VALL-E
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
- 5 January 2023
- Authors
- Chengyi Wang, Sanyuan Chen, Yu Wu, Ziqiang Zhang, Long Zhou, Shujie Liu, Zhuo Chen, Yanqing Liu, Huaming Wang, Jinyu Li, Lei He, Sheng Zhao, Furu Wei
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Audio, Speech
- Task
- Speech synthesis, Text-to-speech (TTS)
- 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
- 353M
- Training data
- 76,800,000,000 tokens
"Both the AR model and the NAR model have the same transformer architecture with 12 layers, 16 attention heads, an embedding dimension of 1024, a feed-forward layer dimension of 4096, and a dropout of 0.1" Ben's script says that's 353M parameters, using n_block 12, d_model 1024, d_ff 4096, encoder only False https://github.com/bencottier/ml-parameter-count/blob/main/parameter_count.py
60k hours ~13,680 words/hour * 60,000 = 820800000 words https://docs.google.com/document/d/1G3vvQkn4x_W71MKg0GmHVtzfd9m0y3_Ofcoew0v902Q/edit#heading=h.3pbt0hfgv7pq
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 × 10¹⁹ FLOP
- How it was established
- Operation counting
"The models are trained using 16 NVIDIA TESLA V100 32GB GPUs with a batch size of 6k acoustic tokens per GPU for 800k steps" 353M * 800k * 6k * 6 = 1.01e19 16 V100s is 2080 teraFLOP or 2e15 FLOP so 1e19 would take 1.5 hours at 100% utilization or ~5 hours at 30%. Is that plausible?
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
- Compute cost
- $11
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
onlly demos https://www.microsoft.com/en-us/research/project/vall-e-x/vall-e/
How it is classified
Labels the source dataset applies when tracking notable models, and how confident it is in the entry.
- Foundation model
- Yes
- Why it is tracked
- SOTA improvement
- Record confidence
- Speculative
- Citations
- 1,180
"VALL-E significantly outperforms the state-of-the-art zero-shot TTS system [Casanova et al., 2022b] in terms of speech naturalness and speaker similarity, with +0.12 comparative mean option score (CMOS) and +0.93 similarity mean option score (SMOS) improvement on LibriSpeech"
Sources
Where this record came from and when it was last checked.
- Reference
- Neural Codec Language Models are Zero-Shot Text to Speech Synthesizers
- Last updated
- 25 May 2026
What the numbers mean
About this model
VALL-E was published by Microsoft, in United States of America, in January 2023. The organisation is categorised as industry.
It works in Audio, Speech, and is recorded as doing speech synthesis, Text-to-speech (TTS).
Its weights were never published, so it can only be reached through its provider. No graphics card changes that.
What went into building it
Training it took roughly 1 × 10¹⁹ FLOP of computation, on NVIDIA V100 — a measure of what producing the model cost, not of how fast it answers.
The training set ran to roughly 76,800,000,000 tokens.
It is tracked in the underlying dataset for one reason in particular: sOTA improvement.
Answers
VALL-E — common questions
What GPU do I need to run VALL-E?
None. VALL-E 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 VALL-E open source?
No. VALL-E has not had its weights published, so it exists only as a service controlled by its owner.
How many parameters does VALL-E have?
VALL-E has 353M parameters. "Both the AR model and the NAR model have the same transformer architecture with 12 layers, 16 attention heads, an embedding dimension of 1024, a feed-forward layer dimension of 4096, and a dropout of 0.1" Ben's script says that's 353M parameters, using n_block 12, d_model 1024, d_ff 4096, encoder only False https://github.com/bencottier/ml-parameter-count/blob/main/parameter_count.py. 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 VALL-E?
VALL-E was published by Microsoft, based in United States of America, categorised as industry.
When was VALL-E released?
VALL-E was published in January 2023. 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 VALL-E used for?
VALL-E works in Audio, Speech, and is recorded as handling speech synthesis, Text-to-speech (TTS). 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 VALL-E?
Around 1 × 10¹⁹ FLOP, on 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.
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.