tzervas/bwsk-vit-base

BWSK ViT-base

ViT-base (86M params) trained in 6 variants (3 BWSK modes x 2 experiments) on CIFAR-10 with full convergence training and early stopping.

This repo contains all model weights, configs, and training results in a single consolidated repository.

What is BWSK?

BWSK is a framework that classifies every neural network operation as S-type (information-preserving, reversible, coordination-free) or K-type (information-erasing, synchronization point) using combinator logic. This classification enables reversible backpropagation through S-phases to save memory, and CALM-based parallelism analysis.

Model Overview

Property	Value
Base Model	google/vit-base-patch16-224
Architecture	Vit (image_cls)
Parameters	86M
Dataset	CIFAR-10
Eval Metric	Accuracy

S/K Classification

Type	Ratio
S-type (information-preserving)	72.1%
K-type (information-erasing)	27.9%

Fine-tune Results

Mode	Final Loss	Val Accuracy	Test Accuracy	Peak Memory	Time	Epochs
Conventional	0.0022	97.8%	97.6%	3.1 GB	3.8m	1
BWSK Analyzed	0.3425	98.0%	98.2%	3.1 GB	8.4m	2
BWSK Reversible	0.0019	97.7%	97.3%	2.0 GB	4.5m	1

Memory savings (reversible vs conventional): 37.3%

From Scratch Results

Mode	Final Loss	Val Accuracy	Test Accuracy	Peak Memory	Time	Epochs
Conventional	1.5347	37.9%	37.5%	3.1 GB	7.6m	2
BWSK Analyzed	1.8406	38.0%	36.9%	3.1 GB	4.3m	1
BWSK Reversible	1.8934	39.6%	37.8%	2.0 GB	6.4m	2

Memory savings (reversible vs conventional): 37.3%

Repository Structure

├── README.md
├── results.json
├── finetune-conventional/
│   ├── model.safetensors
│   ├── config.json
│   └── training_results.json
├── finetune-bwsk-analyzed/
│   ├── model.safetensors
│   ├── config.json
│   └── training_results.json
├── finetune-bwsk-reversible/
│   ├── model.safetensors
│   ├── config.json
│   └── training_results.json
├── scratch-conventional/
│   ├── model.safetensors
│   ├── config.json
│   └── training_results.json
├── scratch-bwsk-analyzed/
│   ├── model.safetensors
│   ├── config.json
│   └── training_results.json
├── scratch-bwsk-reversible/
│   ├── model.safetensors
│   ├── config.json
│   └── training_results.json

Usage

Load a specific variant:

from transformers import AutoModelForImageClassification, AutoFeatureExtractor

# Load fine-tuned conventional variant
model = AutoModelForImageClassification.from_pretrained(
    "tzervas/bwsk-vit-base", subfolder="finetune-conventional"
)

Training Configuration

Setting	Value
Optimizer	AdamW
LR (fine-tune)	5e-05
LR (from-scratch)	3e-04
LR Schedule	Cosine with warmup
Max Grad Norm	1.0
Mixed Precision	AMP (float16)
Early Stopping	Patience 3
Batch Size	16

Links

• GitHub Repository
• Whitepaper
• Full Training Report

Citation

@software{zervas2026bwsk,
  author = {Zervas, Tyler},
  title = {BWSK: Combinator-Typed Neural Network Analysis},
  year = {2026},
  url = {https://github.com/tzervas/ai-s-combinator},
}

License

MIT