as ex:
+            #  ex.set_num_threads(10)
+            ex.input("in0", ncnn.Mat(x.numpy()).clone())
+            x_lens = torch.tensor([x.size(0)], dtype=torch.float32)
+            ex.input("in1", ncnn.Mat(x_lens.numpy()).clone())
+            ex.input("in2", ncnn.Mat(states[0].numpy()).clone())
+            ex.input("in3", ncnn.Mat(states[1].numpy()).clone())
+
+            ret, ncnn_out0 = ex.extract("out0")
+            assert ret == 0, ret
+
+            ret, ncnn_out1 = ex.extract("out1")
+            assert ret == 0, ret
+
+            ret, ncnn_out2 = ex.extract("out2")
+            assert ret == 0, ret
+
+            ret, ncnn_out3 = ex.extract("out3")
+            assert ret == 0, ret
+
+            encoder_out = torch.from_numpy(ncnn_out0.numpy()).clone()
+            encoder_out_lens = torch.from_numpy(ncnn_out1.numpy()).to(
+                torch.int32
+            )
+            hx = torch.from_numpy(ncnn_out2.numpy()).clone()
+            cx = torch.from_numpy(ncnn_out3.numpy()).clone()
+            return encoder_out, encoder_out_lens, hx, cx
+
+    def run_decoder(self, decoder_input):
+        assert decoder_input.dtype == torch.int32
+
+        with self.decoder_net.create_extractor() as ex:
+            #  ex.set_num_threads(10)
+            ex.input("in0", ncnn.Mat(decoder_input.numpy()).clone())
+            ret, ncnn_out0 = ex.extract("out0")
+            assert ret == 0, ret
+            decoder_out = torch.from_numpy(ncnn_out0.numpy()).clone()
+            return decoder_out
+
+    def run_joiner(self, encoder_out, decoder_out):
+        with self.joiner_net.create_extractor() as ex:
+            #  ex.set_num_threads(10)
+            ex.input("in0", ncnn.Mat(encoder_out.numpy()).clone())
+            ex.input("in1", ncnn.Mat(decoder_out.numpy()).clone())
+            ret, ncnn_out0 = ex.extract("out0")
+            assert ret == 0, ret
+            joiner_out = torch.from_numpy(ncnn_out0.numpy()).clone()
+            return joiner_out
+
+
+def read_sound_files(
+    filenames: List[str], expected_sample_rate: float
+) -> List[torch.Tensor]:
+    """Read a list of sound files into a list 1-D float32 torch tensors.
+    Args:
+      filenames:
+        A list of sound filenames.
+      expected_sample_rate:
+        The expected sample rate of the sound files.
+    Returns:
+      Return a list of 1-D float32 torch tensors.
+    """
+    ans = []
+    for f in filenames:
+        wave, sample_rate = torchaudio.load(f)
+        assert sample_rate == expected_sample_rate, (
+            f"expected sample rate: {expected_sample_rate}. "
+            f"Given: {sample_rate}"
+        )
+        # We use only the first channel
+        ans.append(wave[0])
+    return ans
+
+
+def create_streaming_feature_extractor() -> OnlineFeature:
+    """Create a CPU streaming feature extractor.
+
+    At present, we assume it returns a fbank feature extractor with
+    fixed options. In the future, we will support passing in the options
+    from outside.
+
+    Returns:
+      Return a CPU streaming feature extractor.
+    """
+    opts = FbankOptions()
+    opts.device = "cpu"
+    opts.frame_opts.dither = 0
+    opts.frame_opts.snip_edges = False
+    opts.frame_opts.samp_freq = 16000
+    opts.mel_opts.num_bins = 80
+    return OnlineFbank(opts)
+
+
+def greedy_search(
+    model: Model,
+    encoder_out: torch.Tensor,
+    decoder_out: Optional[torch.Tensor] = None,
+    hyp: Optional[List[int]] = None,
+):
+    assert encoder_out.ndim == 1
+    context_size = 2
+    blank_id = 0
+
+    if decoder_out is None:
+        assert hyp is None, hyp
+        hyp = [blank_id] * context_size
+        decoder_input = torch.tensor(
+            hyp, dtype=torch.int32
+        )  # (1, context_size)
+        decoder_out = model.run_decoder(decoder_input).squeeze(0)
+
+    else:
+        assert decoder_out.ndim == 1
+        assert hyp is not None, hyp
+
+    joiner_out = model.run_joiner(encoder_out, decoder_out)
+    y = joiner_out.argmax(dim=0).tolist()
+    if y != blank_id:
+        hyp.append(y)
+        decoder_input = hyp[-context_size:]
+        decoder_input = torch.tensor(decoder_input, dtype=torch.int32)
+        decoder_out = model.run_decoder(decoder_input).squeeze(0)
+
+    return hyp, decoder_out
+
+
+def main():
+    args = get_args()
+    logging.info(vars(args))
+
+    model = Model(args)
+
+    sp = spm.SentencePieceProcessor()
+    sp.load(args.bpe_model_filename)
+
+    sound_file = args.sound_filename
+
+    sample_rate = 16000
+
+    logging.info("Constructing Fbank computer")
+    online_fbank = create_streaming_feature_extractor()
+
+    logging.info(f"Reading sound files: {sound_file}")
+    wave_samples = read_sound_files(
+        filenames=[sound_file],
+        expected_sample_rate=sample_rate,
+    )[0]
+    logging.info(wave_samples.shape)
+
+    num_encoder_layers = 12
+    batch_size = 1
+    d_model = 512
+    rnn_hidden_size = 1024
+
+    states = (
+        torch.zeros(num_encoder_layers, batch_size, d_model),
+        torch.zeros(
+            num_encoder_layers,
+            batch_size,
+            rnn_hidden_size,
+        ),
+    )
+
+    hyp = None
+    decoder_out = None
+
+    num_processed_frames = 0
+    segment = 9
+    offset = 4
+
+    chunk = 3200  # 0.2 second
+
+    start = 0
+    while start < wave_samples.numel():
+        end = min(start + chunk, wave_samples.numel())
+        samples = wave_samples[start:end]
+        start += chunk
+
+        online_fbank.accept_waveform(
+            sampling_rate=sample_rate,
+            waveform=samples,
+        )
+        while online_fbank.num_frames_ready - num_processed_frames >= segment:
+            frames = []
+            for i in range(segment):
+                frames.append(online_fbank.get_frame(num_processed_frames + i))
+            num_processed_frames += offset
+            frames = torch.cat(frames, dim=0)
+            encoder_out, encoder_out_lens, hx, cx = model.run_encoder(
+                frames, states
+            )
+            states = (hx, cx)
+            hyp, decoder_out = greedy_search(
+                model, encoder_out.squeeze(0), decoder_out, hyp
+            )
+    online_fbank.accept_waveform(
+        sampling_rate=sample_rate, waveform=torch.zeros(8000, dtype=torch.int32)
+    )
+
+    online_fbank.input_finished()
+    while online_fbank.num_frames_ready - num_processed_frames >= segment:
+        frames = []
+        for i in range(segment):
+            frames.append(online_fbank.get_frame(num_processed_frames + i))
+        num_processed_frames += offset
+        frames = torch.cat(frames, dim=0)
+        encoder_out, encoder_out_lens, hx, cx = model.run_encoder(
+            frames, states
+        )
+        states = (hx, cx)
+        hyp, decoder_out = greedy_search(
+            model, encoder_out.squeeze(0), decoder_out, hyp
+        )
+
+    context_size = 2
+
+    logging.info(sound_file)
+    logging.info(sp.decode(hyp[context_size:]))
+
+
+if __name__ == "__main__":
+    formatter = (
+        "%(asctime)s %(levelname)s [%(filename)s:%(lineno)d] %(message)s"
+    )
+
+    logging.basicConfig(format=formatter, level=logging.INFO)
+
+    main()

egs/librispeech/ASR/lstm_transducer_stateless2/train.py

@@ -406,6 +406,8 @@ def get_params() -> AttributeDict:
             "decoder_dim": 512,
             # parameters for joiner
             "joiner_dim": 512,
+            # True to generate a model that can be exported via PNNX
+            "is_pnnx": False,
             # parameters for Noam
             "model_warm_step": 3000,  # arg given to model, not for lrate
             "env_info": get_env_info(),
@@ -424,6 +426,7 @@ def get_encoder_model(params: AttributeDict) -> nn.Module:
         dim_feedforward=params.dim_feedforward,
         num_encoder_layers=params.num_encoder_layers,
         aux_layer_period=params.aux_layer_period,
+        is_pnnx=params.is_pnnx,
     )
     return encoder
 

egs/librispeech/ASR/pruned_transducer_stateless3/scaling_converter.py

@@ -30,6 +30,7 @@ from typing import List
 import torch
 import torch.nn as nn
 from scaling import (
+    BasicNorm,
     ScaledConv1d,
     ScaledConv2d,
     ScaledEmbedding,
@@ -38,6 +39,29 @@ from scaling import (
 )
 
 
+class NonScaledNorm(nn.Module):
+    """See BasicNorm for doc"""
+
+    def __init__(
+        self,
+        num_channels: int,
+        eps_exp: float,
+        channel_dim: int = -1,  # CAUTION: see documentation.
+    ):
+        super().__init__()
+        self.num_channels = num_channels
+        self.channel_dim = channel_dim
+        self.eps_exp = eps_exp
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        if not torch.jit.is_tracing():
+            assert x.shape[self.channel_dim] == self.num_channels
+        scales = (
+            torch.mean(x * x, dim=self.channel_dim, keepdim=True) + self.eps_exp
+        ).pow(-0.5)
+        return x * scales
+
+
 def scaled_linear_to_linear(scaled_linear: ScaledLinear) -> nn.Linear:
     """Convert an instance of ScaledLinear to nn.Linear.
 
@@ -174,6 +198,16 @@ def scaled_embedding_to_embedding(
     return embedding
 
 
+def convert_basic_norm(basic_norm: BasicNorm) -> NonScaledNorm:
+    assert isinstance(basic_norm, BasicNorm), type(BasicNorm)
+    norm = NonScaledNorm(
+        num_channels=basic_norm.num_channels,
+        eps_exp=basic_norm.eps.data.exp().item(),
+        channel_dim=basic_norm.channel_dim,
+    )
+    return norm
+
+
 def scaled_lstm_to_lstm(scaled_lstm: ScaledLSTM) -> nn.LSTM:
     """Convert an instance of ScaledLSTM to nn.LSTM.
 
@@ -256,6 +290,8 @@ def convert_scaled_to_non_scaled(model: nn.Module, inplace: bool = False):
             d[name] = scaled_conv2d_to_conv2d(m)
         elif isinstance(m, ScaledEmbedding):
             d[name] = scaled_embedding_to_embedding(m)
+        elif isinstance(m, BasicNorm):
+            d[name] = convert_basic_norm(m)
         elif isinstance(m, ScaledLSTM):
             d[name] = scaled_lstm_to_lstm(m)