RNN for speech

Closed weights National Chiao Tung University 7.5K parameters May 1998

No estimate

No hardware requirements for this model

The weights for this model have not been published, so it cannot be downloaded or run on your own hardware at any size. It is reachable only through its provider, and no graphics card changes that.

On record

Full specification

Everything on record for this model. Most of it describes how it was trained rather than how it runs — useful context for judging how much work went into it, and how it compares with models built at a different scale.

Origin

Who built this model, where, and when it was published.

Organisation
National Chiao Tung University
Organisation type
Academia
Country
Taiwan
Published
15 May 1998
Authors
SH Chen, SH Hwang, YR Wang

What it does

The problem areas the model was built for. A model can carry several of each.

Domain
Speech
Task
Speech synthesis

Size

How large the model is and how much data it was trained on. Parameters are the figure that decides whether it fits on a given graphics card.

Parameters
7.5K

"The RNN generated a total of eigt output prosodic parameters. [...] The numbers of nodes in the first and second hidden layers were determined empirically and set to be 35 and 30, respectively" Figure 1 contains an overview of the architecture. Layer 1: (102 + 35 + 1)*35 parameters Layer 2: (43 + 35 + 1)*30 parameters Output layer: (30+8+1)*8 parameters

Training data
tokens

The data base was divided into two parts: a training set and an open test set. These two sets consisted of 28 191 and 7051 syllables, respectively. Of the top 10,000 Chinese words, 15% have 1 syllable, 78% have 2 syllables, and 7% have more than two syllables. Assuming 2 syllables per word, the training set is around 14100 words.

Training compute

The arithmetic performed to train the model, measured in floating-point operations. It is a measure of what the training run cost, not of how fast the finished model answers you.

Training compute
2.3 × 10¹¹ FLOP

Extracted from AI and Compute (https://openai.com/blog/ai-and-compute/) charts by using https://automeris.io/WebPlotDigitizer/.

How it was established
Third-party estimation

How it is classified

Labels the source dataset applies when tracking notable models, and how confident it is in the entry.

Citations
231

Sources

Where this record came from and when it was last checked.

Reference
An RNN-based prosodic information synthesizer for Mandarin text-to-speech
Last updated
28 November 2025

What the numbers mean

Background

RNN for speech was published by National Chiao Tung University, in Taiwan, in May 1998. academia is the category the publisher falls under.

It works in Speech, and is recorded as doing speech synthesis.

Because the weights are not available, none of the hardware figures elsewhere on this site apply to it.

What went into building it

Training it took roughly 2.3 × 10¹¹ FLOP of computation — a measure of what producing the model cost, not of how fast it answers.

Answers

RNN for speech — common questions

01

How much compute was used to train RNN for speech?

Around 2.3 × 10¹¹ FLOP. That measures what producing the model cost and says nothing about how quickly it answers once trained — inference speed comes from memory bandwidth, not from the training budget.

02

What GPU do I need to run RNN for speech?

None. RNN for speech is a closed model — its weights were never published, so it cannot be downloaded or run on your own hardware at any price. It is reachable only through its provider.

03

Is RNN for speech open source?

The licensing for RNN for speech was never recorded in our source data. We treat unstated licensing as closed, because an unrecorded licence is not one to rely on.

04

How many parameters does RNN for speech have?

RNN for speech has 7.5K parameters. "The RNN generated a total of eigt output prosodic parameters. [...] The numbers of nodes in the first and second hidden layers were determined empirically and set to be 35 and 30, respectively" Figure 1 contains an overview of the architecture. Layer 1: (102 + 35 + 1)*35 parameters Layer 2: (43 + 35 + 1)*30 parameters Output layer: (30+8+1)*8 parameters. That figure is the total, and it is what decides how much memory the model needs — roughly half a gigabyte per billion at the compression most people use.

05

Who created RNN for speech?

RNN for speech was published by National Chiao Tung University, based in Taiwan, categorised as academia.

06

When was RNN for speech released?

RNN for speech was published in May 1998. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

07

What is RNN for speech used for?

RNN for speech works in Speech, and is recorded as handling speech synthesis. These are the areas it was designed around; they describe intent rather than a hard boundary.

Source

Original publication

Record last updated 28 November 2025

The other direction

Looking at it from the other side?

This page starts from the model. If you already own a card and want to know everything it will run, start from the hardware instead.