MTDP
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
- Chinese University of Hong Kong (CUHK),City University of Hong Kong
- Organisation type
- Academia,Academia
- Country
- Hong Kong
- Published
- 24 September 2024
- Authors
- Jiayu Shang, Cheng Peng, Yongxin Ji, Jiaojiao Guan, Dehan Cai, Xubo Tang, Yanni Sun
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Biology
- Task
- Protein embedding
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
- 20M
- Training data
- tokens
- Epochs
- 10
- Batch size
- 16
"In particular, the student model in MTDP has a significantly smaller parameter scale (~20 million)" from 2.1 Model Structure
"MTDP is pre-trained on ~500 000 proteins from UniProtKB (Swiss-Prot) provided by the teacher model (Elnaggar et al. 2021)." from 2.3.1 Pre-training data
"We conducted the pre-training process on four RTX 3080 24 GB Nvidia GPUs, with a batch size of 16." from 2.3.1 Pre-training data
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
- 6 × 10¹⁴ FLOP
- How it was established
- Operation counting
6 * 20,000,000 connections * 500,000 training examples * 10 epochs = 6.0e14 FLOP
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 GeForce RTX 3080
- Chips used
- 4
- Power draw
- 2.5 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
Academic Free License v3.0 License. Inference and fine-tuning code + datasets, I don't see model weights and training code https://github.com/KennthShang/MTDP
How it is classified
Labels the source dataset applies when tracking notable models, and how confident it is in the entry.
- Record confidence
- Confident
Sources
Where this record came from and when it was last checked.
- Reference
- Accurate and efficient protein embedding using multi-teacher distillation learning
- Last updated
- 28 November 2025
What the numbers mean
Where it came from
MTDP was published by Chinese University of Hong Kong (CUHK),City University of Hong Kong, in the country recorded as Hong Kong, during September 2024. It comes out of an organisation categorised as academia,Academia.
It works in the domain of Biology, and is recorded as performing the task of protein embedding.
Because the weights are not available, none of the hardware figures elsewhere on this site apply to it.
How it was trained
The training run consumed about 6 × 10¹⁴ FLOP, on hardware recorded as NVIDIA GeForce RTX 3080. That figure measures what producing the model cost, and has no bearing on how fast it answers.
Answers
MTDP — common questions
MTDP— what is it used for?
It works in the domain of Biology, and is recorded as handling the task of protein embedding. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.
MTDP— how much compute was used to train it?
Training consumed around 6 × 10¹⁴ FLOP, on hardware recorded as NVIDIA GeForce RTX 3080. 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.
MTDP— 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.
MTDP— is it open source?
No. Its weights have not been published, so it exists only as a service controlled by its owner.
MTDP— how many parameters does it have?
It has a parameter count of 20M. "In particular, the student model in MTDP has a significantly smaller parameter scale (~20 million)" from 2.1 Model Structure. 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.
MTDP— who created it?
It was published by Chinese University of Hong Kong (CUHK),City University of Hong Kong, based in Hong Kong, an organisation categorised as academia,Academia.
MTDP— when was it released?
It was published in September 2024. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.
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.