Spaces:
Runtime error
Runtime error
Commit
•
cd84165
1
Parent(s):
51e109c
French MTEB
Browse files- EXTERNAL_MODEL_RESULTS.json +0 -0
- app.py +467 -61
EXTERNAL_MODEL_RESULTS.json
CHANGED
The diff for this file is too large to render.
See raw diff
|
|
app.py
CHANGED
@@ -38,8 +38,6 @@ TASK_LIST_CLASSIFICATION = [
|
|
38 |
"TweetSentimentExtractionClassification",
|
39 |
]
|
40 |
|
41 |
-
TASK_LIST_CLASSIFICATION_NORM = [x.replace(" (en)", "") for x in TASK_LIST_CLASSIFICATION]
|
42 |
-
|
43 |
TASK_LIST_CLASSIFICATION_DA = [
|
44 |
"AngryTweetsClassification",
|
45 |
"DanishPoliticalCommentsClassification",
|
@@ -51,6 +49,15 @@ TASK_LIST_CLASSIFICATION_DA = [
|
|
51 |
"ScalaDaClassification",
|
52 |
]
|
53 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
54 |
TASK_LIST_CLASSIFICATION_NB = [
|
55 |
"NoRecClassification",
|
56 |
"NordicLangClassification",
|
@@ -115,6 +122,16 @@ TASK_LIST_CLUSTERING_DE = [
|
|
115 |
"TenKGnadClusteringS2S",
|
116 |
]
|
117 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
118 |
TASK_LIST_CLUSTERING_PL = [
|
119 |
"8TagsClustering",
|
120 |
]
|
@@ -132,6 +149,11 @@ TASK_LIST_PAIR_CLASSIFICATION = [
|
|
132 |
"TwitterURLCorpus",
|
133 |
]
|
134 |
|
|
|
|
|
|
|
|
|
|
|
135 |
TASK_LIST_PAIR_CLASSIFICATION_PL = [
|
136 |
"CDSC-E",
|
137 |
"PPC",
|
@@ -151,6 +173,11 @@ TASK_LIST_RERANKING = [
|
|
151 |
"StackOverflowDupQuestions",
|
152 |
]
|
153 |
|
|
|
|
|
|
|
|
|
|
|
154 |
TASK_LIST_RERANKING_ZH = [
|
155 |
"CMedQAv1",
|
156 |
"CMedQAv2",
|
@@ -176,6 +203,15 @@ TASK_LIST_RETRIEVAL = [
|
|
176 |
"TRECCOVID",
|
177 |
]
|
178 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
179 |
TASK_LIST_RETRIEVAL_PL = [
|
180 |
"ArguAna-PL",
|
181 |
"DBPedia-PL",
|
@@ -229,6 +265,12 @@ TASK_LIST_STS = [
|
|
229 |
"STSBenchmark",
|
230 |
]
|
231 |
|
|
|
|
|
|
|
|
|
|
|
|
|
232 |
TASK_LIST_STS_PL = [
|
233 |
"CDSC-R",
|
234 |
"SICK-R-PL",
|
@@ -247,11 +289,13 @@ TASK_LIST_STS_ZH = [
|
|
247 |
]
|
248 |
|
249 |
TASK_LIST_STS_OTHER = ["STS17 (ar-ar)", "STS17 (en-ar)", "STS17 (en-de)", "STS17 (en-tr)", "STS17 (es-en)", "STS17 (es-es)", "STS17 (fr-en)", "STS17 (it-en)", "STS17 (ko-ko)", "STS17 (nl-en)", "STS22 (ar)", "STS22 (de)", "STS22 (de-en)", "STS22 (de-fr)", "STS22 (de-pl)", "STS22 (es)", "STS22 (es-en)", "STS22 (es-it)", "STS22 (fr)", "STS22 (fr-pl)", "STS22 (it)", "STS22 (pl)", "STS22 (pl-en)", "STS22 (ru)", "STS22 (tr)", "STS22 (zh-en)", "STSBenchmark",]
|
250 |
-
TASK_LIST_STS_NORM = [x.replace(" (en)", "").replace(" (en-en)", "") for x in TASK_LIST_STS]
|
251 |
|
252 |
TASK_LIST_SUMMARIZATION = ["SummEval",]
|
253 |
|
|
|
|
|
254 |
TASK_LIST_EN = TASK_LIST_CLASSIFICATION + TASK_LIST_CLUSTERING + TASK_LIST_PAIR_CLASSIFICATION + TASK_LIST_RERANKING + TASK_LIST_RETRIEVAL + TASK_LIST_STS + TASK_LIST_SUMMARIZATION
|
|
|
255 |
TASK_LIST_PL = TASK_LIST_CLASSIFICATION_PL + TASK_LIST_CLUSTERING_PL + TASK_LIST_PAIR_CLASSIFICATION_PL + TASK_LIST_RETRIEVAL_PL + TASK_LIST_STS_PL
|
256 |
TASK_LIST_ZH = TASK_LIST_CLASSIFICATION_ZH + TASK_LIST_CLUSTERING_ZH + TASK_LIST_PAIR_CLASSIFICATION_ZH + TASK_LIST_RERANKING_ZH + TASK_LIST_RETRIEVAL_ZH + TASK_LIST_STS_ZH
|
257 |
|
@@ -276,11 +320,22 @@ def make_clickable_model(model_name, link=None):
|
|
276 |
|
277 |
# Models without metadata, thus we cannot fetch their results naturally
|
278 |
EXTERNAL_MODELS = [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
279 |
"all-MiniLM-L12-v2",
|
280 |
"all-MiniLM-L6-v2",
|
281 |
"all-mpnet-base-v2",
|
282 |
"allenai-specter",
|
283 |
-
"
|
|
|
|
|
|
|
|
|
284 |
"bert-base-swedish-cased",
|
285 |
"bert-base-uncased",
|
286 |
"bge-base-zh-v1.5",
|
@@ -292,18 +347,21 @@ EXTERNAL_MODELS = [
|
|
292 |
"dfm-encoder-large-v1",
|
293 |
"dfm-sentence-encoder-large-1",
|
294 |
"distiluse-base-multilingual-cased-v2",
|
295 |
-
"DanskBERT",
|
296 |
"e5-base",
|
297 |
"e5-large",
|
298 |
-
"e5-
|
|
|
299 |
"electra-small-nordic",
|
300 |
"electra-small-swedish-cased-discriminator",
|
|
|
|
|
|
|
301 |
"gbert-base",
|
302 |
"gbert-large",
|
303 |
"gelectra-base",
|
304 |
"gelectra-large",
|
305 |
-
"gottbert-base",
|
306 |
"glove.6B.300d",
|
|
|
307 |
"gtr-t5-base",
|
308 |
"gtr-t5-large",
|
309 |
"gtr-t5-xl",
|
@@ -311,11 +369,11 @@ EXTERNAL_MODELS = [
|
|
311 |
"herbert-base-retrieval-v2",
|
312 |
"komninos",
|
313 |
"luotuo-bert-medium",
|
314 |
-
"LASER2",
|
315 |
-
"LaBSE",
|
316 |
"m3e-base",
|
317 |
-
"m3e-large",
|
|
|
318 |
"msmarco-bert-co-condensor",
|
|
|
319 |
"multilingual-e5-base",
|
320 |
"multilingual-e5-large",
|
321 |
"multilingual-e5-small",
|
@@ -330,14 +388,19 @@ EXTERNAL_MODELS = [
|
|
330 |
"paraphrase-multilingual-MiniLM-L12-v2",
|
331 |
"paraphrase-multilingual-mpnet-base-v2",
|
332 |
"sentence-bert-swedish-cased",
|
|
|
|
|
|
|
333 |
"sentence-t5-base",
|
334 |
"sentence-t5-large",
|
335 |
"sentence-t5-xl",
|
336 |
"sentence-t5-xxl",
|
|
|
337 |
"sup-simcse-bert-base-uncased",
|
338 |
"st-polish-paraphrase-from-distilroberta",
|
339 |
-
"st-polish-paraphrase-from-mpnet",
|
340 |
"text2vec-base-chinese",
|
|
|
341 |
"text2vec-large-chinese",
|
342 |
"text-embedding-3-small",
|
343 |
"text-embedding-3-large",
|
@@ -353,38 +416,62 @@ EXTERNAL_MODELS = [
|
|
353 |
"text-search-curie-001",
|
354 |
"text-search-davinci-001",
|
355 |
"titan-embed-text-v1",
|
|
|
|
|
|
|
|
|
356 |
"unsup-simcse-bert-base-uncased",
|
357 |
"use-cmlm-multilingual",
|
|
|
|
|
358 |
"voyage-lite-01-instruct",
|
359 |
-
"voyage-lite-02-instruct",
|
360 |
"xlm-roberta-base",
|
361 |
-
"xlm-roberta-large",
|
362 |
]
|
363 |
|
364 |
EXTERNAL_MODEL_TO_LINK = {
|
|
|
|
|
365 |
"allenai-specter": "https://huggingface.co/sentence-transformers/allenai-specter",
|
366 |
"allenai-specter": "https://huggingface.co/sentence-transformers/allenai-specter",
|
367 |
"all-MiniLM-L12-v2": "https://huggingface.co/sentence-transformers/all-MiniLM-L12-v2",
|
368 |
"all-MiniLM-L6-v2": "https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2",
|
369 |
"all-mpnet-base-v2": "https://huggingface.co/sentence-transformers/all-mpnet-base-v2",
|
370 |
"Baichuan-text-embedding": "https://platform.baichuan-ai.com/docs/text-Embedding",
|
|
|
|
|
|
|
|
|
|
|
371 |
"bert-base-swedish-cased": "https://huggingface.co/KB/bert-base-swedish-cased",
|
372 |
"bert-base-uncased": "https://huggingface.co/bert-base-uncased",
|
373 |
"bge-base-zh-v1.5": "https://huggingface.co/BAAI/bge-base-zh-v1.5",
|
374 |
"bge-large-zh-v1.5": "https://huggingface.co/BAAI/bge-large-zh-v1.5",
|
375 |
"bge-large-zh-noinstruct": "https://huggingface.co/BAAI/bge-large-zh-noinstruct",
|
376 |
"bge-small-zh-v1.5": "https://huggingface.co/BAAI/bge-small-zh-v1.5",
|
|
|
|
|
377 |
"contriever-base-msmarco": "https://huggingface.co/nthakur/contriever-base-msmarco",
|
378 |
"cross-en-de-roberta-sentence-transformer": "https://huggingface.co/T-Systems-onsite/cross-en-de-roberta-sentence-transformer",
|
379 |
"DanskBERT": "https://huggingface.co/vesteinn/DanskBERT",
|
|
|
|
|
|
|
|
|
|
|
380 |
"distiluse-base-multilingual-cased-v2": "https://huggingface.co/sentence-transformers/distiluse-base-multilingual-cased-v2",
|
381 |
"dfm-encoder-large-v1": "https://huggingface.co/chcaa/dfm-encoder-large-v1",
|
382 |
"dfm-sentence-encoder-large-1": "https://huggingface.co/chcaa/dfm-encoder-large-v1",
|
383 |
"e5-base": "https://huggingface.co/intfloat/e5-base",
|
384 |
"e5-large": "https://huggingface.co/intfloat/e5-large",
|
|
|
385 |
"e5-small": "https://huggingface.co/intfloat/e5-small",
|
386 |
"electra-small-nordic": "https://huggingface.co/jonfd/electra-small-nordic",
|
387 |
"electra-small-swedish-cased-discriminator": "https://huggingface.co/KBLab/electra-small-swedish-cased-discriminator",
|
|
|
|
|
|
|
388 |
"gbert-base": "https://huggingface.co/deepset/gbert-base",
|
389 |
"gbert-large": "https://huggingface.co/deepset/gbert-large",
|
390 |
"gelectra-base": "https://huggingface.co/deepset/gelectra-base",
|
@@ -402,7 +489,9 @@ EXTERNAL_MODEL_TO_LINK = {
|
|
402 |
"LaBSE": "https://huggingface.co/sentence-transformers/LaBSE",
|
403 |
"m3e-base": "https://huggingface.co/moka-ai/m3e-base",
|
404 |
"m3e-large": "https://huggingface.co/moka-ai/m3e-large",
|
|
|
405 |
"msmarco-bert-co-condensor": "https://huggingface.co/sentence-transformers/msmarco-bert-co-condensor",
|
|
|
406 |
"multilingual-e5-base": "https://huggingface.co/intfloat/multilingual-e5-base",
|
407 |
"multilingual-e5-large": "https://huggingface.co/intfloat/multilingual-e5-large",
|
408 |
"multilingual-e5-small": "https://huggingface.co/intfloat/multilingual-e5-small",
|
@@ -414,13 +503,18 @@ EXTERNAL_MODEL_TO_LINK = {
|
|
414 |
"nomic-embed-text-v1.5-512": "https://huggingface.co/nomic-ai/nomic-embed-text-v1.5",
|
415 |
"norbert3-base": "https://huggingface.co/ltg/norbert3-base",
|
416 |
"norbert3-large": "https://huggingface.co/ltg/norbert3-large",
|
|
|
417 |
"paraphrase-multilingual-mpnet-base-v2": "https://huggingface.co/sentence-transformers/paraphrase-multilingual-mpnet-base-v2",
|
418 |
"paraphrase-multilingual-MiniLM-L12-v2": "https://huggingface.co/sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2",
|
|
|
|
|
|
|
419 |
"sentence-bert-swedish-cased": "https://huggingface.co/KBLab/sentence-bert-swedish-cased",
|
420 |
"sentence-t5-base": "https://huggingface.co/sentence-transformers/sentence-t5-base",
|
421 |
"sentence-t5-large": "https://huggingface.co/sentence-transformers/sentence-t5-large",
|
422 |
"sentence-t5-xl": "https://huggingface.co/sentence-transformers/sentence-t5-xl",
|
423 |
"sentence-t5-xxl": "https://huggingface.co/sentence-transformers/sentence-t5-xxl",
|
|
|
424 |
"sup-simcse-bert-base-uncased": "https://huggingface.co/princeton-nlp/sup-simcse-bert-base-uncased",
|
425 |
"st-polish-paraphrase-from-distilroberta": "https://huggingface.co/sdadas/st-polish-paraphrase-from-distilroberta",
|
426 |
"st-polish-paraphrase-from-mpnet": "https://huggingface.co/sdadas/st-polish-paraphrase-from-mpnet",
|
@@ -441,8 +535,14 @@ EXTERNAL_MODEL_TO_LINK = {
|
|
441 |
"text-search-babbage-001": "https://openai.com/blog/introducing-text-and-code-embeddings",
|
442 |
"text-search-davinci-001": "https://openai.com/blog/introducing-text-and-code-embeddings",
|
443 |
"titan-embed-text-v1": "https://docs.aws.amazon.com/bedrock/latest/userguide/embeddings.html",
|
|
|
|
|
|
|
|
|
444 |
"unsup-simcse-bert-base-uncased": "https://huggingface.co/princeton-nlp/unsup-simcse-bert-base-uncased",
|
445 |
"use-cmlm-multilingual": "https://huggingface.co/sentence-transformers/use-cmlm-multilingual",
|
|
|
|
|
446 |
"voyage-lite-01-instruct": "https://docs.voyageai.com/embeddings/",
|
447 |
"voyage-lite-02-instruct": "https://docs.voyageai.com/embeddings/",
|
448 |
"xlm-roberta-base": "https://huggingface.co/xlm-roberta-base",
|
@@ -450,28 +550,46 @@ EXTERNAL_MODEL_TO_LINK = {
|
|
450 |
}
|
451 |
|
452 |
EXTERNAL_MODEL_TO_DIM = {
|
|
|
|
|
453 |
"all-MiniLM-L12-v2": 384,
|
454 |
"all-MiniLM-L6-v2": 384,
|
455 |
"all-mpnet-base-v2": 768,
|
456 |
"allenai-specter": 768,
|
457 |
"Baichuan-text-embedding": 1024,
|
|
|
|
|
|
|
|
|
|
|
458 |
"bert-base-swedish-cased": 768,
|
459 |
"bert-base-uncased": 768,
|
460 |
"bge-base-zh-v1.5": 768,
|
461 |
"bge-large-zh-v1.5": 1024,
|
462 |
"bge-large-zh-noinstruct": 1024,
|
463 |
"bge-small-zh-v1.5": 512,
|
|
|
|
|
464 |
"contriever-base-msmarco": 768,
|
465 |
"cross-en-de-roberta-sentence-transformer": 768,
|
466 |
"DanskBERT": 768,
|
|
|
|
|
|
|
|
|
|
|
467 |
"distiluse-base-multilingual-cased-v2": 512,
|
468 |
"dfm-encoder-large-v1": 1024,
|
469 |
"dfm-sentence-encoder-large-1": 1024,
|
470 |
"e5-base": 768,
|
|
|
|
|
471 |
"e5-small": 384,
|
472 |
-
"e5-large": 1024,
|
473 |
"electra-small-nordic": 256,
|
474 |
"electra-small-swedish-cased-discriminator": 256,
|
|
|
|
|
|
|
475 |
"luotuo-bert-medium": 768,
|
476 |
"LASER2": 1024,
|
477 |
"LaBSE": 768,
|
@@ -489,7 +607,9 @@ EXTERNAL_MODEL_TO_DIM = {
|
|
489 |
"komninos": 300,
|
490 |
"m3e-base": 768,
|
491 |
"m3e-large": 768,
|
|
|
492 |
"msmarco-bert-co-condensor": 768,
|
|
|
493 |
"multilingual-e5-base": 768,
|
494 |
"multilingual-e5-small": 384,
|
495 |
"multilingual-e5-large": 1024,
|
@@ -501,13 +621,18 @@ EXTERNAL_MODEL_TO_DIM = {
|
|
501 |
"nomic-embed-text-v1.5-512": 512,
|
502 |
"norbert3-base": 768,
|
503 |
"norbert3-large": 1024,
|
|
|
504 |
"paraphrase-multilingual-MiniLM-L12-v2": 384,
|
505 |
"paraphrase-multilingual-mpnet-base-v2": 768,
|
|
|
|
|
|
|
506 |
"sentence-bert-swedish-cased": 768,
|
507 |
"sentence-t5-base": 768,
|
508 |
"sentence-t5-large": 768,
|
509 |
"sentence-t5-xl": 768,
|
510 |
"sentence-t5-xxl": 768,
|
|
|
511 |
"sup-simcse-bert-base-uncased": 768,
|
512 |
"st-polish-paraphrase-from-distilroberta": 768,
|
513 |
"st-polish-paraphrase-from-mpnet": 768,
|
@@ -528,8 +653,14 @@ EXTERNAL_MODEL_TO_DIM = {
|
|
528 |
"text-search-curie-001": 4096,
|
529 |
"text-search-davinci-001": 12288,
|
530 |
"titan-embed-text-v1": 1536,
|
|
|
|
|
|
|
|
|
531 |
"unsup-simcse-bert-base-uncased": 768,
|
532 |
"use-cmlm-multilingual": 768,
|
|
|
|
|
533 |
"voyage-lite-01-instruct": 1024,
|
534 |
"voyage-lite-02-instruct": 1024,
|
535 |
"xlm-roberta-base": 768,
|
@@ -537,28 +668,46 @@ EXTERNAL_MODEL_TO_DIM = {
|
|
537 |
}
|
538 |
|
539 |
EXTERNAL_MODEL_TO_SEQLEN = {
|
|
|
|
|
540 |
"all-MiniLM-L12-v2": 512,
|
541 |
"all-MiniLM-L6-v2": 512,
|
542 |
"all-mpnet-base-v2": 514,
|
543 |
"allenai-specter": 512,
|
544 |
"Baichuan-text-embedding": 512,
|
|
|
|
|
|
|
|
|
|
|
545 |
"bert-base-swedish-cased": 512,
|
546 |
"bert-base-uncased": 512,
|
547 |
"bge-base-zh-v1.5": 512,
|
548 |
"bge-large-zh-v1.5": 512,
|
549 |
"bge-large-zh-noinstruct": 512,
|
550 |
-
"bge-small-zh-v1.5": 512,
|
|
|
|
|
551 |
"contriever-base-msmarco": 512,
|
552 |
"cross-en-de-roberta-sentence-transformer": 514,
|
|
|
|
|
|
|
|
|
|
|
553 |
"DanskBERT": 514,
|
554 |
"dfm-encoder-large-v1": 512,
|
555 |
"dfm-sentence-encoder-large-1": 512,
|
556 |
"distiluse-base-multilingual-cased-v2": 512,
|
557 |
"e5-base": 512,
|
558 |
"e5-large": 512,
|
|
|
559 |
"e5-small": 512,
|
560 |
"electra-small-nordic": 512,
|
561 |
"electra-small-swedish-cased-discriminator": 512,
|
|
|
|
|
|
|
562 |
"gbert-base": 512,
|
563 |
"gbert-large": 512,
|
564 |
"gelectra-base": 512,
|
@@ -575,8 +724,10 @@ EXTERNAL_MODEL_TO_SEQLEN = {
|
|
575 |
"LASER2": "N/A",
|
576 |
"LaBSE": 512,
|
577 |
"m3e-base": 512,
|
578 |
-
"m3e-large": 512,
|
|
|
579 |
"msmarco-bert-co-condensor": 512,
|
|
|
580 |
"multilingual-e5-base": 514,
|
581 |
"multilingual-e5-large": 514,
|
582 |
"multilingual-e5-small": 512,
|
@@ -588,13 +739,18 @@ EXTERNAL_MODEL_TO_SEQLEN = {
|
|
588 |
"nomic-embed-text-v1.5-512": 8192,
|
589 |
"norbert3-base": 512,
|
590 |
"norbert3-large": 512,
|
|
|
591 |
"paraphrase-multilingual-MiniLM-L12-v2": 512,
|
592 |
"paraphrase-multilingual-mpnet-base-v2": 514,
|
|
|
|
|
|
|
593 |
"sentence-bert-swedish-cased": 512,
|
594 |
"sentence-t5-base": 512,
|
595 |
"sentence-t5-large": 512,
|
596 |
"sentence-t5-xl": 512,
|
597 |
"sentence-t5-xxl": 512,
|
|
|
598 |
"sup-simcse-bert-base-uncased": 512,
|
599 |
"st-polish-paraphrase-from-distilroberta": 514,
|
600 |
"st-polish-paraphrase-from-mpnet": 514,
|
@@ -615,8 +771,14 @@ EXTERNAL_MODEL_TO_SEQLEN = {
|
|
615 |
"text-search-curie-001": 2046,
|
616 |
"text-search-davinci-001": 2046,
|
617 |
"titan-embed-text-v1": 8000,
|
|
|
|
|
|
|
|
|
618 |
"use-cmlm-multilingual": 512,
|
619 |
"unsup-simcse-bert-base-uncased": 512,
|
|
|
|
|
620 |
"voyage-lite-01-instruct": 4000,
|
621 |
"voyage-lite-02-instruct": 4000,
|
622 |
"xlm-roberta-base": 514,
|
@@ -628,23 +790,39 @@ EXTERNAL_MODEL_TO_SIZE = {
|
|
628 |
"all-MiniLM-L12-v2": 0.13,
|
629 |
"all-MiniLM-L6-v2": 0.09,
|
630 |
"all-mpnet-base-v2": 0.44,
|
|
|
|
|
|
|
|
|
|
|
631 |
"bert-base-uncased": 0.44,
|
632 |
"bert-base-swedish-cased": 0.50,
|
633 |
"bge-base-zh-v1.5": 0.41,
|
634 |
"bge-large-zh-v1.5": 1.30,
|
635 |
"bge-large-zh-noinstruct": 1.30,
|
636 |
-
"bge-small-zh-v1.5": 0.10,
|
|
|
|
|
637 |
"cross-en-de-roberta-sentence-transformer": 1.11,
|
638 |
"contriever-base-msmarco": 0.44,
|
|
|
|
|
|
|
|
|
|
|
639 |
"DanskBERT": 0.50,
|
640 |
"distiluse-base-multilingual-cased-v2": 0.54,
|
641 |
"dfm-encoder-large-v1": 1.42,
|
642 |
"dfm-sentence-encoder-large-1": 1.63,
|
643 |
"e5-base": 0.44,
|
644 |
-
"e5-small": 0.13,
|
645 |
"e5-large": 1.34,
|
|
|
|
|
646 |
"electra-small-nordic": 0.09,
|
647 |
"electra-small-swedish-cased-discriminator": 0.06,
|
|
|
|
|
|
|
648 |
"gbert-base": 0.44,
|
649 |
"gbert-large": 1.35,
|
650 |
"gelectra-base": 0.44,
|
@@ -663,6 +841,7 @@ EXTERNAL_MODEL_TO_SIZE = {
|
|
663 |
"m3e-base": 0.41,
|
664 |
"m3e-large": 0.41,
|
665 |
"msmarco-bert-co-condensor": 0.44,
|
|
|
666 |
"multilingual-e5-base": 1.11,
|
667 |
"multilingual-e5-small": 0.47,
|
668 |
"multilingual-e5-large": 2.24,
|
@@ -676,11 +855,15 @@ EXTERNAL_MODEL_TO_SIZE = {
|
|
676 |
"norbert3-large": 1.47,
|
677 |
"paraphrase-multilingual-mpnet-base-v2": 1.11,
|
678 |
"paraphrase-multilingual-MiniLM-L12-v2": 0.47,
|
|
|
|
|
|
|
679 |
"sentence-bert-swedish-cased": 0.50,
|
680 |
"sentence-t5-base": 0.22,
|
681 |
"sentence-t5-large": 0.67,
|
682 |
"sentence-t5-xl": 2.48,
|
683 |
"sentence-t5-xxl": 9.73,
|
|
|
684 |
"sup-simcse-bert-base-uncased": 0.44,
|
685 |
"st-polish-paraphrase-from-distilroberta": 0.50,
|
686 |
"st-polish-paraphrase-from-mpnet": 0.50,
|
@@ -807,16 +990,9 @@ MODELS_TO_SKIP = {
|
|
807 |
"atian-chapters/Chapters-SFR-Embedding-Mistral", # Copy
|
808 |
"rlsChapters/Chapters-SFR-Embedding-Mistral", # Copy
|
809 |
"TitanML/jina-v2-base-en-embed", # Copy
|
810 |
-
"MaziyarPanahi/GritLM-8x7B-GGUF", # GGUF variant
|
811 |
}
|
812 |
|
813 |
-
|
814 |
-
if os.path.exists("EXTERNAL_MODEL_RESULTS.json"):
|
815 |
-
with open("EXTERNAL_MODEL_RESULTS.json") as f:
|
816 |
-
EXTERNAL_MODEL_RESULTS = json.load(f)
|
817 |
-
else:
|
818 |
-
EXTERNAL_MODEL_RESULTS = {model: {k: {v: []} for k, v in TASK_TO_METRIC.items()} for model in EXTERNAL_MODELS}
|
819 |
-
|
820 |
def add_lang(examples):
|
821 |
if not(examples["eval_language"]):
|
822 |
examples["mteb_dataset_name_with_lang"] = examples["mteb_dataset_name"]
|
@@ -824,48 +1000,62 @@ def add_lang(examples):
|
|
824 |
examples["mteb_dataset_name_with_lang"] = examples["mteb_dataset_name"] + f' ({examples["eval_language"]})'
|
825 |
return examples
|
826 |
|
|
|
|
|
827 |
def add_task(examples):
|
828 |
# Could be added to the dataset loading script instead
|
829 |
-
if examples["mteb_dataset_name"] in
|
830 |
examples["mteb_task"] = "Classification"
|
831 |
-
elif examples["mteb_dataset_name"] in TASK_LIST_CLUSTERING + TASK_LIST_CLUSTERING_DE + TASK_LIST_CLUSTERING_PL + TASK_LIST_CLUSTERING_ZH:
|
832 |
examples["mteb_task"] = "Clustering"
|
833 |
-
elif examples["mteb_dataset_name"] in TASK_LIST_PAIR_CLASSIFICATION + TASK_LIST_PAIR_CLASSIFICATION_PL + TASK_LIST_PAIR_CLASSIFICATION_ZH:
|
834 |
examples["mteb_task"] = "PairClassification"
|
835 |
-
elif examples["mteb_dataset_name"] in TASK_LIST_RERANKING + TASK_LIST_RERANKING_ZH:
|
836 |
examples["mteb_task"] = "Reranking"
|
837 |
-
elif examples["mteb_dataset_name"] in TASK_LIST_RETRIEVAL_NORM + TASK_LIST_RETRIEVAL_PL + TASK_LIST_RETRIEVAL_ZH:
|
838 |
examples["mteb_task"] = "Retrieval"
|
839 |
-
elif examples["mteb_dataset_name"] in
|
840 |
examples["mteb_task"] = "STS"
|
841 |
-
elif examples["mteb_dataset_name"] in TASK_LIST_SUMMARIZATION:
|
842 |
examples["mteb_task"] = "Summarization"
|
843 |
-
elif examples["mteb_dataset_name"] in
|
844 |
examples["mteb_task"] = "BitextMining"
|
845 |
else:
|
846 |
print("WARNING: Task not found for dataset", examples["mteb_dataset_name"])
|
847 |
examples["mteb_task"] = "Unknown"
|
848 |
return examples
|
849 |
|
850 |
-
if
|
851 |
-
|
852 |
-
|
853 |
-
|
854 |
-
|
855 |
-
|
856 |
-
|
857 |
-
|
858 |
-
|
859 |
-
|
860 |
-
|
861 |
-
|
862 |
-
|
863 |
-
|
864 |
-
|
865 |
-
|
866 |
-
|
867 |
-
|
868 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
869 |
|
870 |
def get_dim_seq_size(model):
|
871 |
filenames = [sib.rfilename for sib in model.siblings]
|
@@ -1136,6 +1326,68 @@ def get_mteb_average_zh():
|
|
1136 |
|
1137 |
return DATA_OVERALL_ZH
|
1138 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1139 |
def get_mteb_average_pl():
|
1140 |
global DATA_OVERALL_PL, DATA_CLASSIFICATION_PL, DATA_CLUSTERING_PL, DATA_PAIR_CLASSIFICATION_PL, DATA_RETRIEVAL_PL, DATA_STS_PL
|
1141 |
DATA_OVERALL_PL = get_mteb_data(
|
@@ -1191,6 +1443,7 @@ def get_mteb_average_pl():
|
|
1191 |
return DATA_OVERALL_PL
|
1192 |
|
1193 |
get_mteb_average()
|
|
|
1194 |
get_mteb_average_pl()
|
1195 |
get_mteb_average_zh()
|
1196 |
DATA_BITEXT_MINING = get_mteb_data(["BitextMining"], [], TASK_LIST_BITEXT_MINING)
|
@@ -1212,6 +1465,7 @@ for d in [
|
|
1212 |
DATA_BITEXT_MINING_OTHER,
|
1213 |
DATA_CLASSIFICATION_EN,
|
1214 |
DATA_CLASSIFICATION_DA,
|
|
|
1215 |
DATA_CLASSIFICATION_NB,
|
1216 |
DATA_CLASSIFICATION_PL,
|
1217 |
DATA_CLASSIFICATION_SV,
|
@@ -1219,21 +1473,27 @@ for d in [
|
|
1219 |
DATA_CLASSIFICATION_OTHER,
|
1220 |
DATA_CLUSTERING,
|
1221 |
DATA_CLUSTERING_DE,
|
|
|
1222 |
DATA_CLUSTERING_PL,
|
1223 |
DATA_CLUSTERING_ZH,
|
1224 |
DATA_PAIR_CLASSIFICATION,
|
|
|
1225 |
DATA_PAIR_CLASSIFICATION_PL,
|
1226 |
DATA_PAIR_CLASSIFICATION_ZH,
|
1227 |
DATA_RERANKING,
|
|
|
1228 |
DATA_RERANKING_ZH,
|
1229 |
DATA_RETRIEVAL,
|
|
|
1230 |
DATA_RETRIEVAL_PL,
|
1231 |
DATA_RETRIEVAL_ZH,
|
1232 |
DATA_STS_EN,
|
|
|
1233 |
DATA_STS_PL,
|
1234 |
DATA_STS_ZH,
|
1235 |
DATA_STS_OTHER,
|
1236 |
DATA_SUMMARIZATION,
|
|
|
1237 |
]:
|
1238 |
# NUM_SCORES += d.iloc[:, 1:].apply(lambda x: sum([1 for y in x if isinstance(y, float) and not np.isnan(y)]), axis=1).sum()
|
1239 |
cols_to_ignore = 3 if "Average" in d.columns else 2
|
@@ -1308,7 +1568,26 @@ with block:
|
|
1308 |
)
|
1309 |
with gr.Row():
|
1310 |
data_run_overall_zh = gr.Button("Refresh")
|
1311 |
-
data_run_overall_zh.click(get_mteb_average_zh, inputs=None, outputs=data_overall_zh)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1312 |
with gr.TabItem("Polish"):
|
1313 |
with gr.Row():
|
1314 |
gr.Markdown("""
|
@@ -1433,6 +1712,27 @@ with block:
|
|
1433 |
partial(get_mteb_data, tasks=["Classification"], datasets=TASK_LIST_CLASSIFICATION_DA),
|
1434 |
outputs=data_run_classification_da,
|
1435 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1436 |
with gr.TabItem("Norwegian"):
|
1437 |
with gr.Row():
|
1438 |
gr.Markdown("""
|
@@ -1558,6 +1858,27 @@ with block:
|
|
1558 |
partial(get_mteb_data, tasks=["Clustering"], datasets=TASK_LIST_CLUSTERING_ZH),
|
1559 |
outputs=data_clustering_zh,
|
1560 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1561 |
with gr.TabItem("German"):
|
1562 |
with gr.Row():
|
1563 |
gr.Markdown("""
|
@@ -1642,6 +1963,27 @@ with block:
|
|
1642 |
partial(get_mteb_data, tasks=["PairClassification"], datasets=TASK_LIST_PAIR_CLASSIFICATION_ZH),
|
1643 |
outputs=data_pair_classification_zh,
|
1644 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1645 |
with gr.TabItem("Polish"):
|
1646 |
with gr.Row():
|
1647 |
gr.Markdown("""
|
@@ -1705,6 +2047,27 @@ with block:
|
|
1705 |
partial(get_mteb_data, tasks=["Reranking"], datasets=TASK_LIST_RERANKING_ZH),
|
1706 |
outputs=data_reranking_zh,
|
1707 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1708 |
with gr.TabItem("Retrieval"):
|
1709 |
with gr.TabItem("English"):
|
1710 |
with gr.Row():
|
@@ -1737,18 +2100,40 @@ with block:
|
|
1737 |
- **Credits:** [FlagEmbedding](https://github.com/FlagOpen/FlagEmbedding)
|
1738 |
""")
|
1739 |
with gr.Row():
|
1740 |
-
|
1741 |
-
|
1742 |
# Add support for more columns than existing as a buffer for CQADupstack & other Retrieval tasks (e.g. MSMARCOv2)
|
1743 |
-
datatype=["number", "markdown"] + ["number"] * len(
|
1744 |
type="pandas",
|
1745 |
)
|
1746 |
with gr.Row():
|
1747 |
-
|
1748 |
-
|
1749 |
-
partial(get_mteb_data, tasks=["Retrieval"], datasets=
|
1750 |
-
outputs=
|
1751 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1752 |
with gr.TabItem("Polish"):
|
1753 |
with gr.Row():
|
1754 |
gr.Markdown("""
|
@@ -1813,6 +2198,27 @@ with block:
|
|
1813 |
partial(get_mteb_data, tasks=["STS"], datasets=TASK_LIST_STS_ZH),
|
1814 |
outputs=data_sts_zh,
|
1815 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1816 |
with gr.TabItem("Polish"):
|
1817 |
with gr.Row():
|
1818 |
gr.Markdown("""
|
|
|
38 |
"TweetSentimentExtractionClassification",
|
39 |
]
|
40 |
|
|
|
|
|
41 |
TASK_LIST_CLASSIFICATION_DA = [
|
42 |
"AngryTweetsClassification",
|
43 |
"DanishPoliticalCommentsClassification",
|
|
|
49 |
"ScalaDaClassification",
|
50 |
]
|
51 |
|
52 |
+
TASK_LIST_CLASSIFICATION_FR = [
|
53 |
+
"AmazonReviewsClassification (fr)",
|
54 |
+
"MasakhaNEWSClassification (fra)",
|
55 |
+
"MassiveIntentClassification (fr)",
|
56 |
+
"MassiveScenarioClassification (fr)",
|
57 |
+
"MTOPDomainClassification (fr)",
|
58 |
+
"MTOPIntentClassification (fr)",
|
59 |
+
]
|
60 |
+
|
61 |
TASK_LIST_CLASSIFICATION_NB = [
|
62 |
"NoRecClassification",
|
63 |
"NordicLangClassification",
|
|
|
122 |
"TenKGnadClusteringS2S",
|
123 |
]
|
124 |
|
125 |
+
TASK_LIST_CLUSTERING_FR = [
|
126 |
+
"AlloProfClusteringP2P",
|
127 |
+
"AlloProfClusteringS2S",
|
128 |
+
"HALClusteringS2S",
|
129 |
+
"MLSUMClusteringP2P",
|
130 |
+
"MLSUMClusteringS2S",
|
131 |
+
"MasakhaNEWSClusteringP2P (fra)",
|
132 |
+
"MasakhaNEWSClusteringS2S (fra)",
|
133 |
+
]
|
134 |
+
|
135 |
TASK_LIST_CLUSTERING_PL = [
|
136 |
"8TagsClustering",
|
137 |
]
|
|
|
149 |
"TwitterURLCorpus",
|
150 |
]
|
151 |
|
152 |
+
TASK_LIST_PAIR_CLASSIFICATION_FR = [
|
153 |
+
"OpusparcusPC (fr)",
|
154 |
+
"PawsX (fr)",
|
155 |
+
]
|
156 |
+
|
157 |
TASK_LIST_PAIR_CLASSIFICATION_PL = [
|
158 |
"CDSC-E",
|
159 |
"PPC",
|
|
|
173 |
"StackOverflowDupQuestions",
|
174 |
]
|
175 |
|
176 |
+
TASK_LIST_RERANKING_FR = [
|
177 |
+
"AlloprofReranking",
|
178 |
+
"SyntecReranking",
|
179 |
+
]
|
180 |
+
|
181 |
TASK_LIST_RERANKING_ZH = [
|
182 |
"CMedQAv1",
|
183 |
"CMedQAv2",
|
|
|
203 |
"TRECCOVID",
|
204 |
]
|
205 |
|
206 |
+
TASK_LIST_RETRIEVAL_FR = [
|
207 |
+
"AlloprofRetrieval",
|
208 |
+
"BSARDRetrieval",
|
209 |
+
"MintakaRetrieval (fr)",
|
210 |
+
# "MultiLongDocRetrieval",
|
211 |
+
"SyntecRetrieval",
|
212 |
+
"XPQARetrieval (fr)",
|
213 |
+
]
|
214 |
+
|
215 |
TASK_LIST_RETRIEVAL_PL = [
|
216 |
"ArguAna-PL",
|
217 |
"DBPedia-PL",
|
|
|
265 |
"STSBenchmark",
|
266 |
]
|
267 |
|
268 |
+
TASK_LIST_STS_FR = [
|
269 |
+
"STS22 (fr)",
|
270 |
+
"STSBenchmarkMultilingualSTS (fr)",
|
271 |
+
"SICKFr",
|
272 |
+
]
|
273 |
+
|
274 |
TASK_LIST_STS_PL = [
|
275 |
"CDSC-R",
|
276 |
"SICK-R-PL",
|
|
|
289 |
]
|
290 |
|
291 |
TASK_LIST_STS_OTHER = ["STS17 (ar-ar)", "STS17 (en-ar)", "STS17 (en-de)", "STS17 (en-tr)", "STS17 (es-en)", "STS17 (es-es)", "STS17 (fr-en)", "STS17 (it-en)", "STS17 (ko-ko)", "STS17 (nl-en)", "STS22 (ar)", "STS22 (de)", "STS22 (de-en)", "STS22 (de-fr)", "STS22 (de-pl)", "STS22 (es)", "STS22 (es-en)", "STS22 (es-it)", "STS22 (fr)", "STS22 (fr-pl)", "STS22 (it)", "STS22 (pl)", "STS22 (pl-en)", "STS22 (ru)", "STS22 (tr)", "STS22 (zh-en)", "STSBenchmark",]
|
|
|
292 |
|
293 |
TASK_LIST_SUMMARIZATION = ["SummEval",]
|
294 |
|
295 |
+
TASK_LIST_SUMMARIZATION_FR = ["SummEvalFr"]
|
296 |
+
|
297 |
TASK_LIST_EN = TASK_LIST_CLASSIFICATION + TASK_LIST_CLUSTERING + TASK_LIST_PAIR_CLASSIFICATION + TASK_LIST_RERANKING + TASK_LIST_RETRIEVAL + TASK_LIST_STS + TASK_LIST_SUMMARIZATION
|
298 |
+
TASK_LIST_FR = TASK_LIST_CLASSIFICATION_FR + TASK_LIST_CLUSTERING_FR + TASK_LIST_PAIR_CLASSIFICATION_FR + TASK_LIST_RERANKING_FR + TASK_LIST_RETRIEVAL_FR + TASK_LIST_STS_FR + TASK_LIST_SUMMARIZATION_FR
|
299 |
TASK_LIST_PL = TASK_LIST_CLASSIFICATION_PL + TASK_LIST_CLUSTERING_PL + TASK_LIST_PAIR_CLASSIFICATION_PL + TASK_LIST_RETRIEVAL_PL + TASK_LIST_STS_PL
|
300 |
TASK_LIST_ZH = TASK_LIST_CLASSIFICATION_ZH + TASK_LIST_CLUSTERING_ZH + TASK_LIST_PAIR_CLASSIFICATION_ZH + TASK_LIST_RERANKING_ZH + TASK_LIST_RETRIEVAL_ZH + TASK_LIST_STS_ZH
|
301 |
|
|
|
320 |
|
321 |
# Models without metadata, thus we cannot fetch their results naturally
|
322 |
EXTERNAL_MODELS = [
|
323 |
+
"Baichuan-text-embedding",
|
324 |
+
"Cohere-embed-multilingual-v3.0",
|
325 |
+
"Cohere-embed-multilingual-light-v3.0",
|
326 |
+
"DanskBERT",
|
327 |
+
"LASER2",
|
328 |
+
"LaBSE",
|
329 |
+
"OpenSearch-text-hybrid",
|
330 |
"all-MiniLM-L12-v2",
|
331 |
"all-MiniLM-L6-v2",
|
332 |
"all-mpnet-base-v2",
|
333 |
"allenai-specter",
|
334 |
+
"bert-base-10lang-cased",
|
335 |
+
"bert-base-15lang-cased",
|
336 |
+
"bert-base-25lang-cased",
|
337 |
+
"bert-base-multilingual-cased",
|
338 |
+
"bert-base-multilingual-uncased",
|
339 |
"bert-base-swedish-cased",
|
340 |
"bert-base-uncased",
|
341 |
"bge-base-zh-v1.5",
|
|
|
347 |
"dfm-encoder-large-v1",
|
348 |
"dfm-sentence-encoder-large-1",
|
349 |
"distiluse-base-multilingual-cased-v2",
|
|
|
350 |
"e5-base",
|
351 |
"e5-large",
|
352 |
+
"e5-mistral-7b-instruct",
|
353 |
+
"e5-small",
|
354 |
"electra-small-nordic",
|
355 |
"electra-small-swedish-cased-discriminator",
|
356 |
+
"flaubert_base_cased",
|
357 |
+
"flaubert_base_uncased",
|
358 |
+
"flaubert_large_cased",
|
359 |
"gbert-base",
|
360 |
"gbert-large",
|
361 |
"gelectra-base",
|
362 |
"gelectra-large",
|
|
|
363 |
"glove.6B.300d",
|
364 |
+
"gottbert-base",
|
365 |
"gtr-t5-base",
|
366 |
"gtr-t5-large",
|
367 |
"gtr-t5-xl",
|
|
|
369 |
"herbert-base-retrieval-v2",
|
370 |
"komninos",
|
371 |
"luotuo-bert-medium",
|
|
|
|
|
372 |
"m3e-base",
|
373 |
+
"m3e-large",
|
374 |
+
"mistral-embed",
|
375 |
"msmarco-bert-co-condensor",
|
376 |
+
"multi-qa-MiniLM-L6-cos-v1",
|
377 |
"multilingual-e5-base",
|
378 |
"multilingual-e5-large",
|
379 |
"multilingual-e5-small",
|
|
|
388 |
"paraphrase-multilingual-MiniLM-L12-v2",
|
389 |
"paraphrase-multilingual-mpnet-base-v2",
|
390 |
"sentence-bert-swedish-cased",
|
391 |
+
"sentence-camembert-base",
|
392 |
+
"sentence-camembert-large",
|
393 |
+
"sentence-croissant-llm-base",
|
394 |
"sentence-t5-base",
|
395 |
"sentence-t5-large",
|
396 |
"sentence-t5-xl",
|
397 |
"sentence-t5-xxl",
|
398 |
+
"silver-retriever-base-v1",
|
399 |
"sup-simcse-bert-base-uncased",
|
400 |
"st-polish-paraphrase-from-distilroberta",
|
401 |
+
"st-polish-paraphrase-from-mpnet",
|
402 |
"text2vec-base-chinese",
|
403 |
+
"text2vec-base-multilingual",
|
404 |
"text2vec-large-chinese",
|
405 |
"text-embedding-3-small",
|
406 |
"text-embedding-3-large",
|
|
|
416 |
"text-search-curie-001",
|
417 |
"text-search-davinci-001",
|
418 |
"titan-embed-text-v1",
|
419 |
+
"udever-bloom-1b1",
|
420 |
+
"udever-bloom-560m",
|
421 |
+
"universal-sentence-encoder-multilingual-3",
|
422 |
+
"universal-sentence-encoder-multilingual-large-3",
|
423 |
"unsup-simcse-bert-base-uncased",
|
424 |
"use-cmlm-multilingual",
|
425 |
+
"voyage-2",
|
426 |
+
"voyage-code-2",
|
427 |
"voyage-lite-01-instruct",
|
428 |
+
"voyage-lite-02-instruct",
|
429 |
"xlm-roberta-base",
|
430 |
+
"xlm-roberta-large",
|
431 |
]
|
432 |
|
433 |
EXTERNAL_MODEL_TO_LINK = {
|
434 |
+
"Cohere-embed-multilingual-v3.0": "https://huggingface.co/Cohere/Cohere-embed-multilingual-v3.0",
|
435 |
+
"Cohere-embed-multilingual-light-v3.0": "https://huggingface.co/Cohere/Cohere-embed-multilingual-light-v3.0",
|
436 |
"allenai-specter": "https://huggingface.co/sentence-transformers/allenai-specter",
|
437 |
"allenai-specter": "https://huggingface.co/sentence-transformers/allenai-specter",
|
438 |
"all-MiniLM-L12-v2": "https://huggingface.co/sentence-transformers/all-MiniLM-L12-v2",
|
439 |
"all-MiniLM-L6-v2": "https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2",
|
440 |
"all-mpnet-base-v2": "https://huggingface.co/sentence-transformers/all-mpnet-base-v2",
|
441 |
"Baichuan-text-embedding": "https://platform.baichuan-ai.com/docs/text-Embedding",
|
442 |
+
"bert-base-10lang-cased": "https://huggingface.co/Geotrend/bert-base-10lang-cased",
|
443 |
+
"bert-base-15lang-cased": "https://huggingface.co/Geotrend/bert-base-15lang-cased",
|
444 |
+
"bert-base-25lang-cased": "https://huggingface.co/Geotrend/bert-base-25lang-cased",
|
445 |
+
"bert-base-multilingual-cased": "https://huggingface.co/google-bert/bert-base-multilingual-cased",
|
446 |
+
"bert-base-multilingual-uncased": "https://huggingface.co/google-bert/bert-base-multilingual-uncased",
|
447 |
"bert-base-swedish-cased": "https://huggingface.co/KB/bert-base-swedish-cased",
|
448 |
"bert-base-uncased": "https://huggingface.co/bert-base-uncased",
|
449 |
"bge-base-zh-v1.5": "https://huggingface.co/BAAI/bge-base-zh-v1.5",
|
450 |
"bge-large-zh-v1.5": "https://huggingface.co/BAAI/bge-large-zh-v1.5",
|
451 |
"bge-large-zh-noinstruct": "https://huggingface.co/BAAI/bge-large-zh-noinstruct",
|
452 |
"bge-small-zh-v1.5": "https://huggingface.co/BAAI/bge-small-zh-v1.5",
|
453 |
+
"camembert-base": "https://huggingface.co/almanach/camembert-base",
|
454 |
+
"camembert-large": "https://huggingface.co/almanach/camembert-large",
|
455 |
"contriever-base-msmarco": "https://huggingface.co/nthakur/contriever-base-msmarco",
|
456 |
"cross-en-de-roberta-sentence-transformer": "https://huggingface.co/T-Systems-onsite/cross-en-de-roberta-sentence-transformer",
|
457 |
"DanskBERT": "https://huggingface.co/vesteinn/DanskBERT",
|
458 |
+
"distilbert-base-25lang-cased": "https://huggingface.co/Geotrend/distilbert-base-25lang-cased",
|
459 |
+
"distilbert-base-en-fr-cased": "https://huggingface.co/Geotrend/distilbert-base-en-fr-cased",
|
460 |
+
"distilbert-base-en-fr-es-pt-it-cased": "https://huggingface.co/Geotrend/distilbert-base-en-fr-es-pt-it-cased",
|
461 |
+
"distilbert-base-fr-cased": "https://huggingface.co/Geotrend/distilbert-base-fr-cased",
|
462 |
+
"distilbert-base-uncased": "https://huggingface.co/distilbert-base-uncased",
|
463 |
"distiluse-base-multilingual-cased-v2": "https://huggingface.co/sentence-transformers/distiluse-base-multilingual-cased-v2",
|
464 |
"dfm-encoder-large-v1": "https://huggingface.co/chcaa/dfm-encoder-large-v1",
|
465 |
"dfm-sentence-encoder-large-1": "https://huggingface.co/chcaa/dfm-encoder-large-v1",
|
466 |
"e5-base": "https://huggingface.co/intfloat/e5-base",
|
467 |
"e5-large": "https://huggingface.co/intfloat/e5-large",
|
468 |
+
"e5-mistral-7b-instruct": "https://huggingface.co/intfloat/e5-mistral-7b-instruct",
|
469 |
"e5-small": "https://huggingface.co/intfloat/e5-small",
|
470 |
"electra-small-nordic": "https://huggingface.co/jonfd/electra-small-nordic",
|
471 |
"electra-small-swedish-cased-discriminator": "https://huggingface.co/KBLab/electra-small-swedish-cased-discriminator",
|
472 |
+
"flaubert_base_cased": "https://huggingface.co/flaubert/flaubert_base_cased",
|
473 |
+
"flaubert_base_uncased": "https://huggingface.co/flaubert/flaubert_base_uncased",
|
474 |
+
"flaubert_large_cased": "https://huggingface.co/flaubert/flaubert_large_cased",
|
475 |
"gbert-base": "https://huggingface.co/deepset/gbert-base",
|
476 |
"gbert-large": "https://huggingface.co/deepset/gbert-large",
|
477 |
"gelectra-base": "https://huggingface.co/deepset/gelectra-base",
|
|
|
489 |
"LaBSE": "https://huggingface.co/sentence-transformers/LaBSE",
|
490 |
"m3e-base": "https://huggingface.co/moka-ai/m3e-base",
|
491 |
"m3e-large": "https://huggingface.co/moka-ai/m3e-large",
|
492 |
+
"mistral-embed": "https://docs.mistral.ai/guides/embeddings",
|
493 |
"msmarco-bert-co-condensor": "https://huggingface.co/sentence-transformers/msmarco-bert-co-condensor",
|
494 |
+
"multi-qa-MiniLM-L6-cos-v1": "https://huggingface.co/sentence-transformers/multi-qa-MiniLM-L6-cos-v1",
|
495 |
"multilingual-e5-base": "https://huggingface.co/intfloat/multilingual-e5-base",
|
496 |
"multilingual-e5-large": "https://huggingface.co/intfloat/multilingual-e5-large",
|
497 |
"multilingual-e5-small": "https://huggingface.co/intfloat/multilingual-e5-small",
|
|
|
503 |
"nomic-embed-text-v1.5-512": "https://huggingface.co/nomic-ai/nomic-embed-text-v1.5",
|
504 |
"norbert3-base": "https://huggingface.co/ltg/norbert3-base",
|
505 |
"norbert3-large": "https://huggingface.co/ltg/norbert3-large",
|
506 |
+
"OpenSearch-text-hybrid": "https://help.aliyun.com/zh/open-search/vector-search-edition/hybrid-retrieval",
|
507 |
"paraphrase-multilingual-mpnet-base-v2": "https://huggingface.co/sentence-transformers/paraphrase-multilingual-mpnet-base-v2",
|
508 |
"paraphrase-multilingual-MiniLM-L12-v2": "https://huggingface.co/sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2",
|
509 |
+
"sentence-camembert-base": "https://huggingface.co/dangvantuan/sentence-camembert-base",
|
510 |
+
"sentence-camembert-large": "https://huggingface.co/dangvantuan/sentence-camembert-large",
|
511 |
+
"sentence-croissant-llm-base": "https://huggingface.co/Wissam42/sentence-croissant-llm-base",
|
512 |
"sentence-bert-swedish-cased": "https://huggingface.co/KBLab/sentence-bert-swedish-cased",
|
513 |
"sentence-t5-base": "https://huggingface.co/sentence-transformers/sentence-t5-base",
|
514 |
"sentence-t5-large": "https://huggingface.co/sentence-transformers/sentence-t5-large",
|
515 |
"sentence-t5-xl": "https://huggingface.co/sentence-transformers/sentence-t5-xl",
|
516 |
"sentence-t5-xxl": "https://huggingface.co/sentence-transformers/sentence-t5-xxl",
|
517 |
+
"silver-retriever-base-v1": "https://huggingface.co/ipipan/silver-retriever-base-v1",
|
518 |
"sup-simcse-bert-base-uncased": "https://huggingface.co/princeton-nlp/sup-simcse-bert-base-uncased",
|
519 |
"st-polish-paraphrase-from-distilroberta": "https://huggingface.co/sdadas/st-polish-paraphrase-from-distilroberta",
|
520 |
"st-polish-paraphrase-from-mpnet": "https://huggingface.co/sdadas/st-polish-paraphrase-from-mpnet",
|
|
|
535 |
"text-search-babbage-001": "https://openai.com/blog/introducing-text-and-code-embeddings",
|
536 |
"text-search-davinci-001": "https://openai.com/blog/introducing-text-and-code-embeddings",
|
537 |
"titan-embed-text-v1": "https://docs.aws.amazon.com/bedrock/latest/userguide/embeddings.html",
|
538 |
+
"udever-bloom-1b1": "https://huggingface.co/izhx/udever-bloom-1b1",
|
539 |
+
"udever-bloom-560m": "https://huggingface.co/izhx/udever-bloom-560m",
|
540 |
+
"universal-sentence-encoder-multilingual-3": "https://huggingface.co/vprelovac/universal-sentence-encoder-multilingual-3",
|
541 |
+
"universal-sentence-encoder-multilingual-large-3": "https://huggingface.co/vprelovac/universal-sentence-encoder-multilingual-large-3",
|
542 |
"unsup-simcse-bert-base-uncased": "https://huggingface.co/princeton-nlp/unsup-simcse-bert-base-uncased",
|
543 |
"use-cmlm-multilingual": "https://huggingface.co/sentence-transformers/use-cmlm-multilingual",
|
544 |
+
"voyage-2": "https://docs.voyageai.com/embeddings/",
|
545 |
+
"voyage-code-2": "https://docs.voyageai.com/embeddings/",
|
546 |
"voyage-lite-01-instruct": "https://docs.voyageai.com/embeddings/",
|
547 |
"voyage-lite-02-instruct": "https://docs.voyageai.com/embeddings/",
|
548 |
"xlm-roberta-base": "https://huggingface.co/xlm-roberta-base",
|
|
|
550 |
}
|
551 |
|
552 |
EXTERNAL_MODEL_TO_DIM = {
|
553 |
+
"Cohere-embed-multilingual-v3.0": 1024,
|
554 |
+
"Cohere-embed-multilingual-light-v3.0": 384,
|
555 |
"all-MiniLM-L12-v2": 384,
|
556 |
"all-MiniLM-L6-v2": 384,
|
557 |
"all-mpnet-base-v2": 768,
|
558 |
"allenai-specter": 768,
|
559 |
"Baichuan-text-embedding": 1024,
|
560 |
+
"bert-base-10lang-cased": 768,
|
561 |
+
"bert-base-15lang-cased": 768,
|
562 |
+
"bert-base-25lang-cased": 768,
|
563 |
+
"bert-base-multilingual-cased": 768,
|
564 |
+
"bert-base-multilingual-uncased": 768,
|
565 |
"bert-base-swedish-cased": 768,
|
566 |
"bert-base-uncased": 768,
|
567 |
"bge-base-zh-v1.5": 768,
|
568 |
"bge-large-zh-v1.5": 1024,
|
569 |
"bge-large-zh-noinstruct": 1024,
|
570 |
"bge-small-zh-v1.5": 512,
|
571 |
+
"camembert-base": 512,
|
572 |
+
"camembert-large": 768,
|
573 |
"contriever-base-msmarco": 768,
|
574 |
"cross-en-de-roberta-sentence-transformer": 768,
|
575 |
"DanskBERT": 768,
|
576 |
+
"distilbert-base-25lang-cased": 768,
|
577 |
+
"distilbert-base-en-fr-cased": 768,
|
578 |
+
"distilbert-base-en-fr-es-pt-it-cased": 768,
|
579 |
+
"distilbert-base-fr-cased": 768,
|
580 |
+
"distilbert-base-uncased": 768,
|
581 |
"distiluse-base-multilingual-cased-v2": 512,
|
582 |
"dfm-encoder-large-v1": 1024,
|
583 |
"dfm-sentence-encoder-large-1": 1024,
|
584 |
"e5-base": 768,
|
585 |
+
"e5-large": 1024,
|
586 |
+
"e5-mistral-7b-instruct": 4096,
|
587 |
"e5-small": 384,
|
|
|
588 |
"electra-small-nordic": 256,
|
589 |
"electra-small-swedish-cased-discriminator": 256,
|
590 |
+
"flaubert_base_cased": 768,
|
591 |
+
"flaubert_base_uncased": 768,
|
592 |
+
"flaubert_large_cased": 1024,
|
593 |
"luotuo-bert-medium": 768,
|
594 |
"LASER2": 1024,
|
595 |
"LaBSE": 768,
|
|
|
607 |
"komninos": 300,
|
608 |
"m3e-base": 768,
|
609 |
"m3e-large": 768,
|
610 |
+
"mistral-embed": 1024,
|
611 |
"msmarco-bert-co-condensor": 768,
|
612 |
+
"multi-qa-MiniLM-L6-cos-v1": 384,
|
613 |
"multilingual-e5-base": 768,
|
614 |
"multilingual-e5-small": 384,
|
615 |
"multilingual-e5-large": 1024,
|
|
|
621 |
"nomic-embed-text-v1.5-512": 512,
|
622 |
"norbert3-base": 768,
|
623 |
"norbert3-large": 1024,
|
624 |
+
"OpenSearch-text-hybrid": 1792,
|
625 |
"paraphrase-multilingual-MiniLM-L12-v2": 384,
|
626 |
"paraphrase-multilingual-mpnet-base-v2": 768,
|
627 |
+
"sentence-camembert-base": 768,
|
628 |
+
"sentence-camembert-large": 1024,
|
629 |
+
"sentence-croissant-llm-base": 2048,
|
630 |
"sentence-bert-swedish-cased": 768,
|
631 |
"sentence-t5-base": 768,
|
632 |
"sentence-t5-large": 768,
|
633 |
"sentence-t5-xl": 768,
|
634 |
"sentence-t5-xxl": 768,
|
635 |
+
"silver-retriever-base-v1": 768,
|
636 |
"sup-simcse-bert-base-uncased": 768,
|
637 |
"st-polish-paraphrase-from-distilroberta": 768,
|
638 |
"st-polish-paraphrase-from-mpnet": 768,
|
|
|
653 |
"text-search-curie-001": 4096,
|
654 |
"text-search-davinci-001": 12288,
|
655 |
"titan-embed-text-v1": 1536,
|
656 |
+
"udever-bloom-1b1": 1536,
|
657 |
+
"udever-bloom-560m": 1024,
|
658 |
+
"universal-sentence-encoder-multilingual-3": 512,
|
659 |
+
"universal-sentence-encoder-multilingual-large-3": 512,
|
660 |
"unsup-simcse-bert-base-uncased": 768,
|
661 |
"use-cmlm-multilingual": 768,
|
662 |
+
"voyage-2": 1024,
|
663 |
+
"voyage-code-2": 1536,
|
664 |
"voyage-lite-01-instruct": 1024,
|
665 |
"voyage-lite-02-instruct": 1024,
|
666 |
"xlm-roberta-base": 768,
|
|
|
668 |
}
|
669 |
|
670 |
EXTERNAL_MODEL_TO_SEQLEN = {
|
671 |
+
"Cohere-embed-multilingual-v3.0": 512,
|
672 |
+
"Cohere-embed-multilingual-light-v3.0": 512,
|
673 |
"all-MiniLM-L12-v2": 512,
|
674 |
"all-MiniLM-L6-v2": 512,
|
675 |
"all-mpnet-base-v2": 514,
|
676 |
"allenai-specter": 512,
|
677 |
"Baichuan-text-embedding": 512,
|
678 |
+
"bert-base-10lang-cased": 512,
|
679 |
+
"bert-base-15lang-cased": 512,
|
680 |
+
"bert-base-25lang-cased": 512,
|
681 |
+
"bert-base-multilingual-cased": 512,
|
682 |
+
"bert-base-multilingual-uncased": 512,
|
683 |
"bert-base-swedish-cased": 512,
|
684 |
"bert-base-uncased": 512,
|
685 |
"bge-base-zh-v1.5": 512,
|
686 |
"bge-large-zh-v1.5": 512,
|
687 |
"bge-large-zh-noinstruct": 512,
|
688 |
+
"bge-small-zh-v1.5": 512,
|
689 |
+
"camembert-base": 512,
|
690 |
+
"camembert-large": 512,
|
691 |
"contriever-base-msmarco": 512,
|
692 |
"cross-en-de-roberta-sentence-transformer": 514,
|
693 |
+
"distilbert-base-25lang-cased": 512,
|
694 |
+
"distilbert-base-en-fr-cased": 512,
|
695 |
+
"distilbert-base-en-fr-es-pt-it-cased": 512,
|
696 |
+
"distilbert-base-fr-cased": 512,
|
697 |
+
"distilbert-base-uncased": 512,
|
698 |
"DanskBERT": 514,
|
699 |
"dfm-encoder-large-v1": 512,
|
700 |
"dfm-sentence-encoder-large-1": 512,
|
701 |
"distiluse-base-multilingual-cased-v2": 512,
|
702 |
"e5-base": 512,
|
703 |
"e5-large": 512,
|
704 |
+
"e5-mistral-7b-instruct": 32768,
|
705 |
"e5-small": 512,
|
706 |
"electra-small-nordic": 512,
|
707 |
"electra-small-swedish-cased-discriminator": 512,
|
708 |
+
"flaubert_base_cased": 512,
|
709 |
+
"flaubert_base_uncased": 512,
|
710 |
+
"flaubert_large_cased": 512,
|
711 |
"gbert-base": 512,
|
712 |
"gbert-large": 512,
|
713 |
"gelectra-base": 512,
|
|
|
724 |
"LASER2": "N/A",
|
725 |
"LaBSE": 512,
|
726 |
"m3e-base": 512,
|
727 |
+
"m3e-large": 512,
|
728 |
+
# "mistral-embed": "?",
|
729 |
"msmarco-bert-co-condensor": 512,
|
730 |
+
"multi-qa-MiniLM-L6-cos-v1": 512,
|
731 |
"multilingual-e5-base": 514,
|
732 |
"multilingual-e5-large": 514,
|
733 |
"multilingual-e5-small": 512,
|
|
|
739 |
"nomic-embed-text-v1.5-512": 8192,
|
740 |
"norbert3-base": 512,
|
741 |
"norbert3-large": 512,
|
742 |
+
"OpenSearch-text-hybrid": 512,
|
743 |
"paraphrase-multilingual-MiniLM-L12-v2": 512,
|
744 |
"paraphrase-multilingual-mpnet-base-v2": 514,
|
745 |
+
"sentence-camembert-base": 512,
|
746 |
+
"sentence-camembert-large": 512,
|
747 |
+
"sentence-croissant-llm-base": 2048,
|
748 |
"sentence-bert-swedish-cased": 512,
|
749 |
"sentence-t5-base": 512,
|
750 |
"sentence-t5-large": 512,
|
751 |
"sentence-t5-xl": 512,
|
752 |
"sentence-t5-xxl": 512,
|
753 |
+
"silver-retriever-base-v1": 514,
|
754 |
"sup-simcse-bert-base-uncased": 512,
|
755 |
"st-polish-paraphrase-from-distilroberta": 514,
|
756 |
"st-polish-paraphrase-from-mpnet": 514,
|
|
|
771 |
"text-search-curie-001": 2046,
|
772 |
"text-search-davinci-001": 2046,
|
773 |
"titan-embed-text-v1": 8000,
|
774 |
+
"udever-bloom-1b1": 2048,
|
775 |
+
"udever-bloom-560m": 2048,
|
776 |
+
"universal-sentence-encoder-multilingual-3": 512,
|
777 |
+
"universal-sentence-encoder-multilingual-large-3": 512,
|
778 |
"use-cmlm-multilingual": 512,
|
779 |
"unsup-simcse-bert-base-uncased": 512,
|
780 |
+
"voyage-2": 1024,
|
781 |
+
"voyage-code-2": 16000,
|
782 |
"voyage-lite-01-instruct": 4000,
|
783 |
"voyage-lite-02-instruct": 4000,
|
784 |
"xlm-roberta-base": 514,
|
|
|
790 |
"all-MiniLM-L12-v2": 0.13,
|
791 |
"all-MiniLM-L6-v2": 0.09,
|
792 |
"all-mpnet-base-v2": 0.44,
|
793 |
+
"bert-base-10lang-cased": 0.61,
|
794 |
+
"bert-base-15lang-cased": 0.61,
|
795 |
+
"bert-base-25lang-cased": 0.61,
|
796 |
+
"bert-base-multilingual-cased": 0.71,
|
797 |
+
"bert-base-multilingual-uncased": 0.67,
|
798 |
"bert-base-uncased": 0.44,
|
799 |
"bert-base-swedish-cased": 0.50,
|
800 |
"bge-base-zh-v1.5": 0.41,
|
801 |
"bge-large-zh-v1.5": 1.30,
|
802 |
"bge-large-zh-noinstruct": 1.30,
|
803 |
+
"bge-small-zh-v1.5": 0.10,
|
804 |
+
"camembert-base": 0.45,
|
805 |
+
"camembert-large": 1.35,
|
806 |
"cross-en-de-roberta-sentence-transformer": 1.11,
|
807 |
"contriever-base-msmarco": 0.44,
|
808 |
+
"distilbert-base-25lang-cased": 0.44,
|
809 |
+
"distilbert-base-en-fr-cased": 0.44,
|
810 |
+
"distilbert-base-en-fr-es-pt-it-cased": 0.44,
|
811 |
+
"distilbert-base-fr-cased": 0.44,
|
812 |
+
"distilbert-base-uncased": 0.44,
|
813 |
"DanskBERT": 0.50,
|
814 |
"distiluse-base-multilingual-cased-v2": 0.54,
|
815 |
"dfm-encoder-large-v1": 1.42,
|
816 |
"dfm-sentence-encoder-large-1": 1.63,
|
817 |
"e5-base": 0.44,
|
|
|
818 |
"e5-large": 1.34,
|
819 |
+
"e5-mistral-7b-instruct": 14.22,
|
820 |
+
"e5-small": 0.13,
|
821 |
"electra-small-nordic": 0.09,
|
822 |
"electra-small-swedish-cased-discriminator": 0.06,
|
823 |
+
"flaubert_base_cased": 0.55,
|
824 |
+
"flaubert_base_uncased": 0.55,
|
825 |
+
"flaubert_large_cased": 1.49,
|
826 |
"gbert-base": 0.44,
|
827 |
"gbert-large": 1.35,
|
828 |
"gelectra-base": 0.44,
|
|
|
841 |
"m3e-base": 0.41,
|
842 |
"m3e-large": 0.41,
|
843 |
"msmarco-bert-co-condensor": 0.44,
|
844 |
+
"multi-qa-MiniLM-L6-cos-v1": 0.09,
|
845 |
"multilingual-e5-base": 1.11,
|
846 |
"multilingual-e5-small": 0.47,
|
847 |
"multilingual-e5-large": 2.24,
|
|
|
855 |
"norbert3-large": 1.47,
|
856 |
"paraphrase-multilingual-mpnet-base-v2": 1.11,
|
857 |
"paraphrase-multilingual-MiniLM-L12-v2": 0.47,
|
858 |
+
"sentence-camembert-base": 0.44,
|
859 |
+
"sentence-camembert-large": 1.35,
|
860 |
+
"sentence-croissant-llm-base": 5.12,
|
861 |
"sentence-bert-swedish-cased": 0.50,
|
862 |
"sentence-t5-base": 0.22,
|
863 |
"sentence-t5-large": 0.67,
|
864 |
"sentence-t5-xl": 2.48,
|
865 |
"sentence-t5-xxl": 9.73,
|
866 |
+
"silver-retriever-base-v1": 0.50,
|
867 |
"sup-simcse-bert-base-uncased": 0.44,
|
868 |
"st-polish-paraphrase-from-distilroberta": 0.50,
|
869 |
"st-polish-paraphrase-from-mpnet": 0.50,
|
|
|
990 |
"atian-chapters/Chapters-SFR-Embedding-Mistral", # Copy
|
991 |
"rlsChapters/Chapters-SFR-Embedding-Mistral", # Copy
|
992 |
"TitanML/jina-v2-base-en-embed", # Copy
|
993 |
+
"MaziyarPanahi/GritLM-8x7B-GGUF", # GGUF variant
|
994 |
}
|
995 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
996 |
def add_lang(examples):
|
997 |
if not(examples["eval_language"]):
|
998 |
examples["mteb_dataset_name_with_lang"] = examples["mteb_dataset_name"]
|
|
|
1000 |
examples["mteb_dataset_name_with_lang"] = examples["mteb_dataset_name"] + f' ({examples["eval_language"]})'
|
1001 |
return examples
|
1002 |
|
1003 |
+
def norm(names): return set([name.split(" ")[0] for name in names])
|
1004 |
+
|
1005 |
def add_task(examples):
|
1006 |
# Could be added to the dataset loading script instead
|
1007 |
+
if examples["mteb_dataset_name"] in norm(TASK_LIST_CLASSIFICATION + TASK_LIST_CLASSIFICATION_DA + TASK_LIST_CLASSIFICATION_FR + TASK_LIST_CLASSIFICATION_NB + TASK_LIST_CLASSIFICATION_PL + TASK_LIST_CLASSIFICATION_SV + TASK_LIST_CLASSIFICATION_ZH):
|
1008 |
examples["mteb_task"] = "Classification"
|
1009 |
+
elif examples["mteb_dataset_name"] in norm(TASK_LIST_CLUSTERING + TASK_LIST_CLUSTERING_DE + TASK_LIST_CLUSTERING_FR + TASK_LIST_CLUSTERING_PL + TASK_LIST_CLUSTERING_ZH):
|
1010 |
examples["mteb_task"] = "Clustering"
|
1011 |
+
elif examples["mteb_dataset_name"] in norm(TASK_LIST_PAIR_CLASSIFICATION + TASK_LIST_PAIR_CLASSIFICATION_FR + TASK_LIST_PAIR_CLASSIFICATION_PL + TASK_LIST_PAIR_CLASSIFICATION_ZH):
|
1012 |
examples["mteb_task"] = "PairClassification"
|
1013 |
+
elif examples["mteb_dataset_name"] in norm(TASK_LIST_RERANKING + TASK_LIST_RERANKING_FR + TASK_LIST_RERANKING_ZH):
|
1014 |
examples["mteb_task"] = "Reranking"
|
1015 |
+
elif examples["mteb_dataset_name"] in norm(TASK_LIST_RETRIEVAL_NORM + TASK_LIST_RETRIEVAL_FR + TASK_LIST_RETRIEVAL_PL + TASK_LIST_RETRIEVAL_ZH):
|
1016 |
examples["mteb_task"] = "Retrieval"
|
1017 |
+
elif examples["mteb_dataset_name"] in norm(TASK_LIST_STS + TASK_LIST_STS_FR + TASK_LIST_STS_PL + TASK_LIST_STS_ZH):
|
1018 |
examples["mteb_task"] = "STS"
|
1019 |
+
elif examples["mteb_dataset_name"] in norm(TASK_LIST_SUMMARIZATION + TASK_LIST_SUMMARIZATION_FR):
|
1020 |
examples["mteb_task"] = "Summarization"
|
1021 |
+
elif examples["mteb_dataset_name"] in norm(TASK_LIST_BITEXT_MINING + TASK_LIST_BITEXT_MINING_OTHER):
|
1022 |
examples["mteb_task"] = "BitextMining"
|
1023 |
else:
|
1024 |
print("WARNING: Task not found for dataset", examples["mteb_dataset_name"])
|
1025 |
examples["mteb_task"] = "Unknown"
|
1026 |
return examples
|
1027 |
|
1028 |
+
if os.path.exists("EXTERNAL_MODEL_RESULTS.json"):
|
1029 |
+
with open("EXTERNAL_MODEL_RESULTS.json") as f:
|
1030 |
+
EXTERNAL_MODEL_RESULTS = json.load(f)
|
1031 |
+
# Update with models not contained
|
1032 |
+
models_to_run = []
|
1033 |
+
for model in EXTERNAL_MODELS:
|
1034 |
+
if model not in EXTERNAL_MODEL_RESULTS:
|
1035 |
+
models_to_run.append(model)
|
1036 |
+
EXTERNAL_MODEL_RESULTS[model] = {k: {v: []} for k, v in TASK_TO_METRIC.items()}
|
1037 |
+
else:
|
1038 |
+
EXTERNAL_MODEL_RESULTS = {model: {k: {v: []} for k, v in TASK_TO_METRIC.items()} for model in EXTERNAL_MODELS}
|
1039 |
+
models_to_run = EXTERNAL_MODELS
|
1040 |
+
|
1041 |
+
pbar = tqdm(models_to_run, desc="Fetching external model results")
|
1042 |
+
for model in pbar:
|
1043 |
+
pbar.set_description(f"Fetching external model results for {model!r}")
|
1044 |
+
ds = load_dataset("mteb/results", model, trust_remote_code=True)
|
1045 |
+
# For local debugging:
|
1046 |
+
#, download_mode='force_redownload', verification_mode="no_checks")
|
1047 |
+
ds = ds.map(add_lang)
|
1048 |
+
ds = ds.map(add_task)
|
1049 |
+
base_dict = {"Model": make_clickable_model(model, link=EXTERNAL_MODEL_TO_LINK.get(model, "https://huggingface.co/spaces/mteb/leaderboard"))}
|
1050 |
+
# For now only one metric per task - Could add more metrics lateron
|
1051 |
+
for task, metric in TASK_TO_METRIC.items():
|
1052 |
+
ds_dict = ds.filter(lambda x: (x["mteb_task"] == task) and (x["metric"] == metric))["test"].to_dict()
|
1053 |
+
ds_dict = {k: round(v, 2) for k, v in zip(ds_dict["mteb_dataset_name_with_lang"], ds_dict["score"])}
|
1054 |
+
EXTERNAL_MODEL_RESULTS[model][task][metric].append({**base_dict, **ds_dict})
|
1055 |
+
|
1056 |
+
# Save & cache EXTERNAL_MODEL_RESULTS
|
1057 |
+
with open("EXTERNAL_MODEL_RESULTS.json", "w") as f:
|
1058 |
+
json.dump(EXTERNAL_MODEL_RESULTS, f)
|
1059 |
|
1060 |
def get_dim_seq_size(model):
|
1061 |
filenames = [sib.rfilename for sib in model.siblings]
|
|
|
1326 |
|
1327 |
return DATA_OVERALL_ZH
|
1328 |
|
1329 |
+
def get_mteb_average_fr():
|
1330 |
+
global DATA_OVERALL_FR, DATA_CLASSIFICATION_FR, DATA_CLUSTERING_FR, DATA_PAIR_CLASSIFICATION_FR, DATA_RERANKING_FR, DATA_RETRIEVAL_FR, DATA_STS_FR, DATA_SUMMARIZATION_FR
|
1331 |
+
DATA_OVERALL_FR = get_mteb_data(
|
1332 |
+
tasks=[
|
1333 |
+
"Classification",
|
1334 |
+
"Clustering",
|
1335 |
+
"PairClassification",
|
1336 |
+
"Reranking",
|
1337 |
+
"Retrieval",
|
1338 |
+
"STS",
|
1339 |
+
"Summarization"
|
1340 |
+
],
|
1341 |
+
datasets=TASK_LIST_CLASSIFICATION_FR + TASK_LIST_CLUSTERING_FR + TASK_LIST_PAIR_CLASSIFICATION_FR + TASK_LIST_RERANKING_FR + TASK_LIST_RETRIEVAL_FR + TASK_LIST_STS_FR + TASK_LIST_SUMMARIZATION_FR,
|
1342 |
+
fillna=False,
|
1343 |
+
add_emb_dim=True,
|
1344 |
+
rank=False,
|
1345 |
+
)
|
1346 |
+
# Debugging:
|
1347 |
+
# DATA_OVERALL_FR.to_csv("overall.csv")
|
1348 |
+
|
1349 |
+
DATA_OVERALL_FR.insert(1, f"Average ({len(TASK_LIST_FR)} datasets)", DATA_OVERALL_FR[TASK_LIST_FR].mean(axis=1, skipna=False))
|
1350 |
+
DATA_OVERALL_FR.insert(2, f"Classification Average ({len(TASK_LIST_CLASSIFICATION_FR)} datasets)", DATA_OVERALL_FR[TASK_LIST_CLASSIFICATION_FR].mean(axis=1, skipna=False))
|
1351 |
+
DATA_OVERALL_FR.insert(3, f"Clustering Average ({len(TASK_LIST_CLUSTERING_FR)} datasets)", DATA_OVERALL_FR[TASK_LIST_CLUSTERING_FR].mean(axis=1, skipna=False))
|
1352 |
+
DATA_OVERALL_FR.insert(4, f"Pair Classification Average ({len(TASK_LIST_PAIR_CLASSIFICATION_FR)} datasets)", DATA_OVERALL_FR[TASK_LIST_PAIR_CLASSIFICATION_FR].mean(axis=1, skipna=False))
|
1353 |
+
DATA_OVERALL_FR.insert(5, f"Reranking Average ({len(TASK_LIST_RERANKING_FR)} datasets)", DATA_OVERALL_FR[TASK_LIST_RERANKING_FR].mean(axis=1, skipna=False))
|
1354 |
+
DATA_OVERALL_FR.insert(6, f"Retrieval Average ({len(TASK_LIST_RETRIEVAL_FR)} datasets)", DATA_OVERALL_FR[TASK_LIST_RETRIEVAL_FR].mean(axis=1, skipna=False))
|
1355 |
+
DATA_OVERALL_FR.insert(7, f"STS Average ({len(TASK_LIST_STS_FR)} datasets)", DATA_OVERALL_FR[TASK_LIST_STS_FR].mean(axis=1, skipna=False))
|
1356 |
+
DATA_OVERALL_FR.insert(8, f"Summarization Average ({len(TASK_LIST_SUMMARIZATION_FR)} dataset)", DATA_OVERALL_FR[TASK_LIST_SUMMARIZATION_FR].mean(axis=1, skipna=False))
|
1357 |
+
DATA_OVERALL_FR.sort_values(f"Average ({len(TASK_LIST_FR)} datasets)", ascending=False, inplace=True)
|
1358 |
+
# Start ranking from 1
|
1359 |
+
DATA_OVERALL_FR.insert(0, "Rank", list(range(1, len(DATA_OVERALL_FR) + 1)))
|
1360 |
+
DATA_OVERALL_FR = DATA_OVERALL_FR.round(2)
|
1361 |
+
|
1362 |
+
DATA_CLASSIFICATION_FR = add_rank(DATA_OVERALL_FR[["Model"] + TASK_LIST_CLASSIFICATION_FR])
|
1363 |
+
DATA_CLASSIFICATION_FR = DATA_CLASSIFICATION_FR[DATA_CLASSIFICATION_FR.iloc[:, 2:].ne("").any(axis=1)]
|
1364 |
+
|
1365 |
+
DATA_CLUSTERING_FR = add_rank(DATA_OVERALL_FR[["Model"] + TASK_LIST_CLUSTERING_FR])
|
1366 |
+
DATA_CLUSTERING_FR = DATA_CLUSTERING_FR[DATA_CLUSTERING_FR.iloc[:, 2:].ne("").any(axis=1)]
|
1367 |
+
|
1368 |
+
DATA_PAIR_CLASSIFICATION_FR = add_rank(DATA_OVERALL_FR[["Model"] + TASK_LIST_PAIR_CLASSIFICATION_FR])
|
1369 |
+
DATA_PAIR_CLASSIFICATION_FR = DATA_PAIR_CLASSIFICATION_FR[DATA_PAIR_CLASSIFICATION_FR.iloc[:, 2:].ne("").any(axis=1)]
|
1370 |
+
|
1371 |
+
DATA_RERANKING_FR = add_rank(DATA_OVERALL_FR[["Model"] + TASK_LIST_RERANKING_FR])
|
1372 |
+
DATA_RERANKING_FR = DATA_RERANKING_FR[DATA_RERANKING_FR.iloc[:, 2:].ne("").any(axis=1)]
|
1373 |
+
|
1374 |
+
DATA_RETRIEVAL_FR = add_rank(DATA_OVERALL_FR[["Model"] + TASK_LIST_RETRIEVAL_FR])
|
1375 |
+
DATA_RETRIEVAL_FR = DATA_RETRIEVAL_FR[DATA_RETRIEVAL_FR.iloc[:, 2:].ne("").any(axis=1)]
|
1376 |
+
|
1377 |
+
DATA_STS_FR = add_rank(DATA_OVERALL_FR[["Model"] + TASK_LIST_STS_FR])
|
1378 |
+
DATA_STS_FR = DATA_STS_FR[DATA_STS_FR.iloc[:, 2:].ne("").any(axis=1)]
|
1379 |
+
|
1380 |
+
DATA_SUMMARIZATION_FR = add_rank(DATA_OVERALL_FR[["Model"] + TASK_LIST_SUMMARIZATION_FR])
|
1381 |
+
DATA_SUMMARIZATION_FR = DATA_SUMMARIZATION_FR[DATA_SUMMARIZATION_FR.iloc[:, 1:].ne("").any(axis=1)]
|
1382 |
+
|
1383 |
+
# Fill NaN after averaging
|
1384 |
+
DATA_OVERALL_FR.fillna("", inplace=True)
|
1385 |
+
|
1386 |
+
DATA_OVERALL_FR = DATA_OVERALL_FR[["Rank", "Model", "Model Size (GB)", "Embedding Dimensions", "Max Tokens", f"Average ({len(TASK_LIST_FR)} datasets)", f"Classification Average ({len(TASK_LIST_CLASSIFICATION_FR)} datasets)", f"Clustering Average ({len(TASK_LIST_CLUSTERING_FR)} datasets)", f"Pair Classification Average ({len(TASK_LIST_PAIR_CLASSIFICATION_FR)} datasets)", f"Reranking Average ({len(TASK_LIST_RERANKING_FR)} datasets)", f"Retrieval Average ({len(TASK_LIST_RETRIEVAL_FR)} datasets)", f"STS Average ({len(TASK_LIST_STS_FR)} datasets)", f"Summarization Average ({len(TASK_LIST_SUMMARIZATION_FR)} dataset)"]]
|
1387 |
+
DATA_OVERALL_FR = DATA_OVERALL_FR[DATA_OVERALL_FR.iloc[:, 5:].ne("").any(axis=1)]
|
1388 |
+
|
1389 |
+
return DATA_OVERALL_FR
|
1390 |
+
|
1391 |
def get_mteb_average_pl():
|
1392 |
global DATA_OVERALL_PL, DATA_CLASSIFICATION_PL, DATA_CLUSTERING_PL, DATA_PAIR_CLASSIFICATION_PL, DATA_RETRIEVAL_PL, DATA_STS_PL
|
1393 |
DATA_OVERALL_PL = get_mteb_data(
|
|
|
1443 |
return DATA_OVERALL_PL
|
1444 |
|
1445 |
get_mteb_average()
|
1446 |
+
get_mteb_average_fr()
|
1447 |
get_mteb_average_pl()
|
1448 |
get_mteb_average_zh()
|
1449 |
DATA_BITEXT_MINING = get_mteb_data(["BitextMining"], [], TASK_LIST_BITEXT_MINING)
|
|
|
1465 |
DATA_BITEXT_MINING_OTHER,
|
1466 |
DATA_CLASSIFICATION_EN,
|
1467 |
DATA_CLASSIFICATION_DA,
|
1468 |
+
DATA_CLASSIFICATION_FR,
|
1469 |
DATA_CLASSIFICATION_NB,
|
1470 |
DATA_CLASSIFICATION_PL,
|
1471 |
DATA_CLASSIFICATION_SV,
|
|
|
1473 |
DATA_CLASSIFICATION_OTHER,
|
1474 |
DATA_CLUSTERING,
|
1475 |
DATA_CLUSTERING_DE,
|
1476 |
+
DATA_CLUSTERING_FR,
|
1477 |
DATA_CLUSTERING_PL,
|
1478 |
DATA_CLUSTERING_ZH,
|
1479 |
DATA_PAIR_CLASSIFICATION,
|
1480 |
+
DATA_PAIR_CLASSIFICATION_FR,
|
1481 |
DATA_PAIR_CLASSIFICATION_PL,
|
1482 |
DATA_PAIR_CLASSIFICATION_ZH,
|
1483 |
DATA_RERANKING,
|
1484 |
+
DATA_RERANKING_FR,
|
1485 |
DATA_RERANKING_ZH,
|
1486 |
DATA_RETRIEVAL,
|
1487 |
+
DATA_RETRIEVAL_FR,
|
1488 |
DATA_RETRIEVAL_PL,
|
1489 |
DATA_RETRIEVAL_ZH,
|
1490 |
DATA_STS_EN,
|
1491 |
+
DATA_STS_FR,
|
1492 |
DATA_STS_PL,
|
1493 |
DATA_STS_ZH,
|
1494 |
DATA_STS_OTHER,
|
1495 |
DATA_SUMMARIZATION,
|
1496 |
+
DATA_SUMMARIZATION_FR,
|
1497 |
]:
|
1498 |
# NUM_SCORES += d.iloc[:, 1:].apply(lambda x: sum([1 for y in x if isinstance(y, float) and not np.isnan(y)]), axis=1).sum()
|
1499 |
cols_to_ignore = 3 if "Average" in d.columns else 2
|
|
|
1568 |
)
|
1569 |
with gr.Row():
|
1570 |
data_run_overall_zh = gr.Button("Refresh")
|
1571 |
+
data_run_overall_zh.click(get_mteb_average_zh, inputs=None, outputs=data_overall_zh)
|
1572 |
+
with gr.TabItem("French"):
|
1573 |
+
with gr.Row():
|
1574 |
+
gr.Markdown("""
|
1575 |
+
**Overall MTEB French leaderboard (F-MTEB)** 🔮🇫🇷
|
1576 |
+
|
1577 |
+
- **Metric:** Various, refer to task tabs
|
1578 |
+
- **Languages:** French
|
1579 |
+
- **Credits:** [Lyon-NLP](https://github.com/Lyon-NLP): [Gabriel Sequeira](https://github.com/GabrielSequeira), [Imene Kerboua](https://github.com/imenelydiaker), [wissam-sib](https://github.com/wissam-sib), [Sunalwing](https://github.com/Sunalwing)
|
1580 |
+
""")
|
1581 |
+
with gr.Row():
|
1582 |
+
data_overall_fr = gr.components.Dataframe(
|
1583 |
+
DATA_OVERALL_FR,
|
1584 |
+
datatype=["number", "markdown"] + ["number"] * len(DATA_OVERALL_FR.columns),
|
1585 |
+
type="pandas",
|
1586 |
+
height=600,
|
1587 |
+
)
|
1588 |
+
with gr.Row():
|
1589 |
+
data_overall_fr = gr.Button("Refresh")
|
1590 |
+
data_overall_fr.click(get_mteb_average_fr, inputs=None, outputs=data_overall_fr)
|
1591 |
with gr.TabItem("Polish"):
|
1592 |
with gr.Row():
|
1593 |
gr.Markdown("""
|
|
|
1712 |
partial(get_mteb_data, tasks=["Classification"], datasets=TASK_LIST_CLASSIFICATION_DA),
|
1713 |
outputs=data_run_classification_da,
|
1714 |
)
|
1715 |
+
with gr.TabItem("French"):
|
1716 |
+
with gr.Row():
|
1717 |
+
gr.Markdown("""
|
1718 |
+
**Classification French Leaderboard** 💙🇫🇷
|
1719 |
+
|
1720 |
+
- **Metric:** [Accuracy](https://huggingface.co/spaces/evaluate-metric/accuracy)
|
1721 |
+
- **Languages:** French
|
1722 |
+
- **Credits:**
|
1723 |
+
""")
|
1724 |
+
with gr.Row():
|
1725 |
+
data_classification_fr = gr.components.Dataframe(
|
1726 |
+
DATA_CLASSIFICATION_FR,
|
1727 |
+
datatype=["number", "markdown"] + ["number"] * len(DATA_CLASSIFICATION_FR.columns),
|
1728 |
+
type="pandas",
|
1729 |
+
)
|
1730 |
+
with gr.Row():
|
1731 |
+
data_run_classification_fr = gr.Button("Refresh")
|
1732 |
+
data_run_classification_fr.click(
|
1733 |
+
partial(get_mteb_data, tasks=["Classification"], datasets=TASK_LIST_CLASSIFICATION_FR),
|
1734 |
+
outputs=data_run_classification_fr,
|
1735 |
+
)
|
1736 |
with gr.TabItem("Norwegian"):
|
1737 |
with gr.Row():
|
1738 |
gr.Markdown("""
|
|
|
1858 |
partial(get_mteb_data, tasks=["Clustering"], datasets=TASK_LIST_CLUSTERING_ZH),
|
1859 |
outputs=data_clustering_zh,
|
1860 |
)
|
1861 |
+
with gr.TabItem("French"):
|
1862 |
+
with gr.Row():
|
1863 |
+
gr.Markdown("""
|
1864 |
+
**Clustering French Leaderboard** ✨🇫🇷
|
1865 |
+
|
1866 |
+
- **Metric:** Validity Measure (v_measure)
|
1867 |
+
- **Languages:** French
|
1868 |
+
- **Credits:**
|
1869 |
+
""")
|
1870 |
+
with gr.Row():
|
1871 |
+
data_clustering_fr = gr.components.Dataframe(
|
1872 |
+
DATA_CLUSTERING_FR,
|
1873 |
+
datatype=["number", "markdown"] + ["number"] * len(DATA_CLUSTERING_FR.columns),
|
1874 |
+
type="pandas",
|
1875 |
+
)
|
1876 |
+
with gr.Row():
|
1877 |
+
data_run_clustering_fr = gr.Button("Refresh")
|
1878 |
+
data_run_clustering_fr.click(
|
1879 |
+
partial(get_mteb_data, tasks=["Clustering"], datasets=TASK_LIST_CLUSTERING_FR),
|
1880 |
+
outputs=data_clustering_fr,
|
1881 |
+
)
|
1882 |
with gr.TabItem("German"):
|
1883 |
with gr.Row():
|
1884 |
gr.Markdown("""
|
|
|
1963 |
partial(get_mteb_data, tasks=["PairClassification"], datasets=TASK_LIST_PAIR_CLASSIFICATION_ZH),
|
1964 |
outputs=data_pair_classification_zh,
|
1965 |
)
|
1966 |
+
with gr.TabItem("French"):
|
1967 |
+
with gr.Row():
|
1968 |
+
gr.Markdown("""
|
1969 |
+
**Pair Classification French Leaderboard** 🎭🇫🇷
|
1970 |
+
|
1971 |
+
- **Metric:** Average Precision based on Cosine Similarities (cos_sim_ap)
|
1972 |
+
- **Languages:** French
|
1973 |
+
- **Credits:**
|
1974 |
+
""")
|
1975 |
+
with gr.Row():
|
1976 |
+
data_pair_classification_fr = gr.components.Dataframe(
|
1977 |
+
DATA_PAIR_CLASSIFICATION_FR,
|
1978 |
+
datatype=["number", "markdown"] + ["number"] * len(DATA_PAIR_CLASSIFICATION_FR.columns),
|
1979 |
+
type="pandas",
|
1980 |
+
)
|
1981 |
+
with gr.Row():
|
1982 |
+
data_run_pair_classification_fr = gr.Button("Refresh")
|
1983 |
+
data_run_pair_classification_fr.click(
|
1984 |
+
partial(get_mteb_data, tasks=["PairClassification"], datasets=TASK_LIST_PAIR_CLASSIFICATION_FR),
|
1985 |
+
outputs=data_pair_classification_fr,
|
1986 |
+
)
|
1987 |
with gr.TabItem("Polish"):
|
1988 |
with gr.Row():
|
1989 |
gr.Markdown("""
|
|
|
2047 |
partial(get_mteb_data, tasks=["Reranking"], datasets=TASK_LIST_RERANKING_ZH),
|
2048 |
outputs=data_reranking_zh,
|
2049 |
)
|
2050 |
+
with gr.TabItem("French"):
|
2051 |
+
with gr.Row():
|
2052 |
+
gr.Markdown("""
|
2053 |
+
**Reranking French Leaderboard** 🥈🇫🇷
|
2054 |
+
|
2055 |
+
- **Metric:** Mean Average Precision (MAP)
|
2056 |
+
- **Languages:** French
|
2057 |
+
- **Credits:**
|
2058 |
+
""")
|
2059 |
+
with gr.Row():
|
2060 |
+
data_reranking_fr = gr.components.Dataframe(
|
2061 |
+
DATA_RERANKING_FR,
|
2062 |
+
datatype=["number", "markdown"] + ["number"] * len(DATA_RERANKING_FR.columns),
|
2063 |
+
type="pandas",
|
2064 |
+
)
|
2065 |
+
with gr.Row():
|
2066 |
+
data_run_reranking_fr = gr.Button("Refresh")
|
2067 |
+
data_run_reranking_fr.click(
|
2068 |
+
partial(get_mteb_data, tasks=["Reranking"], datasets=TASK_LIST_RERANKING_FR),
|
2069 |
+
outputs=data_reranking_fr,
|
2070 |
+
)
|
2071 |
with gr.TabItem("Retrieval"):
|
2072 |
with gr.TabItem("English"):
|
2073 |
with gr.Row():
|
|
|
2100 |
- **Credits:** [FlagEmbedding](https://github.com/FlagOpen/FlagEmbedding)
|
2101 |
""")
|
2102 |
with gr.Row():
|
2103 |
+
data_retrieval_fr = gr.components.Dataframe(
|
2104 |
+
DATA_RETRIEVAL_FR,
|
2105 |
# Add support for more columns than existing as a buffer for CQADupstack & other Retrieval tasks (e.g. MSMARCOv2)
|
2106 |
+
datatype=["number", "markdown"] + ["number"] * len(DATA_RETRIEVAL_FR.columns) * 2,
|
2107 |
type="pandas",
|
2108 |
)
|
2109 |
with gr.Row():
|
2110 |
+
data_run_retrieval_fr = gr.Button("Refresh")
|
2111 |
+
data_run_retrieval_fr.click(
|
2112 |
+
partial(get_mteb_data, tasks=["Retrieval"], datasets=TASK_LIST_RETRIEVAL_FR),
|
2113 |
+
outputs=data_retrieval_fr,
|
2114 |
)
|
2115 |
+
with gr.TabItem("French"):
|
2116 |
+
with gr.Row():
|
2117 |
+
gr.Markdown("""
|
2118 |
+
**Retrieval French Leaderboard** 🔎🇫🇷
|
2119 |
+
|
2120 |
+
- **Metric:** Normalized Discounted Cumulative Gain @ k (ndcg_at_10)
|
2121 |
+
- **Languages:** French
|
2122 |
+
- **Credits:**
|
2123 |
+
""")
|
2124 |
+
with gr.Row():
|
2125 |
+
data_retrieval_fr = gr.components.Dataframe(
|
2126 |
+
DATA_RETRIEVAL_FR,
|
2127 |
+
# Add support for more columns than existing as a buffer for CQADupstack & other Retrieval tasks (e.g. MSMARCOv2)
|
2128 |
+
datatype=["number", "markdown"] + ["number"] * len(DATA_RETRIEVAL_FR.columns) * 2,
|
2129 |
+
type="pandas",
|
2130 |
+
)
|
2131 |
+
with gr.Row():
|
2132 |
+
data_run_retrieval_fr = gr.Button("Refresh")
|
2133 |
+
data_run_retrieval_fr.click(
|
2134 |
+
partial(get_mteb_data, tasks=["Retrieval"], datasets=TASK_LIST_RETRIEVAL_FR),
|
2135 |
+
outputs=data_retrieval_fr,
|
2136 |
+
)
|
2137 |
with gr.TabItem("Polish"):
|
2138 |
with gr.Row():
|
2139 |
gr.Markdown("""
|
|
|
2198 |
partial(get_mteb_data, tasks=["STS"], datasets=TASK_LIST_STS_ZH),
|
2199 |
outputs=data_sts_zh,
|
2200 |
)
|
2201 |
+
with gr.TabItem("French"):
|
2202 |
+
with gr.Row():
|
2203 |
+
gr.Markdown("""
|
2204 |
+
**STS French Leaderboard** 🤖🇫🇷
|
2205 |
+
|
2206 |
+
- **Metric:** Spearman correlation based on cosine similarity
|
2207 |
+
- **Languages:** French
|
2208 |
+
- **Credits:**
|
2209 |
+
""")
|
2210 |
+
with gr.Row():
|
2211 |
+
data_sts_fr = gr.components.Dataframe(
|
2212 |
+
DATA_STS_FR,
|
2213 |
+
datatype=["number", "markdown"] + ["number"] * len(DATA_STS_FR.columns),
|
2214 |
+
type="pandas",
|
2215 |
+
)
|
2216 |
+
with gr.Row():
|
2217 |
+
data_run_sts_fr = gr.Button("Refresh")
|
2218 |
+
data_run_sts_fr.click(
|
2219 |
+
partial(get_mteb_data, tasks=["STS"], datasets=TASK_LIST_STS_FR),
|
2220 |
+
outputs=data_sts_fr,
|
2221 |
+
)
|
2222 |
with gr.TabItem("Polish"):
|
2223 |
with gr.Row():
|
2224 |
gr.Markdown("""
|