Skip to page content
SAVRN
Search Plan a project

Open-weight model

deduplicator

by Ebisuke ebisuke/deduplicator

deduplicator is an open-weight model from Ebisuke. It has 68M parameters and a 4,096-token context. At 16-bit it needs about 0.2 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index.

Parameters68M
Context4,096
Weights794.4 MB
License—
AccessOpen weights
Monthly Downloads—

Runs On

What it takes to serve deduplicator (68M parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.

PrecisionWeightsMemory neededCheapest setupPer hourAlso fits
16-bit 0.1 GB 0.2 GB 1x MI300X (192 GB)
Vultr
$1.85 1x H100 $1.99 · 1x MI325X $2.00
8-bit 0.1 GB 0.1 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 20, 2026.

deduplicator on every accelerator the SAVRN Index prices, at every precision

Model Card

The publisher has not written a card for this model.

Configuration

Architecture
StreamingDeduplicatorModel
Context length (tokens)
4,096
Layers
4
Hidden size
256
Feed-forward size
1,024
Attention heads
4
Key/value heads
2
Head dimension
64
Vocabulary size
2
Model type
streaming_deduplicator

Identity and Version

Repository
ebisuke/deduplicator
Publisher
Ebisuke
Task
Not stated by the source
Modality
Other
Library
Not stated by the source
Parameters
68M parameters
Languages
Not stated by the source
Revision
ed5218d1e0a0e821dc4d850b115d549a2d3eee80
First published
2026-09-29
Last updated
2026-09-29

Files and Weights

5 files, 794.4 MB in total. The weights are 2 files totalling 794.4 MB in pt, safetensors.

Weights2 files · 794.4 MB
Configuration2 files · 876 B
Repository1 file · 1.5 KB
Every file
FileTypeSizeSHA-256
model.safetensorsWeights270.0 MB c4e9f8cfad7f
optimizer.ptWeights524.3 MB 898165be512a
config.jsonConfiguration856 B —
state.jsonConfiguration20 B —
.gitattributesRepository1.5 KB —

License and Download

License
Not stated by the source
Access
Open weights, no gate
Download size
794.4 MB
Download from Ebisuke

Released by Ebisuke through its official repository on Hugging Face.

Memory Requirements

PrecisionWeights in memory
As published794.4 MB
16-bit0.1 GB
8-bit0.1 GB
4-bit0.0 GB

Weights only, from the published parameter count; the key-value cache and runtime add to this.

Questions About deduplicator

How much GPU memory does deduplicator need?

About 0.2 GB at 16-bit and 0 GB at 4-bit: the weights (68M parameters) plus a working margin. A long context needs more.

What is the cheapest GPU to run deduplicator 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.

What is deduplicator's context length?

4,096 tokens, from the maximum position embeddings in its published configuration.