SGPT-5.8B-weightedmean-msmarco-specb-bitfit
Usage
For usage instructions, refer to our codebase: https://github.com/Muennighoff/sgpt
Evaluation Results
For eval results, refer to our paper: https://arxiv.org/abs/2202.08904
Training
The model was trained with the parameters:
DataLoader:
torch.utils.data.dataloader.DataLoader
of length 249592 with parameters:
{'batch_size': 2, 'sampler': 'torch.utils.data.sampler.RandomSampler', 'batch_sampler': 'torch.utils.data.sampler.BatchSampler'}
Loss:
sentence_transformers.losses.MultipleNegativesRankingLoss.MultipleNegativesRankingLoss
with parameters:
{'scale': 20.0, 'similarity_fct': 'cos_sim'}
Parameters of the fit()-Method:
{
"epochs": 10,
"evaluation_steps": 0,
"evaluator": "NoneType",
"max_grad_norm": 1,
"optimizer_class": "<class 'transformers.optimization.AdamW'>",
"optimizer_params": {
"lr": 5e-05
},
"scheduler": "WarmupLinear",
"steps_per_epoch": null,
"warmup_steps": 1000,
"weight_decay": 0.01
}
Full Model Architecture
SentenceTransformer(
(0): Transformer({'max_seq_length': 300, 'do_lower_case': False}) with Transformer model: GPTJModel
(1): Pooling({'word_embedding_dimension': 4096, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': True, 'pooling_mode_lasttoken': False})
)
Citing & Authors
@article{muennighoff2022sgpt,
title={SGPT: GPT Sentence Embeddings for Semantic Search},
author={Muennighoff, Niklas},
journal={arXiv preprint arXiv:2202.08904},
year={2022}
}
- Downloads last month
- 4,126
Spaces using Muennighoff/SGPT-5.8B-weightedmean-msmarco-specb-bitfit 6
Evaluation results
- accuracy on MTEB AmazonCounterfactualClassification (en)test set self-reported69.224
- ap on MTEB AmazonCounterfactualClassification (en)test set self-reported32.047
- f1 on MTEB AmazonCounterfactualClassification (en)test set self-reported63.257
- accuracy on MTEB AmazonPolarityClassificationtest set self-reported71.261
- ap on MTEB AmazonPolarityClassificationtest set self-reported66.163
- f1 on MTEB AmazonPolarityClassificationtest set self-reported70.897
- accuracy on MTEB AmazonReviewsClassification (en)test set self-reported39.192
- f1 on MTEB AmazonReviewsClassification (en)test set self-reported38.581
- map_at_1 on MTEB ArguAnatest set self-reported27.312
- map_at_10 on MTEB ArguAnatest set self-reported42.620
- map_at_100 on MTEB ArguAnatest set self-reported43.707
- map_at_1000 on MTEB ArguAnatest set self-reported43.715
- map_at_3 on MTEB ArguAnatest set self-reported37.624
- map_at_5 on MTEB ArguAnatest set self-reported40.498
- mrr_at_1 on MTEB ArguAnatest set self-reported27.667
- mrr_at_10 on MTEB ArguAnatest set self-reported42.737
- mrr_at_100 on MTEB ArguAnatest set self-reported43.823
- mrr_at_1000 on MTEB ArguAnatest set self-reported43.831
- mrr_at_3 on MTEB ArguAnatest set self-reported37.743
- mrr_at_5 on MTEB ArguAnatest set self-reported40.616
- ndcg_at_1 on MTEB ArguAnatest set self-reported27.312
- ndcg_at_10 on MTEB ArguAnatest set self-reported51.375
- ndcg_at_100 on MTEB ArguAnatest set self-reported55.778
- ndcg_at_1000 on MTEB ArguAnatest set self-reported55.966
- ndcg_at_3 on MTEB ArguAnatest set self-reported41.087
- ndcg_at_5 on MTEB ArguAnatest set self-reported46.269
- precision_at_1 on MTEB ArguAnatest set self-reported27.312
- precision_at_10 on MTEB ArguAnatest set self-reported7.945
- precision_at_100 on MTEB ArguAnatest set self-reported0.982
- precision_at_1000 on MTEB ArguAnatest set self-reported0.100
- precision_at_3 on MTEB ArguAnatest set self-reported17.046
- precision_at_5 on MTEB ArguAnatest set self-reported12.745
- recall_at_1 on MTEB ArguAnatest set self-reported27.312
- recall_at_10 on MTEB ArguAnatest set self-reported79.445
- recall_at_100 on MTEB ArguAnatest set self-reported98.151
- recall_at_1000 on MTEB ArguAnatest set self-reported99.573
- recall_at_3 on MTEB ArguAnatest set self-reported51.138
- recall_at_5 on MTEB ArguAnatest set self-reported63.727
- v_measure on MTEB ArxivClusteringP2Ptest set self-reported45.590
- v_measure on MTEB ArxivClusteringS2Stest set self-reported38.864
- map on MTEB AskUbuntuDupQuestionstest set self-reported61.626
- mrr on MTEB AskUbuntuDupQuestionstest set self-reported75.833
- cos_sim_pearson on MTEB BIOSSEStest set self-reported89.961
- cos_sim_spearman on MTEB BIOSSEStest set self-reported86.250
- euclidean_pearson on MTEB BIOSSEStest set self-reported82.184
- euclidean_spearman on MTEB BIOSSEStest set self-reported80.148
- manhattan_pearson on MTEB BIOSSEStest set self-reported82.246
- manhattan_spearman on MTEB BIOSSEStest set self-reported80.004
- accuracy on MTEB Banking77Classificationtest set self-reported84.494
- f1 on MTEB Banking77Classificationtest set self-reported84.425
- v_measure on MTEB BiorxivClusteringP2Ptest set self-reported36.551