Quark-50M 50M
Our compact 50M parameter model β engineered for extremely hyper-low resource systems.
Lightweight and highly volatile, ideal for basic sequence prediction and embedded units.
Explore Weights βBuilding highly efficient, logic-driven Small Language Models that run natively on edge hardware and consumer devices.
A lightweight bilingual language model optimized for speed and localized logic. Click to expand variants.
Features GQA, SwiGLU, RMSNorm, and RoPE. Trained on 50B+ tokens of ultra-curated data.
Our scaled small model featuring 32 layers and 768 hidden dimensions for advanced reasoning capabilities.
Equipped with a dense 65K vocabulary. Specially designed for multi-turn instruct fine-tuning.
Our compact 50M parameter model β engineered for extremely hyper-low resource systems.
Lightweight and highly volatile, ideal for basic sequence prediction and embedded units.
Explore Weights βInstruction-tuned math & code model. Fine-tuned via SFT on a base checkpoint pre-trained on 5B tokens of math, code, and reasoning data.
14 layers, 512 hidden size, GQA (8Q/2KV), SwiGLU, RoPE. Pre-trained on 5B tokens, then SFTβaligned. License: Apacheβ2.0.
Explore Weights βA high-throughput multi-label moderation engine covering 9 toxicity and cyber-safety categories.
Detects: toxic, severe_toxic, obscene, threat, insult, identity_hate, and advanced content exploits.
View Classifier βMastering sub-1B parameters using Grouped-Query Attention (GQA) architectures.
Integrating step-by-step reasoning logic directly into the pre-training tokens.
High-density blending of localized Italian, English, and technical STEM pipelines.
All weights, configurations, and baseline streaming datasets are entirely open to the world.
| Repository Object | Distribution Target |
|---|---|
| π Quark-135M Base | Foundational localized small baseline language matrix. |
| π Quark-135M Bilingual | Bilingual (IT + EN) checkpoint trained on balanced multi-source pools. |
| π Quark-270M Instruct | Multi-turn conversation alignment model with strict formatting safety. |
| βοΈ Quark-270M Base | Base engine ready for specialized downstream task tokenization. |
| π¦ Quark-50M | Legacy foundational checkpoint for exploratory sequence architectures. |
| π‘οΈ Quark-Mod | Production safety guardrail classifier for modern pipeline filtering. |
| β‘οΈ Complete Collection (v0.1) | Unified access hub to all current generation architectural releases. |
| π» GitHub Organization | Training codebases, data streaming pipelines, and infrastructure layers. |