BiLSTM for Speech
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
- IDSIA,Technical University of Munich
- Organisation type
- Academia,Academia
- Country
- Switzerland, Germany
- Published
- 1 August 2005
- Authors
- A Graves, J Schmidhuber
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Speech
- Task
- Speech recognition (ASR)
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
- 152.1K
- Training data
- tokens
"The hidden layer sizes were chosen to ensure that all networks had roughly the same number of weights W (≈100,000). However, for the MLPs the network grew with the time-window size, and W varied between 22,061 and 152,061."
https://catalog.ldc.upenn.edu/LDC93s1 One sample utterance has around 10 words 3696 utterances * 10 words = around 37k 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
- 2.4 × 10¹³ FLOP
- How it was established
- Third-party estimation
Extracted from AI and Compute (https://openai.com/blog/ai-and-compute/) charts by using https://automeris.io/WebPlotDigitizer/. Forward FLOP: 2*200000=400000 Trained on 4158 utterances (TIMIT) "We found that large networks, of around 200,000 weights, gave good performance" TIMIT has around 5 hours total. Estimated utterance length: 5*60*60/(4620+1680)=2.86s The frame size was 5 ms Frames per utterance: 2.86/0.005=572 Training FLOP for one epoch: 3*400000*572*4158=2854051200000 Epochs unclear, …
How it is classified
Labels the source dataset applies when tracking notable models, and how confident it is in the entry.
- Record confidence
- Likely
- Citations
- 4,706
Sources
Where this record came from and when it was last checked.
- Reference
- Framewise phoneme classification with bidirectional LSTM and other neural network architectures
- Last updated
- 1 January 2026
What the numbers mean
Background
BiLSTM for Speech was published by IDSIA,Technical University of Munich, in the country recorded as Switzerland, during August 2005. It comes out of an organisation categorised as academia,Academia.
It works in the domain of Speech, and is recorded as performing the task of speech recognition (ASR).
This is a closed model: the trained values stayed with whoever produced them, and there is no local version to run.
Training and provenance
The training run consumed about 2.4 × 10¹³ FLOP. That figure measures what producing the model cost, and has no bearing on how fast it answers.
Answers
BiLSTM for Speech — common questions
BiLSTM for Speech— how much compute was used to train it?
Training consumed around 2.4 × 10¹³ 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.
BiLSTM for Speech— what GPU do I need to run it?
None. This 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.
BiLSTM for Speech— is it open source?
The licensing was never recorded in our source data. We treat unstated licensing as closed, because an unrecorded licence is not one to rely on.
BiLSTM for Speech— how many parameters does it have?
It has a parameter count of 152.1K. "The hidden layer sizes were chosen to ensure that all networks had roughly the same number of weights W (≈100,000). However, for the MLPs the network grew with the time-window size, and W varied between 22,061 and 152,061.". 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.
BiLSTM for Speech— who created it?
It was published by IDSIA,Technical University of Munich, based in Switzerland, an organisation categorised as academia,Academia.
BiLSTM for Speech— when was it released?
It was published in August 2005. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.
BiLSTM for Speech— what is it used for?
It works in the domain of Speech, and is recorded as handling the task of speech recognition (ASR). These are the areas it was designed around; they describe intent rather than a hard boundary.
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.