TC-DNN-BLSTM-DNN
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
- Carnegie Mellon University (CMU)
- Organisation type
- Academia
- Country
- United States of America
- Published
- 6 April 2015
- Authors
- William Chan, Ian Lane
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Speech
- Task
- Speech recognition (ASR), Speech-to-text
- Approach
- Supervised
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
- 18.4M
- Training data
- 29,160,000 tokens
- Epochs
- 17
3*40 inputs - 2L DNN (2048) - BLSTM (128*2) - 2L DNN (2048) - 3431 outputs 3*40*2048 + 2048*2048 + 2*4*(2048+128)*128 + 256*2048 + 2048*2048 + 2048*3431=18413568
"We use si284 with approximately 81 hours of speech as the training set,"
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
- 1.9 × 10¹⁷ FLOP
- How it was established
- Hardware
"51 hours in wall clock time with a NVIDIA Tesla K20 GPU" K20 FLOPs: 3524000000000 Compute: 0.3*51*60*60*3524000000000=194101919999999970= 1.94e17
The training run
What it physically took to train: which chips, how many, for how long, and what that drew from the wall.
- Training hardware
- NVIDIA Tesla K20m
- Chips used
- 1
- Wall-clock time
- 51 hours
- Power draw
- 263 W
Availability
Whether you can obtain the model and run it on your own hardware, which is what decides if any of the graphics-card figures on this page apply.
- Weights
- Closed — provider access only
- Model access
- Unreleased
- Training code
- Unreleased
How it is classified
Labels the source dataset applies when tracking notable models, and how confident it is in the entry.
- Why it is tracked
- SOTA improvement
- Record confidence
- Confident
https://paperswithcode.com/sota/speech-recognition-on-wsj-eval92
Sources
Where this record came from and when it was last checked.
- Reference
- Deep Recurrent Neural Networks for Acoustic Modelling
- Last updated
- 28 November 2025
What the numbers mean
About this model
TC-DNN-BLSTM-DNN was published by Carnegie Mellon University (CMU), in United States of America, in April 2015. It comes out of academia.
It works in Speech, and is recorded as doing speech recognition (ASR), Speech-to-text.
Because the weights are not available, none of the hardware figures elsewhere on this site apply to it.
How it was trained
Training it took roughly 1.9 × 10¹⁷ FLOP of computation, on NVIDIA Tesla K20m — a measure of what producing the model cost, not of how fast it answers.
It was trained on about 29,160,000 tokens of text.
The reason it appears in this catalogue at all is sOTA improvement.
Answers
TC-DNN-BLSTM-DNN — common questions
When was TC-DNN-BLSTM-DNN released?
TC-DNN-BLSTM-DNN was published in April 2015. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.
What is TC-DNN-BLSTM-DNN used for?
TC-DNN-BLSTM-DNN works in Speech, and is recorded as handling speech recognition (ASR), Speech-to-text. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.
How much compute was used to train TC-DNN-BLSTM-DNN?
Around 1.9 × 10¹⁷ FLOP, on NVIDIA Tesla K20m. 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.
What GPU do I need to run TC-DNN-BLSTM-DNN?
None. TC-DNN-BLSTM-DNN 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.
Is TC-DNN-BLSTM-DNN open source?
No. TC-DNN-BLSTM-DNN has not had its weights published, so it exists only as a service controlled by its owner.
How many parameters does TC-DNN-BLSTM-DNN have?
TC-DNN-BLSTM-DNN has 18.4M parameters. 3*40 inputs - 2L DNN (2048) - BLSTM (128*2) - 2L DNN (2048) - 3431 outputs 3*40*2048 + 2048*2048 + 2*4*(2048+128)*128 + 256*2048 + 2048*2048 + 2048*3431=18413568. 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.
Who created TC-DNN-BLSTM-DNN?
TC-DNN-BLSTM-DNN was published by Carnegie Mellon University (CMU), based in United States of America, categorised as academia.
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.