license: apache-2.0 language:

• en
• multilingual tags:
• byte-level
• content-classification
• onnx
• edge-ai
• matryoshka
• multi-head
• classifier
• clipboard
• tiny
• fast
• code-detection
• language-detection
• open-source pipeline_tag: text-classification library_name: pico-type inference: parameters: provider: CPUExecutionProvider

pico-type 🔍

A tiny byte-level multi-head content classifier — ~1.5M params, ~209KB ONNX, <6ms inference.

Classifies any content into 7 categories from raw bytes in a single forward pass.

Built by eulogik — AI infrastructure for developers.

✨ Features

• No tokenizer — operates directly on raw UTF-8 bytes (supports all languages, zero pre-processing)
• 7 heads, one forward pass — coarse type, modality, subtype, code lang, text lang, file MIME, risk flags
• 4 Matryoshka tiers — tiny (16d) → small (64d) → base (192d) → pro (576d)
• ~200KB ONNX — deploy on edge devices, serverless functions, browser (WebAssembly)
• <6ms inference on CPU via ONNX Runtime (base tier, 1024 bytes)
• CLI, Gradio Space, MCP server — ready for any integration
• 62 programming languages — Python, JS, TypeScript, Java, C, C++, Go, Rust, SQL, Bash, and 52 more
• 95.2% real-world accuracy — tested against 21 hand-curated inputs across all content types

📊 Performance

Evaluated on 1000 synthetic samples + 21 hand-curated real-world inputs. Base tier, ~5ms inference.

Real-world accuracy: 95.2% (20/21) — The model correctly classifies code, text, markup, config, images, binary archives, and error tracebacks. Only failure: YAML config → predicts error (a fundamental byte-level ambiguity at 2KB context).

🚀 Quick Start

CLI

pip install pico-type

echo "def hello():\n    return 42" | picotype --pretty
picotype --file document.txt
picotype --clip

Python

from picotype import PicoType, PicoTypeConfig, decode_output

model = PicoType(PicoTypeConfig()).eval()
# ... load checkpoint ...
result = decode_output(model(b"input bytes"), tier="base")

MCP Server (Claude/Cursor)

PICOTYPE_MODEL_DIR=./checkpoints python -m model.pico_type.mcp_server

🏗 Architecture

Bytes → ByteEmbed(256→96d) → 3×Conv1D(k=3,5,7) → 2×BiAttention(RoPE) → Pool(mean‖max‖std) → 7×Matryoshka Heads

Total parameters: 1.43M (tiny) / 1.45M (small) / 1.48M (base) / 1.56M (pro)

🔧 Model Tiers

All tiers share the same trunk; only the final linear layers differ. Switch tiers at inference with zero overhead.

🧪 Classification Heads

🌐 Deployment

📚 Documentation

• Model Card — detailed architecture, training, evaluation
• Architecture Plan — full design document
• Walkthrough — development log with all decisions

📄 License

Apache 2.0 — free for commercial and personal use.