SAVRN
Search Contact SAVRN

Open-weight model

multitask-scratch

by Priscila pecarvalho/multitask-scratch

A research-oriented Albef prototype targeting Multitask. The included small setup documents defaults and file formats without presenting unverified performance numbers. - The Python file contains the model and runnable example or training entry point.

Parameters49,600
Context128
Weights198.9 KB
Licensebsd-3-clause
AccessOpen weights
Monthly Downloads

Runs On

What it takes to serve multitask-scratch (49,600 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 0.0 GB 0.0 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 Sep 18, 2026.

Model Card

By Priscila, published under bsd-3-clause, revision 47451889e5fb.

A research-oriented Albef prototype targeting Multitask. The included small setup documents defaults and file formats without presenting unverified performance numbers. - The Python file contains the model and runnable example or training entry point. - config.json records the generated architecture settings. - trainingargs.json records the default experiment recipe. - model.safetensors is a valid initialization checkpoint for smoke tests; it is not presented as a trained benchmark checkpoint. - No benchmark score is claimed in this repository. The included configuration uses adamw with a cosine schedule. These are starting values in the script, not evidence of a completed run. For a…

Read Priscila's full model card

Albef for Multitask

Overview

A research-oriented Albef prototype targeting Multitask. The included small setup documents defaults and file formats without presenting unverified performance numbers.

Repository status

  • The Python file contains the model and runnable example or training entry point.
  • config.json records the generated architecture settings.
  • training_args.json records the default experiment recipe.
  • model.safetensors is a valid initialization checkpoint for smoke tests; it is not presented as a trained benchmark checkpoint.
  • No benchmark score is claimed in this repository.

Architecture

Item Value
Architecture Albef
Scale small
Attention standard
Fusion co attention
Activation approx gelu
Normalization instancenorm

Default experiment recipe

The included configuration uses adamw with a cosine schedule. These are starting values in the script, not evidence of a completed run. For a meaningful evaluation, train all baselines with the same data exposure, tuning budget, and random seeds.

Quick check

python run.py --help

Inspect the script's __main__ block for its generated smoke-test example. Because this is a custom implementation, generic automatic loading APIs require an explicit adapter before use.

Evaluation guidance

A useful first evaluation would use a task-specific held-out set, report the task metric across at least three seeds, and include a matched-capacity baseline. Keep training logs and environment versions with any published result.

Limitations

The initialization checkpoint has not been trained or audited for robustness, fairness, or domain transfer. The implementation should be treated as an experimental starting point. Results from a future trained checkpoint must be documented separately from the defaults shipped here.

Files

  • run.py — primary artifact
  • README.md — this documentation
  • config.json — architecture configuration
  • training_args.json — default experiment settings
  • model.safetensors — initialization checkpoint

License

Released under bsd-3-clause. Review the source-data terms separately when this repository is used with external datasets.

Configuration

Architecture
CustomResearchModel
Context length (tokens)
128
Layers
3
Hidden size
384
Feed-forward size
1,536
Attention heads
4
Model type
albef

Identity and Version

Repository
pecarvalho/multitask-scratch
Publisher
Priscila
Task
Not stated by the source
Modality
Other
Library
Not stated by the source
Parameters
49,600 parameters
Languages
Not stated by the source
Revision
47451889e5fb6d1729b64e62e7078ab86b8e8bb6
First published
2026-09-18
Last updated
2026-09-18

Files and Weights

6 files, 208.1 KB in total. The weights are 1 file totalling 198.9 KB in safetensors.

Weights1 file · 198.9 KB
Configuration3 files · 5.2 KB
Documentation1 file · 2.4 KB
Repository1 file · 1.5 KB
Every file
FileTypeSizeSHA-256
model.safetensorsWeights198.9 KB 1dd709633444
config.jsonConfiguration436 B
run.pyConfiguration4.6 KB
training_args.jsonConfiguration189 B
README.mdDocumentation2.4 KB
.gitattributesRepository1.5 KB

License and Download

License
bsd-3-clause
Access
Open weights, no gate
Download size
198.9 KB
Download from Priscila

Released by Priscila through its official repository on Hugging Face. Read the license.

Memory Requirements

PrecisionWeights in memory
As published198.9 KB
16-bit0.0 GB
8-bit0.0 GB
4-bit0.0 GB

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

Questions About multitask-scratch

How much GPU memory does multitask-scratch need?

About 0 GB at 16-bit and 0 GB at 4-bit: the weights (49,600 parameters) plus a working margin. A long context needs more.

What is the cheapest GPU to run multitask-scratch 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 multitask-scratch commercially?

Yes. multitask-scratch is released under BSD 3-Clause License. The BSD 3-Clause License is permissive. It permits commercial use and redistribution with the copyright notice, and forbids using the authors' names to endorse derived products without permission.

What is multitask-scratch's context length?

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