Upload baseline.txt with huggingface_hub
Browse files- baseline.txt +22 -22
baseline.txt
CHANGED
|
@@ -1,27 +1,27 @@
|
|
| 1 |
-
test bpe-gpt2/encode ... bench: 1,
|
| 2 |
-
test bpe-gpt2/encode-batch ... bench:
|
| 3 |
-
test bpe-gpt2/encode-no-cache ... bench: 2,
|
| 4 |
-
test bpe-gpt2/encode-batch-no-cache ... bench:
|
| 5 |
|
| 6 |
-
test llama3/encode ... bench: 1,
|
| 7 |
-
test llama3/encode-batch ... bench:
|
| 8 |
-
test llama3/encode-fast ... bench: 1,
|
| 9 |
-
test llama3/encode-char-offsets ... bench:
|
| 10 |
-
test llama3/concurrent-4t ... bench:
|
| 11 |
|
| 12 |
-
test decode/llama3 ... bench:
|
| 13 |
-
test decode/llama3-batch ... bench:
|
| 14 |
-
test decode/llama3-stream ... bench:
|
| 15 |
-
test decode/sp-chain ... bench:
|
| 16 |
-
test decode/wordpiece ... bench:
|
| 17 |
|
| 18 |
-
test serialization/from-file-roberta ... bench:
|
| 19 |
-
test serialization/from-file-llama3 ... bench:
|
| 20 |
-
test serialization/from-file-albert ... bench:
|
| 21 |
-
test serialization/bpe-from-file-gpt2 ... bench:
|
| 22 |
-
test serialization/save-llama3 ... bench:
|
| 23 |
-
test serialization/deserialize-roberta ... bench:
|
| 24 |
-
test serialization/deserialize-llama3 ... bench:
|
| 25 |
|
| 26 |
-
test train/bpe-small ... bench:
|
| 27 |
|
|
|
|
| 1 |
+
test bpe-gpt2/encode ... bench: 1,636,755,475 ns/iter (+/- 41,228,529)
|
| 2 |
+
test bpe-gpt2/encode-batch ... bench: 490,593,127 ns/iter (+/- 37,977,404)
|
| 3 |
+
test bpe-gpt2/encode-no-cache ... bench: 2,280,961,713 ns/iter (+/- 23,249,283)
|
| 4 |
+
test bpe-gpt2/encode-batch-no-cache ... bench: 584,604,754 ns/iter (+/- 30,599,336)
|
| 5 |
|
| 6 |
+
test llama3/encode ... bench: 1,693,147,422 ns/iter (+/- 52,409,827)
|
| 7 |
+
test llama3/encode-batch ... bench: 514,246,711 ns/iter (+/- 32,761,026)
|
| 8 |
+
test llama3/encode-fast ... bench: 1,614,659,817 ns/iter (+/- 19,849,046)
|
| 9 |
+
test llama3/encode-char-offsets ... bench: 439,538,627 ns/iter (+/- 12,892,872)
|
| 10 |
+
test llama3/concurrent-4t ... bench: 33,219,021 ns/iter (+/- 969,360)
|
| 11 |
|
| 12 |
+
test decode/llama3 ... bench: 226,576,519 ns/iter (+/- 1,856,451)
|
| 13 |
+
test decode/llama3-batch ... bench: 52,538,827 ns/iter (+/- 2,355,484)
|
| 14 |
+
test decode/llama3-stream ... bench: 142,816,609 ns/iter (+/- 1,326,465)
|
| 15 |
+
test decode/sp-chain ... bench: 503,798,936 ns/iter (+/- 1,634,827)
|
| 16 |
+
test decode/wordpiece ... bench: 995,571,495 ns/iter (+/- 1,024,628)
|
| 17 |
|
| 18 |
+
test serialization/from-file-roberta ... bench: 74,388,980 ns/iter (+/- 4,107,775)
|
| 19 |
+
test serialization/from-file-llama3 ... bench: 450,506,032 ns/iter (+/- 10,257,606)
|
| 20 |
+
test serialization/from-file-albert ... bench: 60,578,981 ns/iter (+/- 15,658,817)
|
| 21 |
+
test serialization/bpe-from-file-gpt2 ... bench: 53,119,294 ns/iter (+/- 1,937,936)
|
| 22 |
+
test serialization/save-llama3 ... bench: 124,862,986 ns/iter (+/- 7,162,323)
|
| 23 |
+
test serialization/deserialize-roberta ... bench: 86,138,751 ns/iter (+/- 4,386,315)
|
| 24 |
+
test serialization/deserialize-llama3 ... bench: 446,240,134 ns/iter (+/- 12,147,370)
|
| 25 |
|
| 26 |
+
test train/bpe-small ... bench: 30,924,718 ns/iter (+/- 712,539)
|
| 27 |
|