File size: 2,540 Bytes
df8b48f | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 | ---
license: mit
tags:
- pytorch
- mobilevit
- multitask
- prototype
---
# Mobilevit experiments: Multitask
## Overview
Working implementation of **Mobilevit** for **Multitask** using a **large** configuration. The repository focuses on transparent code and repeatable smoke tests; benchmark claims are deliberately omitted.
---
## Training details
| Item | Value |
|---|---|
| Framework | PyTorch |
| Precision | fp32 |
| Init scheme | kaiming |
| Hardware | a single consumer GPU is enough |
| Typical run | minutes, not hours |
---
## Changelog
1. 2026-10-05 — initial upload
2. 2026-10-20 — config and docs sync
---
## Acknowledgments
Thanks to the maintainers of the open-source libraries this project leans on.
---
## Repository status
One Python module contains the model and a runnable example.
Generated configuration lives in `config.json`; the recipe in `training_args.json`.
The checkpoint is an initialization artifact and should not be read as a trained model.
Scores are deliberately absent.
---
## Evaluation guidance
A first evaluation could use a task-specific held-out set. Report the metric across at least three seeds and include a matched-capacity baseline; keep logs and library versions with any published number.
---
## 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.
---
## Default experiment recipe
The generated recipe uses adamw with a polynomial schedule.
These values exist so the script runs end-to-end. A meaningful experiment needs its own tuning budget and at least three seeds.
---
## Architecture
| Item | Value |
|---|---|
| Architecture | Mobilevit |
| Scale | large |
| Attention | sparse |
| Fusion | bilinear |
| Activation | swish |
| Normalization | scalenorm |
---
## Files
| File | Purpose |
|---|---|
| `eval.py` | primary artifact |
| `README.md` | this documentation |
| `config.json` | architecture configuration |
| `training_args.json` | default experiment settings |
| `model.safetensors` | initialization checkpoint |
| `USE_POLICY.md` | use and redistribution notes |
---
## License
Released under **mit**. Review the source-data terms separately when this repository is used with external datasets.
|