API reference
API reference
Public exports, exact implementation aliases, and complete module and CLI references.
Public Python surface
This reference covers all 25 names in axosim.__all__ at revision 306a51ed950b, plus the public module functions, configuration fields, dataset methods, and CLI options used by the guides. The package lazily imports these names; implementation aliases are stated explicitly.
| Import from axosim | Implementation | Reference |
|---|---|---|
AXOSIM_POPULATION_PROFILE |
axosim.simulation_contract.AXOSIM_POPULATION_PROFILE |
Open |
MILLION_NEURON_REALTIME_CONTRACT |
axosim.simulation_contract.MILLION_NEURON_REALTIME_CONTRACT |
Open |
AxoMamba |
axosim.axomamba.AxoMamba |
Open |
AxoMambaConfig |
axosim.axomamba.AxoMambaConfig |
Open |
AxoSimGRU |
axosim.block_forecast.CausalBlockForecastModel |
Open |
AxoSimLite |
axosim.support_surrogate.AdaptiveSupportP4Surrogate |
Open |
AxoSimMamba |
axosim.axomamba.AxoMamba |
Open |
AxoSimModelProfile |
axosim.model_family.AxoSimModelProfile |
Open |
AxoSimPopulation |
axosim.interfaces.AxoSimPopulation |
Open |
CudaGraphAdaptationStep |
axosim.adaptation.CudaGraphAdaptationStep |
Open |
HomeostaticThresholdController |
axosim.activity.HomeostaticThresholdController |
Open |
AXOSIM_MODEL_PROFILES |
axosim.model_family.AXOSIM_MODEL_PROFILES |
Open |
BranchELM |
axosim.model.BranchELM |
Open |
BranchELMConfig |
axosim.model.BranchELMConfig |
Open |
LargePopulationSimulationContract |
axosim.simulation_contract.LargePopulationSimulationContract |
Open |
MixedQuantizedNeuronBehaviorBank |
axosim.population.MixedQuantizedNeuronBehaviorBank |
Open |
PopulationInferenceProfile |
axosim.simulation_contract.PopulationInferenceProfile |
Open |
ProceduralMorphologyConnectomeRouter |
axosim.connectome.ProceduralMorphologyConnectomeRouter |
Open |
QuantizedSynapticEfficacyBank |
axosim.synapse.QuantizedSynapticEfficacyBank |
Open |
compact_active_sources |
axosim.connectome.compact_active_sources |
Open |
dequantize_synaptic_efficacies |
axosim.synapse.dequantize_synaptic_efficacies |
Open |
dequantize_mixed_neuron_behavior_parameters |
axosim.population.dequantize_mixed_neuron_behavior_parameters |
Open |
quantize_synaptic_efficacies |
axosim.synapse.quantize_synaptic_efficacies |
Open |
quantize_mixed_neuron_behavior_parameters |
axosim.population.quantize_mixed_neuron_behavior_parameters |
Open |
create_axosim_profile |
axosim.model_family.create_axosim_profile |
Open |
HomeostaticThresholdController is experimental. Configuration recipes and workload constants describe construction or execution contracts; they are not benchmark measurements. Historical checkpoint classes remain documented because the loader supports them.
Modules and configuration
- Population interfaces
- Model profiles
- Mamba models and configuration
- GRU temporal models
- Block-forecast models
- Lite neuron and configuration
- Branch-ELM compatibility
- Checkpoints
- CUDA Graph adaptation
- Population simulation contracts
- Behavior banks and population runners
- Synaptic efficacy banks
- Connectome routing
- Activity and experimental control
- Trace datasets
- NeuronIO conversion
- Local metrics and dataset evaluation
- Training functions
- Inference benchmarking
- AxoBench prediction adapters
- Setup workflow
Command-line tools
Source and inventory
Every source link is pinned to revision 306a51ed950b. Repository access is currently required to open the AxoSim and AxoBench source links because these repositories are private. The machine-readable API inventory records source locations, exact signatures, parameters/defaults, fields, methods, exported aliases, and CLI options. It is generated using AST parsing without importing PyTorch.