GPT-2 (1.5B, Curriculum Learning 45K)

Closed weights Microsoft 1.5B parameters August 2021

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
Microsoft
Organisation type
Industry
Country
United States of America
Published
13 August 2021
Authors
Conglong Li, Minjia Zhang, Yuxiong He

What it does

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

Domain
Language
Task
Language modeling/generation

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
1.5B
Training data
157,000,000,000 tokens

Table 2 SLW 45K bsz4K-seqlen1K 58.8K 157B (1x) training time: 155Hr (2.2x) batch size 4000 sequence length 1000 58.8K steps 157B tokens 4*10^6*58800/157*10^9 = 1.6 epochs

Epochs
1.6
Batch size
4,000,000

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 × 10²¹ FLOP

"Experiments replicating GPT-2 models (117M and 1.5B) show that <..> our method reduces the required number of training tokens and wall clock time by up to 2.2x and 3.7x, respectively." GPT-2 estimated training compute: 1.9200000000009998e+21 FLOP (Speculative confidence) -> 1.9200000000009998e+21 / 3.7 = 5.1891892e+20 FLOP " All of the experiments are performed on 128 NVIDIA V100 GPUs (32GB memory). There are 16 nodes and 8 GPUs per node" 155 hours (Table 2) 125000000000000 FLOP/sec * 128 …

How it was established
Comparison with other models,Hardware,Operation counting

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 V100
Chips used
128
Wall-clock time
155 hours
Power draw
77.6 kW

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

there's a repo for the technique but I don't see training code for this model: https://github.com/microsoft/DeepSpeed

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
55
Benchmark data
GPT-2 (1.5B, Curriculum Learning 45K)

Sources

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

Reference
Curriculum Learning: A Regularization Method for Efficient and Stable Billion-Scale GPT Model Pre-Training
Last updated
25 May 2026

What the numbers mean

What this model is

GPT-2 (1.5B, Curriculum Learning 45K) was published by Microsoft, in the country recorded as United States of America, during August 2021. The category the publisher falls under is industry.

It works in the domain of Language, and is recorded as performing the task of language modeling/generation.

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 a computation budget of roughly 2.4 × 10²¹ FLOP, on hardware recorded as NVIDIA V100. That figure measures what producing the model cost, and has no bearing on how fast it answers.

The training set ran to roughly 157,000,000,000 tokens of text.

Answers

GPT-2 (1.5B, Curriculum Learning 45K) — common questions

01

GPT-2 (1.5B, Curriculum Learning 45K)— when was it released?

It was published in August 2021. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

02

GPT-2 (1.5B, Curriculum Learning 45K)— what is it used for?

It works in the domain of Language, and is recorded as handling the task of language modeling/generation. 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.

03

GPT-2 (1.5B, Curriculum Learning 45K)— how much compute was used to train it?

Training consumed around 2.4 × 10²¹ FLOP, on hardware recorded as NVIDIA V100. 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.

04

GPT-2 (1.5B, Curriculum Learning 45K)— 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.

05

GPT-2 (1.5B, Curriculum Learning 45K)— is it open source?

No. Its weights have not been published, so it exists only as a service controlled by its owner.

06

GPT-2 (1.5B, Curriculum Learning 45K)— how many parameters does it have?

It has a parameter count of 1.5B. 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.

07

GPT-2 (1.5B, Curriculum Learning 45K)— who created it?

It was published by Microsoft, based in United States of America, an organisation categorised as industry.

Source

Original publication

Record last updated 25 May 2026

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.