Model Details
Lim is a general text embedding model(chinese),We are continuously optimizing it.
History
『2023-12-22』Published lim_base_zh_v0 model
Usage (Sentence-Transformers)
Using this model becomes easy when you have sentence-transformers installed:
pip install -U sentence-transformers
Then you can use the model like this:
model_name="liujiarik/lim_base_zh"
from sentence_transformers import SentenceTransformer
sentences = ['我换手机号了', '如果我换手机怎么办?']
model = SentenceTransformer(model_name)
embeddings = model.encode(sentences)
print(embeddings)
- Downloads last month
- 332
Evaluation results
- accuracy on MTEB AmazonReviewsClassification (zh)test set self-reported46.666
- f1 on MTEB AmazonReviewsClassification (zh)test set self-reported43.881
- v_measure on MTEB CLSClusteringP2Ptest set self-reported33.555
- v_measure on MTEB CLSClusteringS2Stest set self-reported36.180
- map on MTEB CMedQAv1test set self-reported83.847
- mrr on MTEB CMedQAv1test set self-reported86.346
- map on MTEB CMedQAv2test set self-reported84.746
- mrr on MTEB CMedQAv2test set self-reported87.416
- cos_sim_accuracy on MTEB Cmnlivalidation set self-reported70.992
- cos_sim_ap on MTEB Cmnlivalidation set self-reported79.584
- cos_sim_f1 on MTEB Cmnlivalidation set self-reported73.012
- cos_sim_precision on MTEB Cmnlivalidation set self-reported67.091
- cos_sim_recall on MTEB Cmnlivalidation set self-reported80.079
- dot_accuracy on MTEB Cmnlivalidation set self-reported70.992
- dot_ap on MTEB Cmnlivalidation set self-reported79.587
- dot_f1 on MTEB Cmnlivalidation set self-reported73.012
- dot_precision on MTEB Cmnlivalidation set self-reported67.091
- dot_recall on MTEB Cmnlivalidation set self-reported80.079
- euclidean_accuracy on MTEB Cmnlivalidation set self-reported70.992
- euclidean_ap on MTEB Cmnlivalidation set self-reported79.584
- euclidean_f1 on MTEB Cmnlivalidation set self-reported73.012
- euclidean_precision on MTEB Cmnlivalidation set self-reported67.091
- euclidean_recall on MTEB Cmnlivalidation set self-reported80.079
- manhattan_accuracy on MTEB Cmnlivalidation set self-reported70.884
- manhattan_ap on MTEB Cmnlivalidation set self-reported79.423
- manhattan_f1 on MTEB Cmnlivalidation set self-reported72.725
- manhattan_precision on MTEB Cmnlivalidation set self-reported65.913
- manhattan_recall on MTEB Cmnlivalidation set self-reported81.108
- max_accuracy on MTEB Cmnlivalidation set self-reported70.992
- max_ap on MTEB Cmnlivalidation set self-reported79.587
- max_f1 on MTEB Cmnlivalidation set self-reported73.012
- accuracy on MTEB IFlyTekvalidation set self-reported47.341
- f1 on MTEB IFlyTekvalidation set self-reported35.500
- accuracy on MTEB JDReviewtest set self-reported85.666
- ap on MTEB JDReviewtest set self-reported53.038
- f1 on MTEB JDReviewtest set self-reported80.147
- map on MTEB MMarcoRerankingself-reported20.564
- mrr on MTEB MMarcoRerankingself-reported19.608
- accuracy on MTEB MassiveIntentClassification (zh-CN)test set self-reported72.384
- f1 on MTEB MassiveIntentClassification (zh-CN)test set self-reported70.330