smolvla-aegis-ft-step10607-full is an open-weight model from H. It has 450M parameters. At 16-bit it needs about 1.1 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index.
Self-contained FULL weights (merged LoRA r16/a32 onto step917, verified to bf16 precision). No peft/adapter path needed. For LIBERO eval: --policy.path with rename camera1/camera2.
Runs On
What it takes to serve smolvla-aegis-ft-step10607-full (450M 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 | 0.9 GB | 1.1 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 0.5 GB | 0.5 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 0.2 GB | 0.3 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 18, 2026.
smolvla-aegis-ft-step10607-full on every accelerator the SAVRN Index prices, at every precision
Model Card
Self-contained FULL weights (merged LoRA r16/a32 onto step917, verified to bf16 precision). No peft/adapter path needed. For LIBERO eval: --policy.path with rename camera1/camera2.
Excerpt from the card by H.
Identity and Version
- Repository
- krishnah27/smolvla-aegis-ft-step10607-full
- Publisher
- H
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- 450M parameters
- Languages
- Not stated by the source
- Revision
- a129dd0ce39d9ed872db711449e09429f6235793
- First published
- 2026-10-08
- Last updated
- 2026-10-09
Files and Weights
6 files, 906.7 MB in total. The weights are 1 file totalling 906.7 MB in safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model.safetensors | Weights | 906.7 MB | 17880524daf9 |
| config.json | Configuration | 2.1 KB | — |
| policy_postprocessor.json | Configuration | 574 B | — |
| policy_preprocessor.json | Configuration | 1.8 KB | — |
| README.md | Documentation | 217 B | — |
| .gitattributes | Repository | 1.5 KB | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 906.7 MB
Released by H through its official repository on Hugging Face.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 906.7 MB |
| 16-bit | 0.9 GB |
| 8-bit | 0.5 GB |
| 4-bit | 0.2 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About smolvla-aegis-ft-step10607-full
How much GPU memory does smolvla-aegis-ft-step10607-full need?
About 1.1 GB at 16-bit and 0.3 GB at 4-bit: the weights (450M parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run smolvla-aegis-ft-step10607-full 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.