LSTM(large)+Sememe+cell
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
- Tsinghua University,Beijing University of Posts and Telecommunications,Huawei Noah's Ark Lab
- Organisation type
- Academia,Academia,Industry
- Country
- China
- Published
- 20 October 2019
- Authors
- Yujia Qin, Fanchao Qi, Sicong Ouyang, Zhiyuan Liu, Cheng Yang, Yasheng Wang, Qun Liu, Maosong Sun
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Language
- Task
- Language modeling
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
- 48M
- Training data
- 2,000,000 tokens
- Epochs
- 40
Based on the details in the paper, here is the calculation for the number of parameters in the LSTM(large)+Sememe+cell model: The model has 2 LSTM layers, each with 1500 hidden units. It uses the "+cell" method to incorporate sememes, which introduces an additional sememe-LSTM cell in each LSTM unit. For each original 1500 unit LSTM layer: Input size is 300 (embedding dimension) Hidden size is 1500 4 weight matrices of size 300 x 1500 = 450,000 params each 4 recurrent weight matrices of size…
"The maximum training epoch number is 40 and the gradient norm clip boundary is 0.25."
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 × 10¹⁶ FLOP
- How it was established
- Operation counting
6 FLOP / token / parameter * 48000000 parameters * 2000000 tokens * 40 epochs = 2.304e+16 FLOP
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
- Open (non-commercial)
code for WT/PTB. no license. https://github.com/thunlp/SememeRNN/blob/master/LM/main.py
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
- 19
- Benchmark data
- LSTM(large)+Sememe+cell
Sources
Where this record came from and when it was last checked.
- Reference
- Improving Sequence Modeling Ability of Recurrent Neural Networks via Sememes
- Last updated
- 11 February 2026
What the numbers mean
What this model is
LSTM(large)+Sememe+cell was published by Tsinghua University,Beijing University of Posts and Telecommunications,Huawei Noah's Ark Lab, in China, in October 2019. The organisation is categorised as academia,Academia,Industry.
It works in Language, and is recorded as doing language modeling.
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 2.3 × 10¹⁶ FLOP of computation — a measure of what producing the model cost, not of how fast it answers.
Around 2,000,000 tokens went into training it.
Answers
LSTM(large)+Sememe+cell — common questions
How much compute was used to train LSTM(large)+Sememe+cell?
Around 2.3 × 10¹⁶ 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.
What GPU do I need to run LSTM(large)+Sememe+cell?
None. LSTM(large)+Sememe+cell 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 LSTM(large)+Sememe+cell open source?
No. LSTM(large)+Sememe+cell has not had its weights published, so it exists only as a service controlled by its owner.
How many parameters does LSTM(large)+Sememe+cell have?
LSTM(large)+Sememe+cell has 48M parameters. Based on the details in the paper, here is the calculation for the number of parameters in the LSTM(large)+Sememe+cell model: The model has 2 LSTM layers, each with 1500 hidden units. It uses the "+cell" method to incorporate sememes, which introduces an additional sememe-LSTM cell in each LSTM unit. For each original 1500 unit LSTM layer: Input size is 300 (embedding dimension) Hidden size is 1500 4 weight matrices of size 300 x 1500 = 450,000 params each 4 recurrent weight matrices of size 1500 x 1500 = 2,250,000 params each 4 bias vectors of size 1500 = 6,000 params each Total = 450,000 x 4 + 2,250,000 x 4 + 6,000 x 4 = 12,006,000 For each sememe-LSTM cell: Input size is 2000 (number of sememes) Hidden size is 1500 Same weight and bias computations as above Total = 12,006,000 Since there are 2 LSTM layers each with a sememe-LSTM cell: Total params = 2 * (12,006,000 + 12,006,000) Total = 48,024,000 Therefore, the total number of parameters for the LSTM(large)+Sememe+cell model is approximately 48 million. 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 LSTM(large)+Sememe+cell?
LSTM(large)+Sememe+cell was published by Tsinghua University,Beijing University of Posts and Telecommunications,Huawei Noah's Ark Lab, based in China, categorised as academia,Academia,Industry.
When was LSTM(large)+Sememe+cell released?
LSTM(large)+Sememe+cell was published in October 2019. 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 LSTM(large)+Sememe+cell used for?
LSTM(large)+Sememe+cell works in Language, and is recorded as handling language modeling. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.
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.