mrwabbit commited on
Commit
4e7f4e0
·
verified ·
1 Parent(s): baa0a69

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +93 -0
README.md ADDED
@@ -0,0 +1,93 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language: en
3
+ license: mit
4
+ library_name: pytorch
5
+ tags:
6
+ - spiking-neural-network
7
+ - neuromorphic
8
+ - surrogate-gradient
9
+ - benchmark
10
+ - catalyst
11
+ - ssc
12
+ datasets:
13
+ - ssc
14
+ metrics:
15
+ - accuracy
16
+ model-index:
17
+ - name: Catalyst SSC SNN Benchmark
18
+ results:
19
+ - task:
20
+ type: audio-classification
21
+ name: Spoken Command Classification
22
+ dataset:
23
+ name: Spiking Speech Commands (SSC)
24
+ type: ssc
25
+ metrics:
26
+ - name: Float Accuracy
27
+ type: accuracy
28
+ value: 72.1
29
+ - name: Quantized Accuracy (int16)
30
+ type: accuracy
31
+ value: 71.6
32
+ ---
33
+
34
+ # Catalyst SSC SNN Benchmark
35
+
36
+ Spiking Neural Network for spoken command classification on SSC.
37
+
38
+ ## Model Description
39
+
40
+ - **Architecture**: 700 → 1024 (recurrent adLIF) → 512 (adLIF) → 35
41
+ - **Neuron model**: Adaptive Leaky Integrate-and-Fire (adLIF) with Symplectic Euler discretization
42
+ - **Training**: Surrogate gradient BPTT, fast-sigmoid surrogate (scale=25)
43
+ - **Hardware target**: Catalyst N1/N2/N3 neuromorphic processors
44
+ - **Quantization**: Float32 weights -> int16, membrane decay -> 12-bit fixed-point
45
+
46
+ ## Results
47
+
48
+ | Metric | Value |
49
+ |--------|-------|
50
+ | Float accuracy | 72.1% |
51
+ | Quantized accuracy (int16) | 71.6% |
52
+ | Parameters | 2,313,763 |
53
+ | Quantization loss | 0.5% |
54
+
55
+ ## Reproduce
56
+
57
+ ```bash
58
+ git clone https://github.com/catalyst-neuromorphic/catalyst-benchmarks.git
59
+ cd catalyst-benchmarks
60
+ pip install -e .
61
+ python ssc/train.py --device cuda:0
62
+ ```
63
+
64
+ ## Deploy to Catalyst Hardware
65
+
66
+ ```python
67
+ import catalyst_cloud
68
+
69
+ client = catalyst_cloud.Client()
70
+ result = client.simulate(
71
+ model="catalyst-neuromorphic/ssc-snn-benchmark",
72
+ input_data=your_spikes,
73
+ processor="n2"
74
+ )
75
+ ```
76
+
77
+ ## Links
78
+
79
+ - **Benchmark repo**: [catalyst-neuromorphic/catalyst-benchmarks](https://github.com/catalyst-neuromorphic/catalyst-benchmarks)
80
+ - **Cloud API**: [catalyst-neuromorphic.com](https://catalyst-neuromorphic.com)
81
+ - **N2 paper**: [Zenodo DOI 10.5281/zenodo.18728256](https://zenodo.org/records/18728256)
82
+ - **N1 paper**: [Zenodo DOI 10.5281/zenodo.18727094](https://zenodo.org/records/18727094)
83
+
84
+ ## Citation
85
+
86
+ ```bibtex
87
+ @misc{catalyst-benchmarks-2026,
88
+ author = {Shulayev Barnes, Henry},
89
+ title = {Catalyst Neuromorphic Benchmarks},
90
+ year = {2026},
91
+ url = {https://github.com/catalyst-neuromorphic/catalyst-benchmarks}
92
+ }
93
+ ```