Fine-tuned-AWD-LSTM-DOC (fin)
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
- Samsung R&D Institute Russia
- Organisation type
- Industry
- Country
- Russia
- Published
- 12 November 2018
- Authors
- Vadim Popov, Mikhail Kudinov
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Language
- Task
- Language modeling
- Base model
- AWD-LSTM-DOC (fin) (23M)
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
- 46M
- Training data
- 1,044,112 tokens
- Epochs
- 15
This is the model trained on Penn Treebank, which uses as a base model the 23M model from Table 7 in https://aclanthology.org/D18-1489.pdf They additionally train a discriminator with the same architecture, so total parameters is 2*23M = 46M
Per https://arxiv.org/pdf/1904.04733: "The most common split of this corpus, where sections from 0 to 18 are used for training (38 219 sentences, 912 344 tokens), sections from 19 to 21 are used for validation (5 527 sentences, 131 768 tokens), and sections from 22 to 24 are used for testing (5 462 sentences, 129 654 tokens)." So dev set is 912,344 + 131768 = 1,044,112
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
- 5.2 × 10¹⁶ FLOP
- How it was established
- Operation counting
Base model uses 4.323e16 FLOPs. They then train a discriminator using the same architecture for 30 epochs, and then use the discriminator to fine-tune the base model for another 15 epochs. Both of these latter training steps require running forward passes on both the discriminator and the language model, but only doing a backward pass on one of them. Discriminator training: 2*23M*30*1044112 + 6*23M*30*1044112 = 5.763e15 LM fine-tuning: 2*23M*15*1044112 + 6*23M*15*1044112 = 2.882e15 Total: 4.…
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
- Citations
- 2
- Benchmark data
- Fine-tuned-AWD-LSTM-DOC(fin)
"The novel approach that we propose allows us to reach state-of-theart quality on Penn Treebank: perplexity decreases from 52.4 to 52.1."
Sources
Where this record came from and when it was last checked.
- Reference
- Fine-tuning of Language Models with Discriminator
- Last updated
- 11 February 2026
What the numbers mean
What this model is
Fine-tuned-AWD-LSTM-DOC (fin) was published by Samsung R&D Institute Russia, in Russia, in November 2018. The organisation is categorised as industry.
It works in Language, and is recorded as doing language modeling.
It builds on AWD-LSTM-DOC (fin) (23M), which is why it shares that model's general shape and size.
Its weights were never published, so it can only be reached through its provider. No graphics card changes that.
How it was trained
Producing it required around 5.2 × 10¹⁶ FLOP of arithmetic, which is a statement about the training budget rather than about inference.
It was trained on about 1,044,112 tokens of text.
It is tracked in the underlying dataset for one reason in particular: sOTA improvement.
Answers
Fine-tuned-AWD-LSTM-DOC (fin) — common questions
When was Fine-tuned-AWD-LSTM-DOC (fin) released?
Fine-tuned-AWD-LSTM-DOC (fin) was published in November 2018. 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 Fine-tuned-AWD-LSTM-DOC (fin) used for?
Fine-tuned-AWD-LSTM-DOC (fin) works in Language, and is recorded as handling language modeling. A model can carry several of each, so these are the areas it was built for rather than a limit on what it will attempt.
How much compute was used to train Fine-tuned-AWD-LSTM-DOC (fin)?
Around 5.2 × 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 Fine-tuned-AWD-LSTM-DOC (fin)?
None. Fine-tuned-AWD-LSTM-DOC (fin) 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 Fine-tuned-AWD-LSTM-DOC (fin) open source?
No. Fine-tuned-AWD-LSTM-DOC (fin) has not had its weights published, so it exists only as a service controlled by its owner.
How many parameters does Fine-tuned-AWD-LSTM-DOC (fin) have?
Fine-tuned-AWD-LSTM-DOC (fin) has 46M parameters. This is the model trained on Penn Treebank, which uses as a base model the 23M model from Table 7 in https://aclanthology.org/D18-1489.pdf They additionally train a discriminator with the same architecture, so total parameters is 2*23M = 46M. 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 Fine-tuned-AWD-LSTM-DOC (fin)?
Fine-tuned-AWD-LSTM-DOC (fin) was published by Samsung R&D Institute Russia, based in Russia, categorised as industry.
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.