You don’t “run a model”
You run Kernels
The model is just a graph
The Inference Engine is scheduler / optimizer / executor
But the actual work? That happens in the Kernels
- MatMul Kernels
- Attention Kernels
- RMSNorm Kernels
- KV cache Kernels
- Quantized linear Kernels
-
31
943
97
126.9K
Snapshot History (26)
| Date | ||||
|---|---|---|---|---|
| Apr 28 · 03:00 | 31 | +3 943 | 97 | +1.3K 126.9K |
| Apr 28 · 02:00 | 31 | +4 940 | +1 97 | +2.9K 125.6K |
| Apr 27 · 23:00 | 31 | +1 936 | 96 | +1.4K 122.8K |
| Apr 27 · 22:00 | 31 | +10 935 | +1 96 | +4.3K 121.4K |
| Apr 27 · 19:00 | 31 | +2 925 | 95 | +1.8K 117.1K |
| Apr 27 · 18:00 | 31 | +15 923 | +3 95 | +5.2K 115.2K |
| Apr 27 · 15:00 | +1 31 | +5 908 | 92 | +2.1K 110.1K |
| Apr 27 · 14:00 | +1 30 | +16 903 | +1 92 | +5.3K 108.0K |
| Apr 27 · 11:00 | 29 | +9 887 | +3 91 | +1.6K 102.7K |
| Apr 27 · 10:00 | 29 | +31 878 | +2 88 | +5.1K 101.1K |
| Apr 27 · 07:00 | 29 | +10 847 | +2 86 | +2.4K 96.0K |
| Apr 27 · 06:00 | 29 | +26 837 | +3 84 | +7.8K 93.6K |
| Apr 27 · 03:00 | 29 | +11 811 | 81 | +2.8K 85.7K |
| Apr 27 · 02:00 | +1 29 | +43 800 | +2 81 | +9.9K 82.9K |
| Apr 26 · 23:00 | 28 | +14 757 | +1 79 | +3.9K 73.0K |
| Apr 26 · 22:00 | +6 28 | +49 743 | +2 78 | +15.2K 69.1K |
| Apr 26 · 19:00 | 22 | +31 694 | +1 76 | +8.8K 53.9K |
| Apr 26 · 18:00 | +2 22 | +34 663 | +3 75 | +9.5K 45.1K |
| Apr 26 · 17:00 | +1 20 | +74 629 | +10 72 | +15.3K 35.7K |
| Apr 26 · 15:00 | +1 19 | +36 555 | +6 62 | +3.2K 20.3K |
| Apr 26 · 14:00 | +2 18 | +50 519 | +5 56 | +1.9K 17.1K |
| Apr 26 · 13:00 | +2 16 | +69 469 | +11 51 | +3.5K 15.2K |
| Apr 26 · 11:00 | +1 14 | +53 400 | +8 40 | +1.8K 11.7K |
| Apr 26 · 10:00 | +2 13 | +62 347 | +6 32 | +2.2K 10.0K |
| Apr 26 · 09:00 | +4 11 | +102 285 | +6 26 | +3.3K 7.8K |
| Apr 26 · 07:00 | +7 7 | +183 183 | +20 20 | +4.5K 4.5K |
| Apr 26 · 02:55 | 0 | 0 | 0 | 0 |