All models integrated into SpikingTSF are listed below. Model implementations are adapted from the source repositories indicated. This library does not claim ownership of any original architecture.
When using a specific model, please cite both SpikingTSF and the corresponding source paper.
| Model | Type | Spike? | Source Paper | Source Repository | File in Repo | Status |
|---|---|---|---|---|---|---|
| SpikF-GO | Spiking Fourier Graph Operator (freq. domain, optional spike-domain CPG-PE) | ✅ | Bakhshaliyev & Landwehr, accepted to ECML PKDD 2026 (our own paper — link to be added on publication) | jafarbakhshaliyev/SpikF-GO | models/SpikF_GO.py |
Runnable, ETTh1+ETTh2 |
| SpikF | Transformer (freq. domain) | ✅ | Wu et al., ICML 2025 | WWJ-creator/SpikF | models/SpikF.py |
Runnable, ETTh1+ETTh2 |
| Spikformer | Transformer (spike-driven SA) | ✅ | SeqSNN, NeurIPS 2025 | microsoft/SeqSNN | models/Spikformer.py |
Runnable, ETTh1 |
| Spikingformer | Transformer (pre-LIF) | ✅ | SeqSNN, NeurIPS 2025 | microsoft/SeqSNN | models/Spikingformer.py |
Runnable, ETTh1+ETTh2 verified |
| QKFormer | Transformer (token-level Q/K attn) | ✅ | SeqSNN, NeurIPS 2025 | microsoft/SeqSNN | models/QKFormer.py |
Runnable, results pending |
| TSGRU | GRU (two-compartment TS-LIF) | ✅ | TS-LIF, ICLR 2025 | kkking-kk/TS-LIF | models/TSGRU.py |
Runnable, ETTh1+ETTh2 verified |
| TSTCN | TCN (two-compartment TS-LIF) | ✅ | TS-LIF, ICLR 2025 | kkking-kk/TS-LIF | models/TSTCN.py |
Runnable, ETTh1 verified |
| TSFormer | Transformer (two-compartment TS-LIF) | ✅ | TS-LIF, ICLR 2025 | kkking-kk/TS-LIF | models/TSFormer.py |
Runnable, ETTh1 verified |
| iSpikformer | Inverted Spiking Transformer | ✅ | SeqSNN, ICML 2024 | microsoft/SeqSNN | models/iSpikformer.py |
Runnable, ETTh1 verified |
| SpikeRNN | Spiking Recurrent Network | ✅ | SeqSNN, ICML 2024 | microsoft/SeqSNN | models/SpikeRNN.py |
Runnable, ETTh1+ETTh2 verified |
| SpikTCN | Spiking Temporal Convolutional Network | ✅ | SeqSNN, ICML 2024 | microsoft/SeqSNN | models/SpikTCN.py |
Runnable, ETTh1 partial |
| SpikGRU | Spiking Gated Recurrent Unit | ✅ | SeqSNN, ICML 2024 | microsoft/SeqSNN | models/SpikGRU.py |
Runnable, ETTh1+ETTh2 verified |
| iTransformer | Inverted Transformer (ANN) | ❌ | Liu et al., ICLR 2024 | thuml/iTransformer | models/ITransformer.py |
Runnable, results pending |
| DLinear | Decomposition Linear (ANN) | ❌ | Zeng et al., AAAI 2023 | thuml/Time-Series-Library | models/DLinear.py |
Runnable, results pending |
- Runnable: Code is integrated and can be executed with the provided scripts.
- Verified: Results have been produced and logged under
Output/. Numbers are in RESULTS.md. - Partial: Some horizons or datasets have been evaluated but not all.
- Pending: Code is present but results have not yet been computed.
See CONTRIBUTING.md and docs/adding_a_model.md.