Skip to content

Commit a741be8

Browse files
authored
Fix Whisper tokenization tests (#68)
After https://huggingface.co/openai/whisper-large-v2/commit/ae4642769ce2ad8fc292556ccea8e901f1530655 The previous version was still cached in my computer :(
1 parent c2e1e0f commit a741be8

File tree

2 files changed

+2
-2
lines changed

2 files changed

+2
-2
lines changed

Tests/TokenizersTests/FactoryTests.swift

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -18,6 +18,6 @@ class FactoryTests: XCTestCase {
1818
func testWhisper() async throws {
1919
let tokenizer = try await AutoTokenizer.from(pretrained: "openai/whisper-large-v2")
2020
let inputIds = tokenizer("Today she took a train to the West")
21-
XCTAssertEqual(inputIds, [50258, 50363, 27676, 750, 220, 83, 1212, 257, 220, 83, 7146, 220, 1353, 220, 3322, 4055, 50257])
21+
XCTAssertEqual(inputIds, [50258, 50363, 27676, 750, 1890, 257, 3847, 281, 264, 4055, 50257])
2222
}
2323
}
Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1 +1 @@
1-
{"text": "Fatouville-Grestain est une commune du Nord-Ouest du d\u00e9partement de l'Eure situ\u00e9e au \nbord de l'estuaire de la Seine et \u00e0 proximit\u00e9 du d\u00e9partement du Calvados. Selon l'atlas des paysages \nde Haute-Normandie, elle appartient \u00e0 la r\u00e9gion naturelle du Lieuvin. Toutefois, l'Agreste, le service \nde la statistique et de la prospective du minist\u00e8re de l'Agriculture, de l'Agroalimentaire et de la For\u00eat, \nla classe au sein du pays d'Auge (en tant que r\u00e9gion agricole).La commune est \u00e0 moins de dix kilom\u00e8tres \u00e0 \nl'est de Honfleur, \u00e0 autant de Beuzeville et \u00e0 environ dix-sept kilom\u00e8tres de Pont-Audemer.", "bpe_tokens": ["F", "at", "ou", "ville", "-", "G", "rest", "ain", "\u0120est", "\u0120une", "\u0120commun", "e", "\u0120du", "\u0120Nord", "-", "O", "uest", "\u0120du", "\u0120d\u00c3\u00a9part", "ement", "\u0120de", "\u0120l", "'", "E", "ure", "\u0120situ", "\u00c3\u00a9e", "\u0120au", "\u0120", "\u010a", "b", "ord", "\u0120de", "\u0120l", "'", "est", "ua", "ire", "\u0120de", "\u0120la", "\u0120Se", "ine", "\u0120et", "\u0120\u00c3\u0142", "\u0120proxim", "it\u00c3\u00a9", "\u0120du", "\u0120d\u00c3\u00a9part", "ement", "\u0120du", "\u0120Cal", "v", "ados", ".", "\u0120Sel", "on", "\u0120l", "'", "at", "las", "\u0120des", "\u0120pays", "ages", "\u0120", "\u010a", "de", "\u0120Ha", "ute", "-", "N", "orm", "and", "ie", ",", "\u0120elle", "\u0120app", "art", "ient", "\u0120\u00c3\u0142", "\u0120la", "\u0120r\u00c3\u00a9gion", "\u0120nature", "lle", "\u0120du", "\u0120Lie", "u", "vin", ".", "\u0120Tout", "ef", "ois", ",", "\u0120l", "'", "A", "gre", "ste", ",", "\u0120le", "\u0120service", "\u0120", "\u010a", "de", "\u0120la", "\u0120statist", "ique", "\u0120et", "\u0120de", "\u0120la", "\u0120prospective", "\u0120du", "\u0120minist", "\u00c3\u00a8re", "\u0120de", "\u0120l", "'", "A", "gr", "ic", "ulture", ",", "\u0120de", "\u0120l", "'", "A", "gro", "al", "iment", "aire", "\u0120et", "\u0120de", "\u0120la", "\u0120For", "\u00c3\u00aat", ",", "\u0120", "\u010a", "la", "\u0120classe", "\u0120au", "\u0120sein", "\u0120du", "\u0120pays", "\u0120d", "'", "A", "uge", "\u0120(", "en", "\u0120", "tant", "\u0120que", "\u0120r\u00c3\u00a9gion", "\u0120agric", "ole", ").", "La", "\u0120commun", "e", "\u0120est", "\u0120\u00c3\u0142", "\u0120moins", "\u0120de", "\u0120d", "ix", "\u0120kil", "om", "\u00c3\u00a8tres", "\u0120\u00c3\u0142", "\u0120", "\u010a", "l", "'", "est", "\u0120de", "\u0120Hon", "fle", "ur", ",", "\u0120\u00c3\u0142", "\u0120autant", "\u0120de", "\u0120Be", "u", "ze", "ville", "\u0120et", "\u0120\u00c3\u0142", "\u0120environ", "\u0120d", "ix", "-", "se", "pt", "\u0120kil", "om", "\u00c3\u00a8tres", "\u0120de", "\u0120Pont", "-", "Aud", "emer", "."], "token_ids": [50258, 50363, 37, 267, 263, 8386, 12, 38, 4149, 491, 871, 2251, 1199, 68, 1581, 16229, 12, 46, 11493, 1581, 37745, 1712, 368, 287, 6, 36, 540, 2054, 3856, 1609, 220, 198, 65, 765, 368, 287, 6, 377, 4398, 621, 368, 635, 1100, 533, 1030, 1531, 21932, 5066, 1581, 37745, 1712, 1581, 3511, 85, 4181, 13, 10736, 266, 287, 6, 267, 7743, 730, 10604, 1660, 220, 198, 1479, 4064, 1169, 12, 45, 687, 474, 414, 11, 8404, 724, 446, 1196, 1531, 635, 42669, 3687, 2447, 1581, 11197, 84, 4796, 13, 20453, 5666, 7376, 11, 287, 6, 32, 33248, 2941, 11, 476, 2643, 220, 198, 1479, 635, 16012, 1925, 1030, 368, 635, 39377, 1581, 16182, 4212, 368, 287, 6, 32, 861, 299, 8389, 11, 368, 287, 6, 32, 17427, 304, 2328, 9020, 1030, 368, 635, 1171, 15519, 11, 220, 198, 875, 32400, 1609, 6195, 1581, 10604, 274, 6, 32, 7181, 522, 268, 220, 35717, 631, 42669, 9682, 4812, 3050, 5478, 1199, 68, 871, 1531, 13099, 368, 274, 970, 5128, 298, 36506, 1531, 220, 198, 75, 6, 377, 368, 6625, 38921, 374, 11, 1531, 34081, 368, 879, 84, 1381, 8386, 1030, 1531, 2571, 274, 970, 12, 405, 662, 5128, 298, 36506, 368, 41127, 12, 15591, 29660, 13, 50257], "decoded_text": "<|startoftranscript|><|notimestamps|>Fatouville-Grestain est une commune du Nord-Ouest du d\u00e9partement de l'Eure situ\u00e9e au \nbord de l'estuaire de la Seine et \u00e0 proximit\u00e9 du d\u00e9partement du Calvados. Selon l'atlas des paysages \nde Haute-Normandie, elle appartient \u00e0 la r\u00e9gion naturelle du Lieuvin. Toutefois, l'Agreste, le service \nde la statistique et de la prospective du minist\u00e8re de l'Agriculture, de l'Agroalimentaire et de la For\u00eat, \nla classe au sein du pays d'Auge (en tant que r\u00e9gion agricole).La commune est \u00e0 moins de dix kilom\u00e8tres \u00e0 \nl'est de Honfleur, \u00e0 autant de Beuzeville et \u00e0 environ dix-sept kilom\u00e8tres de Pont-Audemer.<|endoftext|>"}
1+
{"text": "Fatouville-Grestain est une commune du Nord-Ouest du d\u00e9partement de l'Eure situ\u00e9e au \nbord de l'estuaire de la Seine et \u00e0 proximit\u00e9 du d\u00e9partement du Calvados. Selon l'atlas des paysages \nde Haute-Normandie, elle appartient \u00e0 la r\u00e9gion naturelle du Lieuvin. Toutefois, l'Agreste, le service \nde la statistique et de la prospective du minist\u00e8re de l'Agriculture, de l'Agroalimentaire et de la For\u00eat, \nla classe au sein du pays d'Auge (en tant que r\u00e9gion agricole).La commune est \u00e0 moins de dix kilom\u00e8tres \u00e0 \nl'est de Honfleur, \u00e0 autant de Beuzeville et \u00e0 environ dix-sept kilom\u00e8tres de Pont-Audemer.", "bpe_tokens": ["F", "at", "ou", "ville", "-", "G", "rest", "ain", "\u0120est", "\u0120une", "\u0120commun", "e", "\u0120du", "\u0120Nord", "-", "O", "uest", "\u0120du", "\u0120d\u00c3\u00a9part", "ement", "\u0120de", "\u0120l", "'", "E", "ure", "\u0120situ", "\u00c3\u00a9e", "\u0120au", "\u0120", "\u010a", "b", "ord", "\u0120de", "\u0120l", "'", "est", "ua", "ire", "\u0120de", "\u0120la", "\u0120Se", "ine", "\u0120et", "\u0120\u00c3\u0142", "\u0120proxim", "it\u00c3\u00a9", "\u0120du", "\u0120d\u00c3\u00a9part", "ement", "\u0120du", "\u0120Cal", "v", "ados", ".", "\u0120Sel", "on", "\u0120l", "'", "at", "las", "\u0120des", "\u0120pays", "ages", "\u0120", "\u010a", "de", "\u0120Ha", "ute", "-", "N", "orm", "and", "ie", ",", "\u0120elle", "\u0120app", "art", "ient", "\u0120\u00c3\u0142", "\u0120la", "\u0120r\u00c3\u00a9gion", "\u0120nature", "lle", "\u0120du", "\u0120Lie", "u", "vin", ".", "\u0120Tout", "ef", "ois", ",", "\u0120l", "'", "A", "gre", "ste", ",", "\u0120le", "\u0120service", "\u0120", "\u010a", "de", "\u0120la", "\u0120statist", "ique", "\u0120et", "\u0120de", "\u0120la", "\u0120prospective", "\u0120du", "\u0120minist", "\u00c3\u00a8re", "\u0120de", "\u0120l", "'", "A", "gr", "ic", "ulture", ",", "\u0120de", "\u0120l", "'", "A", "gro", "al", "iment", "aire", "\u0120et", "\u0120de", "\u0120la", "\u0120For", "\u00c3\u00aat", ",", "\u0120", "\u010a", "la", "\u0120classe", "\u0120au", "\u0120sein", "\u0120du", "\u0120pays", "\u0120d", "'", "A", "uge", "\u0120(", "en", "\u0120tant", "\u0120que", "\u0120r\u00c3\u00a9gion", "\u0120agric", "ole", ").", "La", "\u0120commun", "e", "\u0120est", "\u0120\u00c3\u0142", "\u0120moins", "\u0120de", "\u0120d", "ix", "\u0120kil", "om", "\u00c3\u00a8tres", "\u0120\u00c3\u0142", "\u0120", "\u010a", "l", "'", "est", "\u0120de", "\u0120Hon", "fle", "ur", ",", "\u0120\u00c3\u0142", "\u0120autant", "\u0120de", "\u0120Be", "u", "ze", "ville", "\u0120et", "\u0120\u00c3\u0142", "\u0120environ", "\u0120d", "ix", "-", "se", "pt", "\u0120kil", "om", "\u00c3\u00a8tres", "\u0120de", "\u0120Pont", "-", "Aud", "emer", "."], "token_ids": [50258, 50363, 37, 267, 263, 8386, 12, 38, 4149, 491, 871, 2251, 1199, 68, 1581, 16229, 12, 46, 11493, 1581, 37745, 1712, 368, 287, 6, 36, 540, 2054, 3856, 1609, 220, 198, 65, 765, 368, 287, 6, 377, 4398, 621, 368, 635, 1100, 533, 1030, 1531, 21932, 5066, 1581, 37745, 1712, 1581, 3511, 85, 4181, 13, 10736, 266, 287, 6, 267, 7743, 730, 10604, 1660, 220, 198, 1479, 4064, 1169, 12, 45, 687, 474, 414, 11, 8404, 724, 446, 1196, 1531, 635, 42669, 3687, 2447, 1581, 11197, 84, 4796, 13, 20453, 5666, 7376, 11, 287, 6, 32, 33248, 2941, 11, 476, 2643, 220, 198, 1479, 635, 16012, 1925, 1030, 368, 635, 39377, 1581, 16182, 4212, 368, 287, 6, 32, 861, 299, 8389, 11, 368, 287, 6, 32, 17427, 304, 2328, 9020, 1030, 368, 635, 1171, 15519, 11, 220, 198, 875, 32400, 1609, 6195, 1581, 10604, 274, 6, 32, 7181, 522, 268, 12095, 631, 42669, 9682, 4812, 3050, 5478, 1199, 68, 871, 1531, 13099, 368, 274, 970, 5128, 298, 36506, 1531, 220, 198, 75, 6, 377, 368, 6625, 38921, 374, 11, 1531, 34081, 368, 879, 84, 1381, 8386, 1030, 1531, 2571, 274, 970, 12, 405, 662, 5128, 298, 36506, 368, 41127, 12, 15591, 29660, 13, 50257], "decoded_text": "<|startoftranscript|><|notimestamps|>Fatouville-Grestain est une commune du Nord-Ouest du d\u00e9partement de l'Eure situ\u00e9e au \nbord de l'estuaire de la Seine et \u00e0 proximit\u00e9 du d\u00e9partement du Calvados. Selon l'atlas des paysages \nde Haute-Normandie, elle appartient \u00e0 la r\u00e9gion naturelle du Lieuvin. Toutefois, l'Agreste, le service \nde la statistique et de la prospective du minist\u00e8re de l'Agriculture, de l'Agroalimentaire et de la For\u00eat, \nla classe au sein du pays d'Auge (en tant que r\u00e9gion agricole).La commune est \u00e0 moins de dix kilom\u00e8tres \u00e0 \nl'est de Honfleur, \u00e0 autant de Beuzeville et \u00e0 environ dix-sept kilom\u00e8tres de Pont-Audemer.<|endoftext|>"}

0 commit comments

Comments
 (0)