CMU-Wav2Gloss/retriever_Uspanteko

SentenceTransformer based on sentence-transformers/LaBSE

This is a sentence-transformers model finetuned from sentence-transformers/LaBSE. It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: sentence-transformers/LaBSE
• Maximum Sequence Length: 256 tokens
• Output Dimensionality: 768 dimensions
• Similarity Function: Cosine Similarity

Model Sources

• Documentation: Sentence Transformers Documentation
• Repository: Sentence Transformers on GitHub
• Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 256, 'do_lower_case': False, 'architecture': 'BertModel'})
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
  (2): Dense({'in_features': 768, 'out_features': 768, 'bias': True, 'activation_function': 'torch.nn.modules.activation.Tanh'})
  (3): Normalize()
)

Usage

Direct Usage (Sentence Transformers)

First install the Sentence Transformers library:

pip install -U sentence-transformers

Then you can load this model and run inference.

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("sentence_transformers_model_id")
# Run inference
sentences = [
    "Word: r a| Context: r a   j b ' a q i i l   r a   u t i i w .| Translation: El huesito del coyotillo.",
    'Morpheme: r a | Gloss: DIM',
    'Morpheme: p o r k e | Gloss: porque',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[1.0000, 0.8346, 0.1137],
#         [0.8346, 1.0000, 0.0429],
#         [0.1137, 0.0429, 1.0000]])

Evaluation

Metrics

IR

• Dataset: validation
• Evaluated with main.IREvaluatorWithLogging

Metric	Value
cosine_accuracy@1	0.6535
cosine_accuracy@3	0.9351
cosine_accuracy@5	0.9606
cosine_accuracy@10	0.9791
cosine_precision@1	0.6535
cosine_precision@3	0.3449
cosine_precision@5	0.2255
cosine_precision@10	0.1256
cosine_recall@1	0.5838
cosine_recall@3	0.8582
cosine_recall@5	0.9008
cosine_recall@10	0.9527
cosine_ndcg@10	0.8189
cosine_mrr@10	0.7922
cosine_map@100	0.7623

Training Details

Training Dataset

Unnamed Dataset

• Size: 60,444 training samples
• Columns: sentence_0, sentence_1, and label
• Approximate statistics based on the first 1000 samples:

  	sentence_0	sentence_1	label
  type	string	string	int
  details	min: 21 tokens mean: 49.04 tokens max: 116 tokens	min: 11 tokens mean: 13.51 tokens max: 24 tokens	21: ~0.10% 64: ~0.10% 91: ~0.10% 114: ~0.10% 118: ~0.10% 154: ~0.10% 163: ~0.10% 176: ~0.10% 231: ~0.10% 234: ~0.10% 299: ~0.10% 370: ~0.10% 374: ~0.10% 376: ~0.10% 379: ~0.10% 402: ~0.10% 432: ~0.10% 446: ~0.10% 454: ~0.10% 499: ~0.10% 666: ~0.10% 680: ~0.10% 714: ~0.10% 741: ~0.10% 764: ~0.10% 833: ~0.10% 840: ~0.20% 847: ~0.10% 1027: ~0.10% 1044: ~0.10% 1072: ~0.10% 1083: ~0.10% 1085: ~0.10% 1126: ~0.10% 1196: ~0.10% 1277: ~0.10% 1287: ~0.10% 1400: ~0.10% 1402: ~0.10% 1439: ~0.10% 1470: ~0.20% 1483: ~0.10% 1531: ~0.10% 1567: ~0.10% 1570: ~0.10% 1710: ~0.10% 1778: ~0.10% 1786: ~0.10% 1812: ~0.10% 1826: ~0.10% 1854: ~0.10% 1861: ~0.10% 1941: ~0.10% 1950: ~0.10% 1978: ~0.10% 2139: ~0.10% 2146: ~0.10% 2231: ~0.10% 2251: ~0.10% 2255: ~0.10% 2346: ~0.10% 2380: ~0.10% 2409: ~0.10% 2480: ~0.10% 2615: ~0.10% 2709: ~0.10% 2736: ~0.10% 2747: ~0.10% 2778: ~0.10% 2795: ~0.10% 2796: ~0.10% 2800: ~0.10% 2801: ~0.10% 2818: ~0.10% 2872: ~0.10% 2881: ~0.10% 2969: ~0.10% 2997: ~0.10% 3028: ~0.10% 3036: ~0.10% 3054: ~0.10% 3115: ~0.10% 3280: ~0.10% 3290: ~0.20% 3327: ~0.10% 3362: ~0.10% 3375: ~0.10% 3469: ~0.10% 3482: ~0.10% 3507: ~0.10% 3518: ~0.10% 3539: ~0.10% 3568: ~0.10% 3635: ~0.10% 3692: ~0.10% 3712: ~0.10% 3719: ~0.10% 3742: ~0.10% 3761: ~0.10% 3812: ~0.10% 3854: ~0.10% 3867: ~0.10% 3913: ~0.10% 3935: ~0.10% 3948: ~0.10% 3975: ~0.10% 3985: ~0.10% 4015: ~0.10% 4021: ~0.10% 4034: ~0.10% 4178: ~0.10% 4216: ~0.10% 4221: ~0.10% 4250: ~0.10% 4338: ~0.10% 4349: ~0.10% 4362: ~0.10% 4399: ~0.10% 4422: ~0.10% 4440: ~0.10% 4485: ~0.10% 4493: ~0.10% 4496: ~0.10% 4515: ~0.10% 4535: ~0.10% 4568: ~0.10% 4611: ~0.10% 4721: ~0.10% 4885: ~0.10% 5059: ~0.10% 5082: ~0.10% 5102: ~0.20% 5105: ~0.10% 5196: ~0.10% 5197: ~0.10% 5224: ~0.10% 5235: ~0.10% 5252: ~0.10% 5256: ~0.10% 5297: ~0.10% 5394: ~0.10% 5459: ~0.10% 5477: ~0.30% 5498: ~0.20% 5499: ~0.10% 5506: ~0.10% 5512: ~0.10% 5532: ~0.10% 5541: ~0.10% 5586: ~0.10% 5592: ~0.10% 5596: ~0.10% 5605: ~0.10% 5652: ~0.10% 5714: ~0.20% 5719: ~0.10% 5748: ~0.10% 5761: ~0.10% 5785: ~0.10% 5786: ~0.10% 5794: ~0.10% 5825: ~0.10% 5829: ~0.10% 5840: ~0.10% 5844: ~0.10% 5863: ~0.10% 5876: ~0.10% 5881: ~0.10% 5907: ~0.10% 5918: ~0.10% 5924: ~0.10% 5938: ~0.10% 5952: ~0.10% 5991: ~0.10% 5995: ~0.10% 6051: ~0.10% 6056: ~0.10% 6078: ~0.10% 6081: ~0.10% 6145: ~0.10% 6182: ~0.10% 6209: ~0.10% 6219: ~0.10% 6239: ~0.10% 6282: ~0.10% 6287: ~0.20% 6294: ~0.10% 6303: ~0.10% 6340: ~0.10% 6357: ~0.10% 6378: ~0.10% 6437: ~0.10% 6494: ~0.10% 6498: ~0.10% 6504: ~0.10% 6531: ~0.10% 6532: ~0.20% 6593: ~0.10% 6615: ~0.10% 6619: ~0.10% 6631: ~0.10% 6683: ~0.10% 6688: ~0.10% 6698: ~0.10% 6743: ~0.10% 6765: ~0.10% 6817: ~0.10% 6846: ~0.10% 6872: ~0.10% 6874: ~0.10% 6888: ~0.10% 6899: ~0.20% 6909: ~0.10% 6913: ~0.20% 6918: ~0.10% 6920: ~0.10% 6969: ~0.10% 6978: ~0.10% 6979: ~0.10% 6986: ~0.10% 6992: ~0.10% 6997: ~0.10% 7075: ~0.10% 7082: ~0.10% 7095: ~0.10% 7101: ~0.10% 7104: ~0.10% 7106: ~0.10% 7128: ~0.10% 7198: ~0.10% 7219: ~0.10% 7243: ~0.10% 7276: ~0.10% 7292: ~0.10% 7306: ~0.10% 7313: ~0.10% 7315: ~0.20% 7333: ~0.10% 7336: ~0.10% 7357: ~0.10% 7382: ~0.10% 7436: ~0.10% 7441: ~0.10% 7466: ~0.10% 7485: ~0.10% 7497: ~0.10% 7520: ~0.10% 7522: ~0.10% 7571: ~0.10% 7582: ~0.10% 7610: ~0.10% 7709: ~0.10% 7716: ~0.10% 7755: ~0.10% 7770: ~0.10% 7793: ~0.10% 7828: ~0.10% 7871: ~0.10% 8005: ~0.10% 8017: ~0.10% 8037: ~0.10% 8073: ~0.10% 8089: ~0.10% 8111: ~0.10% 8188: ~0.10% 8212: ~0.10% 8231: ~0.10% 8241: ~0.10% 8304: ~0.10% 8334: ~0.10% 8342: ~0.10% 8355: ~0.10% 8395: ~0.10% 8411: ~0.10% 8472: ~0.10% 8496: ~0.10% 8502: ~0.10% 8527: ~0.10% 8609: ~0.10% 8642: ~0.10% 8728: ~0.10% 8803: ~0.10% 8844: ~0.10% 8852: ~0.10% 8878: ~0.10% 8881: ~0.10% 8906: ~0.10% 8968: ~0.10% 8972: ~0.10% 9012: ~0.10% 9070: ~0.10% 9123: ~0.10% 9175: ~0.10% 9208: ~0.10% 9280: ~0.10% 9299: ~0.10% 9335: ~0.10% 9360: ~0.10% 9410: ~0.10% 9419: ~0.10% 9424: ~0.10% 9597: ~0.10% 9605: ~0.10% 9629: ~0.10% 9642: ~0.10% 9671: ~0.10% 9702: ~0.20% 9703: ~0.10% 9741: ~0.10% 9744: ~0.10% 9752: ~0.10% 9769: ~0.10% 9810: ~0.10% 9903: ~0.10% 9905: ~0.10% 10022: ~0.10% 10113: ~0.10% 10168: ~0.10% 10294: ~0.10% 10303: ~0.10% 10348: ~0.10% 10418: ~0.10% 10451: ~0.10% 10503: ~0.10% 10512: ~0.10% 10514: ~0.20% 10592: ~0.10% 10622: ~0.10% 10693: ~0.10% 10815: ~0.10% 10828: ~0.10% 10910: ~0.10% 10958: ~0.10% 10971: ~0.10% 10973: ~0.10% 10994: ~0.20% 11007: ~0.10% 11049: ~0.10% 11050: ~0.10% 11068: ~0.10% 11105: ~0.10% 11115: ~0.10% 11199: ~0.10% 11216: ~0.10% 11232: ~0.10% 11268: ~0.10% 11283: ~0.10% 11366: ~0.10% 11401: ~0.10% 11433: ~0.10% 11558: ~0.10% 11574: ~0.10% 11667: ~0.10% 11719: ~0.10% 11749: ~0.10% 11752: ~0.10% 11799: ~0.10% 11802: ~0.10% 11815: ~0.10% 11826: ~0.10% 11900: ~0.10% 12000: ~0.10% 12111: ~0.10% 12125: ~0.10% 12150: ~0.10% 12212: ~0.10% 12290: ~0.10% 12306: ~0.10% 12319: ~0.10% 12350: ~0.10% 12376: ~0.10% 12458: ~0.10% 12515: ~0.10% 12552: ~0.10% 12597: ~0.10% 12646: ~0.10% 12791: ~0.10% 12810: ~0.10% 12815: ~0.10% 12825: ~0.10% 12840: ~0.10% 12899: ~0.10% 12912: ~0.10% 12931: ~0.10% 12993: ~0.10% 13101: ~0.10% 13124: ~0.10% 13173: ~0.10% 13254: ~0.10% 13255: ~0.10% 13328: ~0.10% 13331: ~0.10% 13414: ~0.10% 13425: ~0.10% 13470: ~0.10% 13479: ~0.10% 13531: ~0.10% 13615: ~0.10% 13637: ~0.10% 13677: ~0.10% 13755: ~0.10% 13799: ~0.10% 13815: ~0.10% 13832: ~0.10% 13898: ~0.10% 13921: ~0.10% 13954: ~0.10% 13958: ~0.10% 14009: ~0.10% 14056: ~0.10% 14059: ~0.10% 14103: ~0.10% 14166: ~0.10% 14288: ~0.10% 14296: ~0.10% 14391: ~0.10% 14443: ~0.10% 14529: ~0.10% 14544: ~0.10% 14567: ~0.10% 14656: ~0.10% 14684: ~0.10% 14727: ~0.10% 14744: ~0.10% 14771: ~0.10% 14793: ~0.10% 14798: ~0.10% 14807: ~0.10% 14815: ~0.10% 14817: ~0.10% 14820: ~0.10% 14948: ~0.10% 14964: ~0.10% 15128: ~0.10% 15155: ~0.10% 15197: ~0.10% 15207: ~0.10% 15279: ~0.10% 15422: ~0.10% 15430: ~0.10% 15454: ~0.10% 15478: ~0.10% 15501: ~0.10% 15504: ~0.10% 15600: ~0.10% 15607: ~0.10% 15643: ~0.10% 15684: ~0.10% 15702: ~0.10% 15735: ~0.10% 15756: ~0.10% 15811: ~0.10% 15926: ~0.10% 15964: ~0.10% 15969: ~0.10% 16083: ~0.10% 16090: ~0.10% 16103: ~0.10% 16116: ~0.10% 16127: ~0.10% 16184: ~0.10% 16223: ~0.10% 16235: ~0.10% 16267: ~0.10% 16350: ~0.10% 16351: ~0.10% 16374: ~0.10% 16405: ~0.10% 16480: ~0.10% 16576: ~0.10% 16686: ~0.10% 16687: ~0.10% 16723: ~0.10% 16780: ~0.10% 16819: ~0.10% 16834: ~0.10% 16920: ~0.10% 16976: ~0.10% 17119: ~0.10% 17234: ~0.10% 17262: ~0.10% 17311: ~0.10% 17312: ~0.10% 17384: ~0.10% 17400: ~0.10% 17406: ~0.10% 17478: ~0.10% 17523: ~0.10% 17542: ~0.10% 17543: ~0.10% 17576: ~0.10% 17641: ~0.10% 17680: ~0.10% 17685: ~0.10% 17714: ~0.10% 17754: ~0.10% 17773: ~0.10% 17782: ~0.10% 17887: ~0.10% 17907: ~0.10% 17939: ~0.10% 18070: ~0.10% 18078: ~0.10% 18180: ~0.10% 18181: ~0.10% 18186: ~0.10% 18260: ~0.10% 18355: ~0.10% 18360: ~0.10% 18362: ~0.10% 18501: ~0.10% 18503: ~0.10% 18511: ~0.10% 18528: ~0.10% 18635: ~0.10% 18638: ~0.10% 18652: ~0.10% 18717: ~0.10% 18725: ~0.10% 18732: ~0.10% 18740: ~0.10% 18769: ~0.10% 18830: ~0.20% 18863: ~0.10% 18893: ~0.10% 18905: ~0.10% 18912: ~0.10% 18930: ~0.10% 18939: ~0.10% 18975: ~0.10% 18984: ~0.10% 18986: ~0.10% 18996: ~0.10% 19034: ~0.10% 19039: ~0.20% 19078: ~0.10% 19080: ~0.10% 19082: ~0.10% 19092: ~0.10% 19119: ~0.10% 19242: ~0.10% 19274: ~0.10% 19323: ~0.10% 19455: ~0.10% 19498: ~0.10% 19558: ~0.10% 19561: ~0.10% 19610: ~0.10% 19639: ~0.10% 19692: ~0.10% 19696: ~0.10% 19757: ~0.10% 19764: ~0.10% 19792: ~0.10% 19822: ~0.10% 19842: ~0.10% 19850: ~0.10% 19861: ~0.10% 19927: ~0.10% 19959: ~0.10% 19971: ~0.10% 19973: ~0.10% 20022: ~0.10% 20085: ~0.10% 20176: ~0.10% 20182: ~0.10% 20349: ~0.10% 20351: ~0.10% 20372: ~0.10% 20382: ~0.10% 20480: ~0.10% 20491: ~0.10% 20493: ~0.10% 20587: ~0.10% 20626: ~0.10% 20680: ~0.10% 20695: ~0.10% 20705: ~0.10% 20770: ~0.10% 20820: ~0.10% 20928: ~0.10% 20943: ~0.10% 21050: ~0.10% 21081: ~0.10% 21086: ~0.10% 21131: ~0.10% 21158: ~0.10% 21230: ~0.10% 21243: ~0.10% 21307: ~0.10% 21403: ~0.10% 21427: ~0.10% 21436: ~0.10% 21478: ~0.10% 21504: ~0.10% 21563: ~0.10% 21610: ~0.10% 21614: ~0.10% 21667: ~0.10% 21695: ~0.10% 21865: ~0.10% 21867: ~0.10% 21895: ~0.10% 21963: ~0.10% 22043: ~0.10% 22050: ~0.10% 22063: ~0.10% 22157: ~0.10% 22184: ~0.10% 22215: ~0.10% 22231: ~0.10% 22249: ~0.10% 22251: ~0.10% 22254: ~0.10% 22301: ~0.10% 22332: ~0.10% 22405: ~0.10% 22406: ~0.10% 22441: ~0.10% 22479: ~0.10% 22496: ~0.10% 22535: ~0.10% 22543: ~0.10% 22592: ~0.10% 22632: ~0.10% 22675: ~0.10% 22697: ~0.10% 22733: ~0.10% 22760: ~0.10% 22773: ~0.10% 22897: ~0.10% 22906: ~0.10% 22908: ~0.10% 22912: ~0.10% 22983: ~0.10% 23033: ~0.10% 23035: ~0.10% 23109: ~0.10% 23200: ~0.10% 23210: ~0.10% 23229: ~0.10% 23251: ~0.10% 23280: ~0.10% 23391: ~0.10% 23403: ~0.10% 23419: ~0.10% 23465: ~0.10% 23475: ~0.10% 23500: ~0.10% 23509: ~0.10% 23514: ~0.10% 23537: ~0.10% 23571: ~0.10% 23609: ~0.10% 23627: ~0.10% 23684: ~0.10% 23744: ~0.10% 23791: ~0.10% 23804: ~0.10% 23817: ~0.10% 23885: ~0.10% 23983: ~0.10% 24038: ~0.10% 24050: ~0.10% 24071: ~0.10% 24088: ~0.10% 24149: ~0.10% 24240: ~0.10% 24366: ~0.10% 24396: ~0.10% 24514: ~0.10% 24522: ~0.10% 24565: ~0.10% 24568: ~0.10% 24571: ~0.10% 24588: ~0.10% 24606: ~0.10% 24616: ~0.10% 24636: ~0.10% 24662: ~0.10% 24680: ~0.10% 24681: ~0.10% 24735: ~0.10% 24737: ~0.10% 24778: ~0.10% 24787: ~0.10% 24789: ~0.10% 24796: ~0.10% 24855: ~0.10% 24972: ~0.10% 25095: ~0.10% 25128: ~0.10% 25173: ~0.10% 25184: ~0.10% 25189: ~0.10% 25207: ~0.10% 25259: ~0.20% 25279: ~0.10% 25344: ~0.10% 25408: ~0.10% 25425: ~0.10% 25477: ~0.10% 25590: ~0.10% 25726: ~0.10% 25937: ~0.10% 25989: ~0.10% 26011: ~0.10% 26012: ~0.10% 26068: ~0.10% 26107: ~0.10% 26230: ~0.10% 26299: ~0.10% 26314: ~0.10% 26330: ~0.10% 26373: ~0.10% 26607: ~0.10% 26619: ~0.10% 26657: ~0.10% 26809: ~0.10% 26824: ~0.10% 26897: ~0.10% 26920: ~0.10% 26930: ~0.10% 26939: ~0.10% 27018: ~0.10% 27035: ~0.10% 27118: ~0.10% 27181: ~0.10% 27225: ~0.10% 27260: ~0.10% 27316: ~0.10% 27352: ~0.10% 27360: ~0.10% 27371: ~0.10% 27376: ~0.10% 27463: ~0.10% 27470: ~0.10% 27534: ~0.10% 27552: ~0.10% 27566: ~0.10% 27659: ~0.10% 27786: ~0.10% 27868: ~0.10% 27889: ~0.10% 27972: ~0.10% 28030: ~0.10% 28046: ~0.10% 28087: ~0.10% 28214: ~0.10% 28263: ~0.10% 28298: ~0.10% 28331: ~0.10% 28346: ~0.10% 28381: ~0.10% 28404: ~0.10% 28417: ~0.10% 28435: ~0.10% 28444: ~0.10% 28448: ~0.10% 28480: ~0.10% 28520: ~0.10% 28538: ~0.10% 28545: ~0.10% 28583: ~0.10% 28614: ~0.10% 28654: ~0.10% 28667: ~0.10% 28671: ~0.10% 28754: ~0.10% 28805: ~0.10% 28806: ~0.10% 28855: ~0.10% 28857: ~0.10% 28862: ~0.10% 28892: ~0.10% 28918: ~0.10% 28922: ~0.10% 29019: ~0.10% 29028: ~0.10% 29059: ~0.10% 29065: ~0.10% 29085: ~0.10% 29100: ~0.10% 29182: ~0.10% 29196: ~0.10% 29232: ~0.10% 29307: ~0.10% 29326: ~0.10% 29368: ~0.10% 29399: ~0.10% 29409: ~0.10% 29415: ~0.10% 29430: ~0.10% 29443: ~0.10% 29447: ~0.10% 29450: ~0.10% 29513: ~0.10% 29580: ~0.10% 29647: ~0.10% 29674: ~0.10% 29748: ~0.10% 29789: ~0.10% 29806: ~0.10% 29844: ~0.10% 29858: ~0.10% 29881: ~0.10% 29974: ~0.10% 29993: ~0.10% 30009: ~0.10% 30011: ~0.10% 30059: ~0.10% 30095: ~0.10% 30099: ~0.10% 30111: ~0.10% 30134: ~0.10% 30152: ~0.10% 30177: ~0.10% 30223: ~0.10% 30238: ~0.10% 30332: ~0.10% 30424: ~0.10% 30456: ~0.10% 30535: ~0.10% 30660: ~0.10% 30672: ~0.10% 30754: ~0.10% 30767: ~0.10% 30777: ~0.10% 30857: ~0.10% 30889: ~0.10% 30963: ~0.10% 30987: ~0.10% 31066: ~0.10% 31082: ~0.10% 31102: ~0.10% 31184: ~0.10% 31263: ~0.10% 31266: ~0.10% 31417: ~0.10% 31421: ~0.10% 31437: ~0.10% 31488: ~0.10% 31544: ~0.10% 31546: ~0.10% 31569: ~0.10% 31781: ~0.10% 31810: ~0.10% 31854: ~0.10% 31871: ~0.10% 31889: ~0.10% 31893: ~0.10% 31894: ~0.10% 31982: ~0.10% 32040: ~0.10% 32049: ~0.10% 32075: ~0.10% 32084: ~0.10% 32108: ~0.10% 32123: ~0.10% 32126: ~0.10% 32199: ~0.10% 32208: ~0.10% 32273: ~0.10% 32330: ~0.10% 32385: ~0.10% 32404: ~0.10% 32413: ~0.10% 32470: ~0.10% 32530: ~0.10% 32542: ~0.10% 32548: ~0.10% 32652: ~0.10% 32653: ~0.10% 32720: ~0.10% 32721: ~0.10% 32763: ~0.10% 32841: ~0.10% 32891: ~0.10% 33012: ~0.10% 33019: ~0.10% 33086: ~0.10% 33228: ~0.10% 33243: ~0.10% 33268: ~0.10% 33311: ~0.10% 33334: ~0.10% 33349: ~0.10% 33366: ~0.10% 33368: ~0.10% 33453: ~0.10% 33572: ~0.10% 33582: ~0.10% 33588: ~0.10% 33633: ~0.10% 33666: ~0.10% 33740: ~0.10% 33805: ~0.10% 33821: ~0.10% 33834: ~0.10% 33898: ~0.10% 33941: ~0.10% 33948: ~0.10% 34051: ~0.10% 34084: ~0.10% 34088: ~0.10% 34107: ~0.10% 34121: ~0.10% 34304: ~0.10% 34375: ~0.10% 34388: ~0.10% 34461: ~0.10% 34490: ~0.10% 34520: ~0.10% 34521: ~0.10% 34639: ~0.10% 34642: ~0.10% 34702: ~0.10% 34732: ~0.10% 34735: ~0.10% 34747: ~0.10% 34827: ~0.10% 34853: ~0.10% 34872: ~0.10% 34957: ~0.10% 34967: ~0.10% 35029: ~0.10% 35157: ~0.10% 35191: ~0.10% 35420: ~0.10% 35440: ~0.10% 35481: ~0.10% 35503: ~0.10% 35563: ~0.10% 35567: ~0.10% 35621: ~0.10% 35656: ~0.10% 35664: ~0.10% 35739: ~0.10% 35790: ~0.20% 35849: ~0.10% 35911: ~0.10% 35924: ~0.10% 35940: ~0.10% 35954: ~0.10% 36057: ~0.10% 36210: ~0.10% 36253: ~0.10% 36345: ~0.10% 36352: ~0.10% 36354: ~0.10% 36472: ~0.10% 36476: ~0.10% 36521: ~0.10% 36528: ~0.10% 36609: ~0.10% 36653: ~0.10% 36664: ~0.10% 36747: ~0.10% 36808: ~0.10% 36862: ~0.10% 36867: ~0.10% 36904: ~0.10% 36912: ~0.10% 36943: ~0.10% 37011: ~0.10% 37026: ~0.10% 37029: ~0.10% 37263: ~0.10% 37274: ~0.10% 37294: ~0.10% 37311: ~0.10% 37314: ~0.10% 37352: ~0.10% 37411: ~0.10% 37427: ~0.10% 37435: ~0.10% 37437: ~0.10% 37442: ~0.10% 37444: ~0.10% 37505: ~0.10%

• Samples:

  sentence_0	sentence_1	label
  Word: l i k ' á y b ' a l| Context: b w e n , n o n q e l i t i b ' i n l o q ' e ' q e l e n w e e l i k ' á y b ' a l| Translation: Bueno asi es compro cosas en el merrcado.	Morpheme: b ' i l | Gloss: INS	11799
  Word: b ' e l e j e b '| Context: r i b ' e l e j e b ' i i k ' ,| Translation: De los nueve meses.	Morpheme: e b ' | Gloss: NUM	34957
  Word: t r a| Context: D e s p w e s t r a j u u t| Translation: Después los jutes.	Morpheme: t r a | Gloss: DIM	9903

• Loss: main.LossLogger

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 64
• per_device_eval_batch_size: 64
• num_train_epochs: 50
• fp16: True
• multi_dataset_batch_sampler: round_robin

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 64
• per_device_eval_batch_size: 64
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 1
• eval_accumulation_steps: None
• torch_empty_cache_steps: None
• learning_rate: 5e-05
• weight_decay: 0.0
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1
• num_train_epochs: 50
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.0
• warmup_steps: 0
• log_level: passive
• log_level_replica: warning
• log_on_each_node: True
• logging_nan_inf_filter: True
• save_safetensors: True
• save_on_each_node: False
• save_only_model: False
• restore_callback_states_from_checkpoint: False
• no_cuda: False
• use_cpu: False
• use_mps_device: False
• seed: 42
• data_seed: None
• jit_mode_eval: False
• use_ipex: False
• bf16: False
• fp16: True
• fp16_opt_level: O1
• half_precision_backend: auto
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: None
• local_rank: 0
• ddp_backend: None
• tpu_num_cores: None
• tpu_metrics_debug: False
• debug: []
• dataloader_drop_last: False
• dataloader_num_workers: 0
• dataloader_prefetch_factor: None
• past_index: -1
• disable_tqdm: False
• remove_unused_columns: True
• label_names: None
• load_best_model_at_end: False
• ignore_data_skip: False
• fsdp: []
• fsdp_min_num_params: 0
• fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
• fsdp_transformer_layer_cls_to_wrap: None
• accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
• parallelism_config: None
• deepspeed: None
• label_smoothing_factor: 0.0
• optim: adamw_torch_fused
• optim_args: None
• adafactor: False
• group_by_length: False
• length_column_name: length
• ddp_find_unused_parameters: None
• ddp_bucket_cap_mb: None
• ddp_broadcast_buffers: False
• dataloader_pin_memory: True
• dataloader_persistent_workers: False
• skip_memory_metrics: True
• use_legacy_prediction_loop: False
• push_to_hub: False
• resume_from_checkpoint: None
• hub_model_id: None
• hub_strategy: every_save
• hub_private_repo: None
• hub_always_push: False
• hub_revision: None
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• include_inputs_for_metrics: False
• include_for_metrics: []
• eval_do_concat_batches: True
• fp16_backend: auto
• push_to_hub_model_id: None
• push_to_hub_organization: None
• mp_parameters:
• auto_find_batch_size: False
• full_determinism: False
• torchdynamo: None
• ray_scope: last
• ddp_timeout: 1800
• torch_compile: False
• torch_compile_backend: None
• torch_compile_mode: None
• include_tokens_per_second: False
• include_num_input_tokens_seen: False
• neftune_noise_alpha: None
• optim_target_modules: None
• batch_eval_metrics: False
• eval_on_start: False
• use_liger_kernel: False
• liger_kernel_config: None
• eval_use_gather_object: False
• average_tokens_across_devices: False
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: round_robin
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Click to expand

Epoch	Step	Training Loss	validation_cosine_ndcg@10
0.5291	500	1.0856	0.6291
1.0	945	-	0.6728
1.0582	1000	0.8022	0.6719
1.5873	1500	0.738	0.6894
2.0	1890	-	0.7071
2.1164	2000	0.707	0.7359
2.6455	2500	0.6807	0.7328
3.0	2835	-	0.7411
3.1746	3000	0.6588	0.7200
3.7037	3500	0.6526	0.7550
4.0	3780	-	0.7424
4.2328	4000	0.6243	0.7574
4.7619	4500	0.6239	0.7488
5.0	4725	-	0.7467
5.2910	5000	0.6009	0.7538
5.8201	5500	0.5984	0.7791
6.0	5670	-	0.7724
6.3492	6000	0.5802	0.7861
6.8783	6500	0.5768	0.8024
7.0	6615	-	0.7780
7.4074	7000	0.5731	0.7770
7.9365	7500	0.5665	0.7898
8.0	7560	-	0.7992
8.4656	8000	0.5451	0.7836
8.9947	8500	0.5611	0.7673
9.0	8505	-	0.7622
9.5238	9000	0.5368	0.7548
10.0	9450	-	0.7742
10.0529	9500	0.5409	0.7843
10.5820	10000	0.5277	0.7467
11.0	10395	-	0.7760
11.1111	10500	0.5323	0.7680
11.6402	11000	0.5261	0.7765
12.0	11340	-	0.7807
12.1693	11500	0.511	0.7805
12.6984	12000	0.5133	0.7817
13.0	12285	-	0.7476
13.2275	12500	0.5185	0.7724
13.7566	13000	0.5133	0.7709
14.0	13230	-	0.7621
14.2857	13500	0.5054	0.7880
14.8148	14000	0.5047	0.7531
15.0	14175	-	0.7696
15.3439	14500	0.4995	0.7659
15.8730	15000	0.4968	0.7615
16.0	15120	-	0.7686
16.4021	15500	0.4841	0.7659
16.9312	16000	0.4848	0.7814
17.0	16065	-	0.7766
17.4603	16500	0.4804	0.7737
17.9894	17000	0.4859	0.7809
18.0	17010	-	0.7789
18.5185	17500	0.4749	0.7832
19.0	17955	-	0.7872
19.0476	18000	0.481	0.7815
19.5767	18500	0.4718	0.7730
20.0	18900	-	0.7746
20.1058	19000	0.4721	0.7807
20.6349	19500	0.4702	0.7973
21.0	19845	-	0.7914
21.1640	20000	0.4671	0.7828
21.6931	20500	0.4707	0.7841
22.0	20790	-	0.7919
22.2222	21000	0.4716	0.7913
22.7513	21500	0.4616	0.7752
23.0	21735	-	0.7854
23.2804	22000	0.4657	0.7830
23.8095	22500	0.4557	0.7713
24.0	22680	-	0.7770
24.3386	23000	0.4683	0.7961
24.8677	23500	0.4537	0.7895
25.0	23625	-	0.7962
25.3968	24000	0.4507	0.7918
25.9259	24500	0.4598	0.7985
26.0	24570	-	0.7862
26.4550	25000	0.458	0.7906
26.9841	25500	0.4523	0.7897
27.0	25515	-	0.7879
27.5132	26000	0.4441	0.7969
28.0	26460	-	0.7845
28.0423	26500	0.4534	0.7974
28.5714	27000	0.4444	0.7906
29.0	27405	-	0.7952
29.1005	27500	0.4393	0.7933
29.6296	28000	0.4408	0.7928
30.0	28350	-	0.8044
30.1587	28500	0.4461	0.8052
30.6878	29000	0.4419	0.7956
31.0	29295	-	0.8047
31.2169	29500	0.4385	0.7957
31.7460	30000	0.446	0.8037
32.0	30240	-	0.7964
32.2751	30500	0.4354	0.7903
32.8042	31000	0.4348	0.7885
33.0	31185	-	0.8039
33.3333	31500	0.4342	0.7933
33.8624	32000	0.4363	0.7974
34.0	32130	-	0.7931
34.3915	32500	0.4245	0.8022
34.9206	33000	0.442	0.8016
35.0	33075	-	0.8013
35.4497	33500	0.4293	0.8052
35.9788	34000	0.4327	0.7908
36.0	34020	-	0.7941
36.5079	34500	0.4267	0.8078
37.0	34965	-	0.7965
37.0370	35000	0.4351	0.7942
37.5661	35500	0.4276	0.8026
38.0	35910	-	0.7991
38.0952	36000	0.4196	0.8083
38.6243	36500	0.4312	0.8014
39.0	36855	-	0.8098
39.1534	37000	0.4268	0.8152
39.6825	37500	0.4224	0.8111
40.0	37800	-	0.7984
40.2116	38000	0.4295	0.8095
40.7407	38500	0.4243	0.8162
41.0	38745	-	0.8097
41.2698	39000	0.4172	0.8139
41.7989	39500	0.4259	0.8128
42.0	39690	-	0.8138
42.3280	40000	0.4261	0.8116
42.8571	40500	0.4156	0.8046
43.0	40635	-	0.8078
43.3862	41000	0.4215	0.8185
43.9153	41500	0.4175	0.8074
44.0	41580	-	0.8053
44.4444	42000	0.4164	0.8108
44.9735	42500	0.4147	0.8054
45.0	42525	-	0.8065
45.5026	43000	0.4141	0.8145
46.0	43470	-	0.8119
46.0317	43500	0.4172	0.8111
46.5608	44000	0.4126	0.8127
47.0	44415	-	0.8155
47.0899	44500	0.4189	0.8130
47.6190	45000	0.4118	0.8163
48.0	45360	-	0.8165
48.1481	45500	0.4146	0.8173
48.6772	46000	0.4103	0.8181
49.0	46305	-	0.8189

Framework Versions

• Python: 3.11.5
• Sentence Transformers: 5.1.1
• Transformers: 4.56.2
• PyTorch: 2.8.0+cu128
• Accelerate: 1.10.1
• Datasets: 4.1.1
• Tokenizers: 0.22.1

Citation

BibTeX

Sentence Transformers

@inproceedings{reimers-2019-sentence-bert,
    title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
    author = "Reimers, Nils and Gurevych, Iryna",
    booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
    month = "11",
    year = "2019",
    publisher = "Association for Computational Linguistics",
    url = "https://arxiv.org/abs/1908.10084",
}