Beta

Explore
Marketplace
Neural Labs
Chat
Wallet
Docs

Beta

Explore Marketplace Neural Labs Wallet Docs

Build

Manage

My assets Manage wallet Organization

Community

Discord X (Twitter)Telegram

Docs Privacy Policy Terms and Conditions

Proud member of

Proud member of

Docs Cookie Policy

Privacy Policy Terms and Conditions

© 2026 AlphaNeural AI. All rights reserved.

commavq_compression – AI Model by SAT-oO | AlphaNeural AI

You can deploy this model and start earning money today!

SAT-oO/
commavq_compression

like

0

Views

No views yet

Commavq Video Compression Challenge

Lossless compression of commaVQ HuggingFace dashcam token sequences for comma.ai challenge evaluation.

Overall performance of 3.05x overall compression with training parameters specified in docs.

Available on HuggingFace: SAT-oO/commavq_compression

Active pipeline

The maintained end-to-end path is:

training/train_global.py — train next-frame predictor + checkpoints
model.py — 4.48M parameter transformer (8 frame context)
coder.py — range coding wrappers (constriction)
compress.py — build submission zip
decompress.py — reconstruct tokens in evaluator
test/evaluate.sh + test/evaluate.py — correctness + ratio check

Legacy/unused approaches are in legacy/.

Output artifacts

Training outputs:

resource/model.pt
resource/model_f16.pt
resource/global_freq.npy
resource/checkpoints/*.pt

Submission output:

submission.zip (or custom --output)

Notes

Training and evaluation download data from Hugging Face directly.
Manual local shard copies are optional.
compress.py can be memory-heavy on CPU; prefer CUDA for packaging.
Implementation details: docs/TECHNICAL_OVERVIEW.md.
To replicate trainig workflow: see docs/QUICK_START.md