@scrypted/libav 1.0.167 → 1.0.169

package/opus/dnn/torch/lpcnet/scripts/loop_run.sh

@@ -0,0 +1,52 @@
+#!/bin/bash
+
+
+case $# in
+    9) SETUP=$1; OUTDIR=$2; NAME=$3; DEVICE=$4; ROUNDS=$5; LPCNEXT=$6; LPCNET=$7; TESTSUITE=$8; TESTITEMS=$9;;
+    *) echo "loop_run.sh setup outdir name device rounds lpcnext_repo lpcnet_repo testsuite_repo testitems"; exit;;
+esac
+
+
+PYTHON="/home/ubuntu/opt/miniconda3/envs/torch/bin/python"
+TESTFEATURES=${LPCNEXT}/testitems/features/all_0_orig_features.f32
+WARPQREFERENCE=${LPCNEXT}/testitems/wav/all_0_orig.wav
+METRICS="warpq,pesq,pitch_error,voicing_error"
+LPCNETDEMO=${LPCNET}/lpcnet_demo
+
+for ((round = 1; round <= $ROUNDS; round++))
+do
+    echo
+    echo round $round
+
+    UUID=$(uuidgen)
+    TRAINOUT=${OUTDIR}/${UUID}/training
+    TESTOUT=${OUTDIR}/${UUID}/testing
+    CHECKPOINT=${TRAINOUT}/checkpoints/checkpoint_last.pth
+    FINALCHECKPOINT=${TRAINOUT}/checkpoints/checkpoint_finalize_last.pth
+
+    # run training
+    echo "starting training..."
+    $PYTHON $LPCNEXT/train_lpcnet.py $SETUP $TRAINOUT --device $DEVICE --test-features $TESTFEATURES --warpq-reference $WARPQREFERENCE
+
+    # run finalization
+    echo "starting finalization..."
+    $PYTHON $LPCNEXT/train_lpcnet.py $SETUP $TRAINOUT \
+            --device $DEVICE --test-features $TESTFEATURES \
+            --warpq-reference $WARPQREFERENCE \
+            --finalize --initial-checkpoint $CHECKPOINT
+
+    # create test configs
+    $PYTHON $LPCNEXT/make_test_config.py ${OUTDIR}/${UUID}/testconfig.yml "$NAME $UUID" $CHECKPOINT --lpcnet-demo $LPCNETDEMO
+    $PYTHON $LPCNEXT/make_test_config.py ${OUTDIR}/${UUID}/testconfig_finalize.yml "$NAME $UUID finalized" $FINALCHECKPOINT --lpcnet-demo $LPCNETDEMO
+
+    # run tests
+    echo "starting test 1 (no finalization)..."
+    $PYTHON $TESTSUITE/run_test.py ${OUTDIR}/${UUID}/testconfig.yml  \
+            $TESTITEMS ${TESTOUT}/prefinal --num-workers 8 \
+            --num-testitems 400 --metrics $METRICS
+
+    echo "starting test 2 (after finalization)..."
+    $PYTHON $TESTSUITE/run_test.py ${OUTDIR}/${UUID}/testconfig_finalize.yml  \
+            $TESTITEMS ${TESTOUT}/final --num-workers 8 \
+            --num-testitems 400 --metrics $METRICS
+done

package/opus/dnn/torch/lpcnet/scripts/make_animation.py

@@ -0,0 +1,67 @@
+"""
+/* Copyright (c) 2023 Amazon
+   Written by Jan Buethe */
+/*
+   Redistribution and use in source and binary forms, with or without
+   modification, are permitted provided that the following conditions
+   are met:
+
+   - Redistributions of source code must retain the above copyright
+   notice, this list of conditions and the following disclaimer.
+
+   - Redistributions in binary form must reproduce the above copyright
+   notice, this list of conditions and the following disclaimer in the
+   documentation and/or other materials provided with the distribution.
+
+   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+   ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
+   OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+   EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+   PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+   PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+   LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+   NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+   SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*/
+"""
+
+
+""" script for creating animations from debug data
+
+"""
+
+
+import argparse
+
+
+import sys
+sys.path.append('./')
+
+from utils.endoscopy import make_animation, read_data
+
+parser = argparse.ArgumentParser()
+
+parser.add_argument('folder', type=str, help='endoscopy folder with debug output')
+parser.add_argument('output', type=str, help='output file (will be auto-extended with .mp4)')
+
+parser.add_argument('--start-index', type=int, help='index of first sample to be considered', default=0)
+parser.add_argument('--stop-index', type=int, help='index of last sample to be considered', default=-1)
+parser.add_argument('--interval', type=int, help='interval between frames in ms', default=20)
+parser.add_argument('--half-window-length', type=int, help='half size of window for displaying signals', default=80)
+
+
+if __name__ == "__main__":
+    args = parser.parse_args()
+
+    filename = args.output if args.output.endswith('.mp4') else args.output + '.mp4'
+    data = read_data(args.folder)
+
+    make_animation(
+        data,
+        filename,
+        start_index=args.start_index,
+        stop_index = args.stop_index,
+        half_signal_window_length=args.half_window_length
+    )

package/opus/dnn/torch/lpcnet/scripts/modify_dataset_target.py

@@ -0,0 +1,17 @@
+import argparse
+
+import numpy as np
+
+
+parser = argparse.ArgumentParser(description="sets s_t to augmented_s_t")
+
+parser.add_argument('datafile', type=str, help='data.s16 file path')
+
+args = parser.parse_args()
+
+data = np.memmap(args.datafile, dtype='int16', mode='readwrite')
+
+# signal is in data[1::2]
+# last augmented signal is in data[0::2]
+
+data[1 : - 1 : 2] = data[2 : : 2]

package/opus/dnn/torch/lpcnet/scripts/multi_run.sh

@@ -0,0 +1,17 @@
+#!/bin/bash
+
+case $# in
+    9) SETUP=$1; OUTDIR=$2; NAME=$3; NUMDEVICES=$4; ROUNDS=$5; LPCNEXT=$6; LPCNET=$7; TESTSUITE=$8; TESTITEMS=$9;;
+    *) echo "multi_run.sh setup outdir name num_devices rounds_per_device lpcnext_repo lpcnet_repo testsuite_repo testitems"; exit;;
+esac
+
+
+LOOPRUN=${LPCNEXT}/loop_run.sh
+
+mkdir -p $OUTDIR
+
+for ((i = 0; i < $NUMDEVICES; i++))
+do
+    echo "launching job queue for device $i"
+    nohup bash $LOOPRUN $SETUP $OUTDIR "$NAME" "cuda:$i" $ROUNDS $LPCNEXT $LPCNET $TESTSUITE $TESTITEMS > $OUTDIR/job_${i}_out.txt &
+done

package/opus/dnn/torch/lpcnet/scripts/run_inference_test.sh

@@ -0,0 +1,22 @@
+#!/bin/bash
+
+
+case $# in
+    3) FEATURES=$1; FOLDER=$2; PYTHON=$3;;
+    *) echo "run_inference_test.sh <features file> <output folder> <python path>"; exit;;
+esac
+
+
+SCRIPTFOLDER=$(dirname "$0")
+
+mkdir -p $FOLDER/inference_test
+
+# update checkpoints
+for fn in $(find $FOLDER -type f -name "checkpoint*.pth")
+do
+    tmp=$(basename $fn)
+    tmp=${tmp%.pth}
+    epoch=${tmp#checkpoint_epoch_}
+    echo "running inference with checkpoint $fn..."
+    $PYTHON $SCRIPTFOLDER/../test_lpcnet.py $FEATURES $fn $FOLDER/inference_test/output_epoch_${epoch}.wav
+done

package/opus/dnn/torch/lpcnet/scripts/update_checkpoints.py

@@ -0,0 +1,54 @@
+"""
+/* Copyright (c) 2023 Amazon
+   Written by Jan Buethe */
+/*
+   Redistribution and use in source and binary forms, with or without
+   modification, are permitted provided that the following conditions
+   are met:
+
+   - Redistributions of source code must retain the above copyright
+   notice, this list of conditions and the following disclaimer.
+
+   - Redistributions in binary form must reproduce the above copyright
+   notice, this list of conditions and the following disclaimer in the
+   documentation and/or other materials provided with the distribution.
+
+   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+   ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
+   OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+   EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+   PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+   PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+   LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+   NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+   SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*/
+"""
+
+""" script for updating checkpoints with new setup entries
+
+    Use this script to update older outputs with newly introduced
+    parameters. (Saves us the trouble of backward compatibility)
+"""
+
+
+import argparse
+
+import torch
+
+parser = argparse.ArgumentParser()
+
+parser.add_argument('checkpoint_file', type=str, help='checkpoint to be updated')
+parser.add_argument('--model', type=str, help='model update', default=None)
+
+args = parser.parse_args()
+
+checkpoint = torch.load(args.checkpoint_file, map_location='cpu')
+
+# update model entry
+if type(args.model) != type(None):
+    checkpoint['setup']['lpcnet']['model'] = args.model
+
+torch.save(checkpoint, args.checkpoint_file)

package/opus/dnn/torch/lpcnet/scripts/update_output_folder.sh

@@ -0,0 +1,22 @@
+#!/bin/bash
+
+
+case $# in
+    3) FOLDER=$1; MODEL=$2; PYTHON=$3;;
+    *) echo "update_output_folder.sh folder model python"; exit;;
+esac
+
+
+SCRIPTFOLDER=$(dirname "$0")
+
+
+# update setup
+echo "updating $FOLDER/setup.py..."
+$PYTHON $SCRIPTFOLDER/update_setups.py $FOLDER/setup.yml --model $MODEL
+
+# update checkpoints
+for fn in $(find $FOLDER -type f -name "checkpoint*.pth")
+do
+    echo "updating $fn..."
+    $PYTHON $SCRIPTFOLDER/update_checkpoints.py $fn --model $MODEL
+done

package/opus/dnn/torch/lpcnet/scripts/update_setups.py

@@ -0,0 +1,57 @@
+"""
+/* Copyright (c) 2023 Amazon
+   Written by Jan Buethe */
+/*
+   Redistribution and use in source and binary forms, with or without
+   modification, are permitted provided that the following conditions
+   are met:
+
+   - Redistributions of source code must retain the above copyright
+   notice, this list of conditions and the following disclaimer.
+
+   - Redistributions in binary form must reproduce the above copyright
+   notice, this list of conditions and the following disclaimer in the
+   documentation and/or other materials provided with the distribution.
+
+   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+   ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
+   OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+   EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+   PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+   PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+   LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+   NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+   SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*/
+"""
+
+""" script for updating setup files with new setup entries
+
+    Use this script to update older outputs with newly introduced
+    parameters. (Saves us the trouble of backward compatibility)
+"""
+
+import argparse
+
+import yaml
+
+parser = argparse.ArgumentParser()
+
+parser.add_argument('setup_file', type=str, help='setup to be updated')
+parser.add_argument('--model', type=str, help='model update', default=None)
+
+args = parser.parse_args()
+
+# load setup
+with open(args.setup_file, 'r') as f:
+    setup = yaml.load(f.read(), yaml.FullLoader)
+
+# update model entry
+if type(args.model) != type(None):
+    setup['lpcnet']['model'] = args.model
+
+# dump result
+with open(args.setup_file, 'w') as f:
+    yaml.dump(setup, f)

package/opus/dnn/torch/lpcnet/test_lpcnet.py

@@ -0,0 +1,89 @@
+"""
+/* Copyright (c) 2023 Amazon
+   Written by Jan Buethe */
+/*
+   Redistribution and use in source and binary forms, with or without
+   modification, are permitted provided that the following conditions
+   are met:
+
+   - Redistributions of source code must retain the above copyright
+   notice, this list of conditions and the following disclaimer.
+
+   - Redistributions in binary form must reproduce the above copyright
+   notice, this list of conditions and the following disclaimer in the
+   documentation and/or other materials provided with the distribution.
+
+   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+   ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
+   OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+   EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+   PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+   PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+   LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+   NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+   SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*/
+"""
+
+import argparse
+
+import torch
+import numpy as np
+
+
+from models import model_dict
+from utils.data import load_features
+from utils.wav import wavwrite16
+
+debug = False
+if debug:
+    args = type('dummy', (object,),
+    {
+        'features'      : 'features.f32',
+        'checkpoint'    : 'checkpoint.pth',
+        'output'        : 'out.wav',
+        'version'       : 2
+    })()
+else:
+    parser = argparse.ArgumentParser()
+
+    parser.add_argument('features', type=str, help='feature file')
+    parser.add_argument('checkpoint', type=str, help='checkpoint file')
+    parser.add_argument('output', type=str, help='output file')
+    parser.add_argument('--version', type=int, help='feature version', default=2)
+
+    args = parser.parse_args()
+
+
+torch.set_num_threads(2)
+
+version = args.version
+feature_file = args.features
+checkpoint_file = args.checkpoint
+
+
+
+output_file = args.output
+if not output_file.endswith('.wav'):
+    output_file += '.wav'
+
+checkpoint = torch.load(checkpoint_file, map_location="cpu")
+
+# check model
+if not 'model' in checkpoint['setup']['lpcnet']:
+    print(f'warning: did not find model entry in setup, using default lpcnet')
+    model_name = 'lpcnet'
+else:
+    model_name = checkpoint['setup']['lpcnet']['model']
+
+model = model_dict[model_name](checkpoint['setup']['lpcnet']['config'])
+
+model.load_state_dict(checkpoint['state_dict'])
+
+data = load_features(feature_file)
+
+output = model.generate(data['features'], data['periods'], data['lpcs'])
+
+wavwrite16(output_file, output.numpy(), 16000)

package/opus/dnn/torch/lpcnet/train_lpcnet.py

@@ -0,0 +1,272 @@
+"""
+/* Copyright (c) 2023 Amazon
+   Written by Jan Buethe */
+/*
+   Redistribution and use in source and binary forms, with or without
+   modification, are permitted provided that the following conditions
+   are met:
+
+   - Redistributions of source code must retain the above copyright
+   notice, this list of conditions and the following disclaimer.
+
+   - Redistributions in binary form must reproduce the above copyright
+   notice, this list of conditions and the following disclaimer in the
+   documentation and/or other materials provided with the distribution.
+
+   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+   ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
+   OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+   EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+   PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+   PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+   LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+   NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+   SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*/
+"""
+
+import os
+import argparse
+import sys
+
+try:
+    import git
+    has_git = True
+except:
+    has_git = False
+
+import yaml
+
+
+import torch
+from torch.optim.lr_scheduler import LambdaLR
+
+from data import LPCNetDataset
+from models import model_dict
+from engine.lpcnet_engine import train_one_epoch, evaluate
+from utils.data import load_features
+from utils.wav import wavwrite16
+
+
+debug = False
+if debug:
+    args = type('dummy', (object,),
+    {
+        'setup' : 'setup.yml',
+        'output' : 'testout',
+        'device' : None,
+        'test_features' : None,
+        'finalize': False,
+        'initial_checkpoint': None,
+        'no-redirect': False
+    })()
+else:
+    parser = argparse.ArgumentParser("train_lpcnet.py")
+    parser.add_argument('setup', type=str, help='setup yaml file')
+    parser.add_argument('output', type=str, help='output path')
+    parser.add_argument('--device', type=str, help='compute device', default=None)
+    parser.add_argument('--test-features', type=str, help='test feature file in v2 format', default=None)
+    parser.add_argument('--finalize', action='store_true', help='run single training round with lr=1e-5')
+    parser.add_argument('--initial-checkpoint', type=str, help='initial checkpoint', default=None)
+    parser.add_argument('--no-redirect', action='store_true', help='disables re-direction of output')
+
+    args = parser.parse_args()
+
+
+torch.set_num_threads(4)
+
+with open(args.setup, 'r') as f:
+    setup = yaml.load(f.read(), yaml.FullLoader)
+
+if args.finalize:
+    if args.initial_checkpoint is None:
+        raise ValueError('finalization requires initial checkpoint')
+
+    if 'sparsification' in setup['lpcnet']['config']:
+        for sp_job in setup['lpcnet']['config']['sparsification'].values():
+            sp_job['start'], sp_job['stop'] = 0, 0
+
+    setup['training']['lr'] = 1.0e-5
+    setup['training']['lr_decay_factor'] = 0.0
+    setup['training']['epochs'] = 1
+
+    checkpoint_prefix = 'checkpoint_finalize'
+    output_prefix = 'output_finalize'
+    setup_name = 'setup_finalize.yml'
+    output_file='out_finalize.txt'
+else:
+    checkpoint_prefix = 'checkpoint'
+    output_prefix = 'output'
+    setup_name = 'setup.yml'
+    output_file='out.txt'
+
+
+# check model
+if not 'model' in setup['lpcnet']:
+    print(f'warning: did not find model entry in setup, using default lpcnet')
+    model_name = 'lpcnet'
+else:
+    model_name = setup['lpcnet']['model']
+
+# prepare output folder
+if os.path.exists(args.output) and not debug and not args.finalize:
+    print("warning: output folder exists")
+
+    reply = input('continue? (y/n): ')
+    while reply not in {'y', 'n'}:
+        reply = input('continue? (y/n): ')
+
+    if reply == 'n':
+        os._exit()
+else:
+    os.makedirs(args.output, exist_ok=True)
+
+checkpoint_dir = os.path.join(args.output, 'checkpoints')
+os.makedirs(checkpoint_dir, exist_ok=True)
+
+
+# add repo info to setup
+if has_git:
+    working_dir = os.path.split(__file__)[0]
+    try:
+        repo = git.Repo(working_dir)
+        setup['repo'] = dict()
+        hash = repo.head.object.hexsha
+        urls = list(repo.remote().urls)
+        is_dirty = repo.is_dirty()
+
+        if is_dirty:
+            print("warning: repo is dirty")
+
+        setup['repo']['hash'] = hash
+        setup['repo']['urls'] = urls
+        setup['repo']['dirty'] = is_dirty
+    except:
+        has_git = False
+
+# dump setup
+with open(os.path.join(args.output, setup_name), 'w') as f:
+    yaml.dump(setup, f)
+
+# prepare inference test if wanted
+run_inference_test = False
+if type(args.test_features) != type(None):
+    test_features = load_features(args.test_features)
+    inference_test_dir = os.path.join(args.output, 'inference_test')
+    os.makedirs(inference_test_dir, exist_ok=True)
+    run_inference_test = True
+
+# training parameters
+batch_size      = setup['training']['batch_size']
+epochs          = setup['training']['epochs']
+lr              = setup['training']['lr']
+lr_decay_factor = setup['training']['lr_decay_factor']
+
+# load training dataset
+lpcnet_config = setup['lpcnet']['config']
+data = LPCNetDataset(   setup['dataset'],
+                        features=lpcnet_config['features'],
+                        input_signals=lpcnet_config['signals'],
+                        target=lpcnet_config['target'],
+                        frames_per_sample=setup['training']['frames_per_sample'],
+                        feature_history=lpcnet_config['feature_history'],
+                        feature_lookahead=lpcnet_config['feature_lookahead'],
+                        lpc_gamma=lpcnet_config.get('lpc_gamma', 1))
+
+# load validation dataset if given
+if 'validation_dataset' in setup:
+    validation_data = LPCNetDataset(   setup['validation_dataset'],
+                        features=lpcnet_config['features'],
+                        input_signals=lpcnet_config['signals'],
+                        target=lpcnet_config['target'],
+                        frames_per_sample=setup['training']['frames_per_sample'],
+                        feature_history=lpcnet_config['feature_history'],
+                        feature_lookahead=lpcnet_config['feature_lookahead'],
+                        lpc_gamma=lpcnet_config.get('lpc_gamma', 1))
+
+    validation_dataloader = torch.utils.data.DataLoader(validation_data, batch_size=batch_size, drop_last=True, num_workers=4)
+
+    run_validation = True
+else:
+    run_validation = False
+
+# create model
+model = model_dict[model_name](setup['lpcnet']['config'])
+
+if args.initial_checkpoint is not None:
+    print(f"loading state dict from {args.initial_checkpoint}...")
+    chkpt = torch.load(args.initial_checkpoint, map_location='cpu')
+    model.load_state_dict(chkpt['state_dict'])
+
+# set compute device
+if type(args.device) == type(None):
+    device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+else:
+    device = torch.device(args.device)
+
+# push model to device
+model.to(device)
+
+# dataloader
+dataloader = torch.utils.data.DataLoader(data, batch_size=batch_size, drop_last=True, shuffle=True, num_workers=4)
+
+# optimizer is introduced to trainable parameters
+parameters = [p for p in model.parameters() if p.requires_grad]
+optimizer = torch.optim.Adam(parameters, lr=lr)
+
+# learning rate scheduler
+scheduler = LambdaLR(optimizer=optimizer, lr_lambda=lambda x : 1 / (1 + lr_decay_factor * x))
+
+# loss
+criterion = torch.nn.NLLLoss()
+
+# model checkpoint
+checkpoint = {
+    'setup'         : setup,
+    'state_dict'    : model.state_dict(),
+    'loss'          : -1
+}
+
+if not args.no_redirect:
+    print(f"re-directing output to {os.path.join(args.output, output_file)}")
+    sys.stdout = open(os.path.join(args.output, output_file), "w")
+
+best_loss = 1e9
+
+for ep in range(1, epochs + 1):
+    print(f"training epoch {ep}...")
+    new_loss = train_one_epoch(model, criterion, optimizer, dataloader, device, scheduler)
+
+
+    # save checkpoint
+    checkpoint['state_dict'] = model.state_dict()
+    checkpoint['loss']       = new_loss
+
+    if run_validation:
+        print("running validation...")
+        validation_loss = evaluate(model, criterion, validation_dataloader, device)
+        checkpoint['validation_loss'] = validation_loss
+
+        if validation_loss < best_loss:
+            torch.save(checkpoint, os.path.join(checkpoint_dir, checkpoint_prefix + f'_best.pth'))
+            best_loss = validation_loss
+
+    torch.save(checkpoint, os.path.join(checkpoint_dir, checkpoint_prefix + f'_epoch_{ep}.pth'))
+    torch.save(checkpoint, os.path.join(checkpoint_dir, checkpoint_prefix + f'_last.pth'))
+
+    # run inference test
+    if run_inference_test:
+        model.to("cpu")
+        print("running inference test...")
+
+        output = model.generate(test_features['features'], test_features['periods'], test_features['lpcs'])
+
+        testfilename = os.path.join(inference_test_dir, output_prefix + f'_epoch_{ep}.wav')
+
+        wavwrite16(testfilename, output.numpy(), 16000)
+
+        model.to(device)
+
+    print()

package/opus/dnn/torch/lpcnet/utils/__init__.py

@@ -0,0 +1,4 @@
+from . import sparsification
+from . import data
+from . import pcm
+from . import sample