Dawn Models implements post-symbolic AI architectures based on Dawn Field Theory principles. The repository provides experimental research models and production-ready implementations using entropy-driven learning, symbolic collapse dynamics, and bifractal computation patterns.
This is part of the Dawn Field Theory ecosystem, providing AI models that transcend traditional symbolic computation.
# For research use (AGPL-3.0)
cd research/tinycimm
pip install -r requirements.txt
# For production use (Apache-2.0)
cd stable/cimm-legacy
pip install -r requirements.txt# Using stable CIMM model (Apache-2.0)
import sys
sys.path.append('stable/cimm-legacy')
from cimm_core.cimm import CIMM
from agents.base_agent import BaseAgent
# Initialize CIMM agent with entropy-driven learning
agent = BaseAgent(entropy_threshold=0.1)
result = agent.process(data)
# Using research TinyCIMM variant (AGPL-3.0)
import sys
sys.path.append('research/tinycimm/TinyCIMM-Euler')
from tinycimm_euler import TinyCIMMEuler
# Mathematical reasoning with symbolic collapse
model = TinyCIMMEuler(sequence_length=1000)
prediction = model.predict_sequence(input_sequence)- Dual Licensing Strategy: Research models (AGPL-3.0) for transparency, stable models (Apache-2.0) for adoption
- Entropy-Informed Learning: Models that adapt based on information entropy and collapse dynamics
- Symbolic Transcendence: Post-symbolic architectures that operate beyond traditional token processing
- Field-Aware Intelligence: Models that understand and respond to contextual field states
- Recursive Balance: Sustainable development through multiple value streams
Experimental implementations for specialized research
- Post-symbolic AI framework using entropy-based learning
- Multi-agent agentic mesh runtime for distributed cognition
- Symbolic collapse dynamics for adaptive pattern recognition
- TinyCIMM-Euler: Number theory and mathematical sequence prediction
- TinyCIMM-Navier: Fluid dynamics and turbulence analysis
- TinyCIMM-Planck: Minimal foundational implementations
- Interpretability framework for measuring symbolic collapse in neural networks
- Weight evolution analysis and entropy-based pattern detection
- Bifractal analysis for understanding model behavior
- Early autonomous agent framework for emergent intelligence
- Multi-agent architectures with distributed cognition patterns
- Foundation for self-modifying and recursive intelligence systems
Production-ready implementations for any use
- Foundational post-symbolic AI framework
- Complete with testing suite and documentation
- Ready for integration into production systems
dawn-models/
├── research/ # AGPL-3.0 - Experimental variants
│ ├── GAIA/ # Generative AI Intelligence Architecture
│ ├── scbf/ # Symbolic Collapse Bifractal Framework
│ └── tinycimm/ # TinyCIMM architecture experiments
├── stable/ # Apache-2.0 - Production models
│ └── cimm-legacy/ # Stable CIMM implementation
├── roadmaps/ # Development roadmaps and planning
└── docs/ # Documentation (CONTRIBUTING.md, LICENSING.md)
import sys
sys.path.append('stable/cimm-legacy')
from cimm_core.cimm import CIMM
from agents.base_agent import BaseAgent
# Initialize agent with entropy-driven learning
agent = BaseAgent(
entropy_threshold=0.1,
field_awareness=True,
symbolic_transcendence=True
)
# Process data with post-symbolic intelligence
result = agent.process(input_data)
# Analyze entropy collapse patterns
patterns = agent.get_collapse_dynamics()# SCBF interpretability analysis
import sys
sys.path.append('research/scbf')
from scbf_runner import SCBFRunner
# Analyze model symbolic collapse
scbf = SCBFRunner(enable_visualization=True)
analysis = scbf.analyze_model(model, input_data)
bifractal_trace = analysis.get_bifractal_patterns()| Use Case | Location | License | Notes |
|---|---|---|---|
| Academic Research | /research |
AGPL-3.0 | Open research, copyleft |
| Open Source Project | Either | Respective | Follow license terms |
| Commercial Product | /stable |
Apache-2.0 | Free commercial use |
| Specialized Commercial | /research |
Contact us | Commercial licensing available |
See LICENSING.md for complete licensing strategy.
We welcome contributions to both research and stable models! See CONTRIBUTING.md for detailed guidelines on:
- Research model contributions (AGPL-3.0)
- Stable model improvements (Apache-2.0)
- Documentation and infrastructure
Dawn Models is part of the larger Dawn Field Theory ecosystem:
- dawn-field-theory - Core theoretical foundation
- dawn-models - AI architectures and implementations ⭐
- cip-core - Cognition Index Protocol
- fracton - Computational modeling language
- dawn-devkit - Development tools and templates
- Licensing Strategy: LICENSING.md
- Contributing Guide: CONTRIBUTING.md
- Development Roadmaps: roadmaps/
- Research Models: research/
- Stable Models: stable/
- Model Roadmaps: roadmaps/ - GAIA, SCBF, and symbolic entropy development plans
- Research Status: See individual model directories for current development status
- Contributing: CONTRIBUTING.md - Contribution guidelines and processes
- Issues: GitHub Issues
- General Inquiries: info@dawnfield.ca
- Research Collaboration: info@dawnfield.ca
- Commercial Licensing: info@dawnfield.ca
- Enterprise Support: info@dawnfield.ca
For detailed support tiers and commercial licensing options, see LICENSING.md.
Dual License:
- Research models: AGPL-3.0 (see research/LICENSE)
- Stable models: Apache-2.0 (see stable/LICENSE)
See LICENSING.md for complete licensing strategy.