SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1703.10135 · 2017

Tacotron: Towards End-to-End Speech Synthesis

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 25 functions out of this paper's own repositories and ran 8 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
copy not recorded — 3 of 3
barronalex/Tacotron reimplementation 2 of 2
andabi/deep-voice-conversion pwc_unofficial 1 of 10
thuhcsi/tacotron reimplementation 1 of 1
fatchord/WaveRNN reimplementation 1 of 1
keithito/tacotron pwc_unofficial 0 of 4
Kyubyong/tacotron pwc_unofficial 0 of 3
mozilla/TTS reimplementation 0 of 1
FunctionStatusWhere it lives
is_end_of_frames Ran thuhcsi/tacotron/model/tacotron.py
code served (permissive licence) · get_code("e882f1d97538b5c1")
load_checkpoint Ran this paper's copy was not recorded; identical code first harvested from anandaswarup/rnn-tts
pointer only · get_code("669fcedc80c2fee3")
load_trained_model Ran this paper's copy was not recorded; identical code first harvested from anandaswarup/rnn-tts
pointer only · get_code("91be03213654f639")
merge_dict Ran andabi/deep-voice-conversion/hparam.py
code served (permissive licence) · get_code("0cfe2baee7e49f75")
np_now Ran this paper's copy was not recorded; identical code first harvested from dipjyoti92/SC-WaveRNN
pointer only · get_code("2af6a131384db5c8")
np_now Ran fatchord/WaveRNN/train_tacotron.py
code served (permissive licence) · get_code("4e3cdb2262f23b36")
pad_to_dense Ran barronalex/Tacotron/preprocess.py
pointer only (licence: NONE) · get_code("f3365f387e58dbeb")
process_char Ran barronalex/Tacotron/preprocess.py
pointer only (licence: NONE) · get_code("57968eca12c787ab")
bn Not yet run Kyubyong/tacotron/modules.py
code served (permissive licence) · get_code("21f623855d3d8bb9")
conv1d Not yet run Kyubyong/tacotron/modules.py
code served (permissive licence) · get_code("6f9a221306ef60fb")
conv1d Not yet run andabi/deep-voice-conversion/modules.py
code served (permissive licence) · get_code("c5f26d8500f4124f")
denormalize_0_1 Not yet run andabi/deep-voice-conversion/utils.py
code served (permissive licence) · get_code("47507862811320b1")
embed Not yet run Kyubyong/tacotron/modules.py
code served (permissive licence) · get_code("648eb7c9f95d169e")
expand_abbreviations Not yet run keithito/tacotron/text/cleaners.py
code served (permissive licence) · get_code("cea1f83e9db28ed8")
expand_numbers Not yet run keithito/tacotron/text/cleaners.py
code served (permissive licence) · get_code("4c6716f6a3894fe8")
format_data Not yet run mozilla/TTS/TTS/bin/train_tacotron.py
pointer only (licence: MPL-2.0) · get_code("9df2870ead7d247b")
load_hparam Not yet run andabi/deep-voice-conversion/hparam.py
code served (permissive licence) · get_code("946edc002ef7c1b6")
lowercase Not yet run keithito/tacotron/text/cleaners.py
code served (permissive licence) · get_code("296274b940cb0b32")
normalize Not yet run andabi/deep-voice-conversion/modules.py
code served (permissive licence) · get_code("1f684f9c02a32fcf")
normalize_0_1 Not yet run andabi/deep-voice-conversion/utils.py
code served (permissive licence) · get_code("e9833fca17340113")
prenet Not yet run keithito/tacotron/models/modules.py
code served (permissive licence) · get_code("8bb2d1c4dd2ef71e")
read_mfcc Not yet run andabi/deep-voice-conversion/audio.py
code served (permissive licence) · get_code("ba9b99f87a010a58")
read_spectrogram Not yet run andabi/deep-voice-conversion/audio.py
code served (permissive licence) · get_code("ff53076f6122f01a")
split_path Not yet run andabi/deep-voice-conversion/utils.py
code served (permissive licence) · get_code("ed55468b3ae5d0b0")
wav_random_crop Not yet run andabi/deep-voice-conversion/data_load.py
code served (permissive licence) · get_code("74e12237b60dba09")

Repositories linked to this paper

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

Abstract

A text-to-speech synthesis system typically consists of multiple stages, such as a text analysis frontend, an acoustic model and an audio synthesis module. Building these components often requires extensive domain expertise and may contain brittle design choices. In this paper, we present Tacotron, an end-to-end generative text-to-speech model that synthesizes speech directly from characters. Given <text, audio> pairs, the model can be trained completely from scratch with random initialization. We present several key techniques to make the sequence-to-sequence framework perform well for this challenging task. Tacotron achieves a 3.82 subjective 5-scale mean opinion score on US English, outperforming a production parametric system in terms of naturalness. In addition, since Tacotron generates speech at the frame level, it's substantially faster than sample-level autoregressive methods.

For agents

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

get_harvested_code_for_paper("1703.10135")
get_code_for_paper("1703.10135")
have("1703.10135")

Connect an agent — have() is free.