foghollowleon/multitask-scratch

Dino for Multitask

Overview

Working implementation of Dino for Multitask using a tiny configuration. The repository focuses on transparent code and repeatable smoke tests; benchmark claims are deliberately omitted.

---

Acknowledgments

This repo borrows ideas freely from public reference implementations; credit to their authors.

---

Repository Status

• Single-file implementation; the __main__ block runs a smoke test.
• config.json and training_args.json hold the generated settings.
• model.safetensors is freshly initialized — plumbing, not results.
• No benchmark numbers are reported on purpose.

---

Limitations

• weights are random-init, so outputs carry no signal
• robustness, fairness and domain transfer are untested
• the API may change without notice while the design settles

---

Environment

Item	Value
Python	>= 3.9
torch	2.x
safetensors	any recent
transformers	optional, for loaders

---

Default Experiment Recipe

The generated recipe uses lion with a step schedule.

These values exist so the script runs end-to-end. A meaningful experiment needs its own tuning budget and at least three seeds.

---

Quick Check

Run python train.py --help for the built-in smoke test. Because this is a custom implementation, generic automatic loading APIs require an explicit adapter before use.

---

Files

File	Purpose
train.py	primary artifact
README.md	this documentation
config.json	architecture configuration
training_args.json	default experiment settings
model.safetensors	initialization checkpoint

---

License

Distributed under the apache-2.0 license. External datasets keep their own terms — check them separately.