mivolo_v2 is an open-weight model from Irina Tolstykh, released under Apache License 2.0. It has 29M parameters. At 16-bit it needs about 0.1 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index. It draws 1.4M downloads a month.
We introduce state-of-the-art multi-input transformer for age and gender estimation. This model was trained on proprietary and open-source datasets.
Runs On
What it takes to serve mivolo_v2 (29M parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.
| Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits |
|---|---|---|---|---|---|
| 16-bit | 0.1 GB | 0.1 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 0.0 GB | 0.0 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 0.0 GB | 0.0 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Oct 9, 2026.
mivolo_v2 on every accelerator the SAVRN Index prices, at every precision
SAVRN's Notes on mivolo_v2
Point it at a photo and it returns an age estimate and a gender label — the entire job for this 29M-parameter multi-input transformer. Weights are 115.1 MB stored in float16; the 16-bit row reads 0.1 GB and the 8-bit and 4-bit rows round to zero, so the cheapest listed setup, a 192 GB MI300X at $1.85 an hour, is theoretical. You batch images against a shared card or run it on CPU-adjacent hardware; 1.8M monthly downloads say pipelines do exactly that.
Apache 2.0 permits commercial use, modification and redistribution, with notices kept, changes stated and an express patent grant. The pinned stack is the mivolo library — the repo ships its own modeling and preprocessing code — so version-lock your serving environment; the repo was last updated 2026-03-11. And do your own bias evaluation before shipping age or gender inference in a product: the card describes the architecture, not its error modes, and does not name its training data.
Model Card
By Irina Tolstykh, published under apache-2.0, revision 53393526c220.
Model Card for MiVOLO V2 model
Configuration
- Architecture
- MiVOLOForImageClassification
- Stored precision
- float16
- Model type
- mivolo
Identity and Version
- Repository
- iitolstykh/mivolo_v2
- Publisher
- Irina Tolstykh
- Task
- Not stated by the source
- Modality
- Other
- Library
- mivolo
- Parameters
- 29M parameters
- Languages
- Not stated by the source
- Revision
- 53393526c220e34cdd7b722b36d22b6f9e5f4241
- First published
- 2024-03-05
- Last updated
- 2026-03-11
Files and Weights
10 files, 115.1 MB in total. The weights are 1 file totalling 115.1 MB in safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model.safetensors | Weights | 115.1 MB | 96efb47051c0 |
| config.json | Configuration | 924 B | — |
| configuration_mivolo.py | Configuration | 1.5 KB | — |
| mivolo_image_processor.py | Configuration | 1.5 KB | — |
| modeling_mivolo.py | Configuration | 5.3 KB | — |
| preprocessor_config.json | Configuration | 96 B | — |
| README.md | Documentation | 3.8 KB | — |
| icon.jpg | Other | 35.3 KB | — |
| requirements.txt | Other | 150 B | — |
| .gitattributes | Repository | 1.6 KB | — |
License and Download
- License
- apache-2.0
- Access
- Open weights, no gate
- Download size
- 115.1 MB
Released by Irina Tolstykh through its official repository on Hugging Face. Read the license.
Built From
- Described by arXiv:2307.04616
- Described by arXiv:2403.02302
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 115.1 MB |
| 16-bit | 0.1 GB |
| 8-bit | 0.0 GB |
| 4-bit | 0.0 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About mivolo_v2
How much GPU memory does mivolo_v2 need?
About 0.1 GB at 16-bit and 0 GB at 4-bit: the weights (29M parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run mivolo_v2 on?
At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
Can I use mivolo_v2 commercially?
Yes. mivolo_v2 is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.