SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1910.11480 · 2019

Parallel WaveGAN: A fast waveform generation model based on generative adversarial networks with multi-resolution spectrogram

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 20 functions out of this paper's own repositories and ran 17 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
bigpon/QPPWG pwc_unofficial 8 of 9
bigpon/vcc20_baseline_cyclevae pwc_unofficial 7 of 8
mukeshv0/ParallelWaveGAN pwc_unofficial 2 of 2
copy not recorded — 0 of 1
FunctionStatusWhere it lives
batch_f0 Ran bigpon/QPPWG/qppwg/utils/features.py
code served (permissive licence) · get_code("0ed54be0f7619aca")
check_hdf5 Ran bigpon/vcc20_baseline_cyclevae/baseline/src/utils/utils.py
code served (permissive licence) · get_code("92d84700541a2536")
dilated_factor Ran bigpon/QPPWG/qppwg/utils/features.py
code served (permissive licence) · get_code("8cd88186efdab6e0")
find_files Ran mukeshv0/ParallelWaveGAN/parallel_wavegan/utils/utils.py
code served (permissive licence) · get_code("562a6efae0225990")
loss_vae_laplace Ran bigpon/vcc20_baseline_cyclevae/baseline/src/nets/gru_vae.py
code served (permissive licence) · get_code("04dd0d78b380bade")
low_cut_filter Ran bigpon/QPPWG/qppwg/bin/preprocess.py
code served (permissive licence) · get_code("bbf58066f7748c5a")
low_cut_filter Ran bigpon/QPPWG/qppwg/utils/filters.py
code served (permissive licence) · get_code("44d3c0be62e4bbab")
low_cut_filter Ran bigpon/vcc20_baseline_cyclevae/baseline/src/bin/feature_extract.py
code served (permissive licence) · get_code("354a3798e3217759")
low_pass_filter Ran bigpon/QPPWG/qppwg/utils/filters.py
code served (permissive licence) · get_code("c28f03204eefdd29")
padding Ran bigpon/vcc20_baseline_cyclevae/baseline/src/utils/dataset.py
code served (permissive licence) · get_code("9164e9f0890be261")
path_replace Ran bigpon/QPPWG/qppwg/bin/preprocess.py
code served (permissive licence) · get_code("a5d255facaefc424")
read_hdf5 Ran bigpon/vcc20_baseline_cyclevae/baseline/src/utils/utils.py
code served (permissive licence) · get_code("c0be466d384f4bf1")
read_hdf5 Ran mukeshv0/ParallelWaveGAN/parallel_wavegan/utils/utils.py
code served (permissive licence) · get_code("4f86bc80523a92c1")
sampling_vae_laplace Ran bigpon/vcc20_baseline_cyclevae/baseline/src/nets/gru_vae.py
code served (permissive licence) · get_code("b45003226b3b8d82")
shape_hdf5 Ran bigpon/vcc20_baseline_cyclevae/baseline/src/utils/utils.py
code served (permissive licence) · get_code("d3302da55afa1057")
spk_division Ran bigpon/QPPWG/qppwg/bin/preprocess.py
code served (permissive licence) · get_code("102ba7c26e81660b")
validate_length Ran bigpon/QPPWG/qppwg/utils/features.py
code served (permissive licence) · get_code("a4fce818fed4b3c9")
load_checkpoint Not yet run this paper's copy was not recorded; identical code first harvested from yanggeng1995/GAN-TTS
pointer only · get_code("f443ef0155d2d898")
sampling_vae_laplace_batch Not yet run bigpon/vcc20_baseline_cyclevae/baseline/src/nets/gru_vae.py
code served (permissive licence) · get_code("6e0be223aee6ee4e")
stft Not yet run bigpon/QPPWG/qppwg/losses/stft_loss.py
code served (permissive licence) · get_code("d014b1a7520bc12b")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

We propose Parallel WaveGAN, a distillation-free, fast, and small-footprint waveform generation method using a generative adversarial network. In the proposed method, a non-autoregressive WaveNet is trained by jointly optimizing multi-resolution spectrogram and adversarial loss functions, which can effectively capture the time-frequency distribution of the realistic speech waveform. As our method does not require density distillation used in the conventional teacher-student framework, the entire model can be easily trained. Furthermore, our model is able to generate high-fidelity speech even with its compact architecture. In particular, the proposed Parallel WaveGAN has only 1.44 M parameters and can generate 24 kHz speech waveform 28.68 times faster than real-time on a single GPU environment. Perceptual listening test results verify that our proposed method achieves 4.16 mean opinion score within a Transformer-based text-to-speech framework, which is comparative to the best distillation-based Parallel WaveNet system.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("1910.11480")
get_code_for_paper("1910.11480")
have("1910.11480")

Connect an agent — have() is free.