diff --git "a/README.md" "b/README.md" --- "a/README.md" +++ "b/README.md" @@ -6,7 +6,7 @@ tags: - generated_from_trainer - dataset_size:449904 - loss:CosineSimilarityLoss -base_model: CocoRoF/ModernBERT-SimCSE_v04 +base_model: x2bee/ModernBERT-SimCSE-multitask_v03 widget: - source_sentence: 우리는 움직이는 동행 우주 정지 좌표계에 비례하여 이동하고 있습니다 ... 약 371km / s에서 별자리 leo 쪽으로. " @@ -50,7 +50,7 @@ metrics: - pearson_max - spearman_max model-index: -- name: SentenceTransformer based on CocoRoF/ModernBERT-SimCSE_v04 +- name: SentenceTransformer based on x2bee/ModernBERT-SimCSE-multitask_v03 results: - task: type: semantic-similarity @@ -60,46 +60,46 @@ model-index: type: sts_dev metrics: - type: pearson_cosine - value: 0.7947107267431892 + value: 0.8352204711407503 name: Pearson Cosine - type: spearman_cosine - value: 0.8008029938863944 + value: 0.840621551045071 name: Spearman Cosine - type: pearson_euclidean - value: 0.7729649224022854 + value: 0.82566517917983 name: Pearson Euclidean - type: spearman_euclidean - value: 0.7731836226956725 + value: 0.8336282961362342 name: Spearman Euclidean - type: pearson_manhattan - value: 0.7728910393964163 + value: 0.8261231592330662 name: Pearson Manhattan - type: spearman_manhattan - value: 0.7732333197709114 + value: 0.8341401840450959 name: Spearman Manhattan - type: pearson_dot - value: 0.6023258275823691 + value: 0.736783144401919 name: Pearson Dot - type: spearman_dot - value: 0.5958009787049323 + value: 0.7201022209834164 name: Spearman Dot - type: pearson_max - value: 0.7947107267431892 + value: 0.8352204711407503 name: Pearson Max - type: spearman_max - value: 0.8008029938863944 + value: 0.840621551045071 name: Spearman Max --- -# SentenceTransformer based on CocoRoF/ModernBERT-SimCSE_v04 +# SentenceTransformer based on x2bee/ModernBERT-SimCSE-multitask_v03 -This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [CocoRoF/ModernBERT-SimCSE_v04](https://huggingface.co/CocoRoF/ModernBERT-SimCSE_v04) on the [misc_sts_pairs_v2_kor_kosimcse](https://huggingface.co/datasets/x2bee/misc_sts_pairs_v2_kor_kosimcse) dataset. 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. +This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [x2bee/ModernBERT-SimCSE-multitask_v03](https://huggingface.co/x2bee/ModernBERT-SimCSE-multitask_v03) on the [misc_sts_pairs_v2_kor_kosimcse](https://huggingface.co/datasets/x2bee/misc_sts_pairs_v2_kor_kosimcse) dataset. 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:** [CocoRoF/ModernBERT-SimCSE_v04](https://huggingface.co/CocoRoF/ModernBERT-SimCSE_v04) +- **Base model:** [x2bee/ModernBERT-SimCSE-multitask_v03](https://huggingface.co/x2bee/ModernBERT-SimCSE-multitask_v03) - **Maximum Sequence Length:** 512 tokens - **Output Dimensionality:** 768 dimensions - **Similarity Function:** Cosine Similarity @@ -139,7 +139,7 @@ Then you can load this model and run inference. from sentence_transformers import SentenceTransformer # Download from the 🤗 Hub -model = SentenceTransformer("CocoRoF/ModernBERT-SimCSE-multitask_v05") +model = SentenceTransformer("x2bee/ModernBERT-SimCSE-multitask_v03-beta") # Run inference sentences = [ '버스가 바쁜 길을 따라 운전한다.', @@ -191,16 +191,16 @@ You can finetune this model on your own dataset. | Metric | Value | |:-------------------|:-----------| -| pearson_cosine | 0.7947 | -| spearman_cosine | 0.8008 | -| pearson_euclidean | 0.773 | -| spearman_euclidean | 0.7732 | -| pearson_manhattan | 0.7729 | -| spearman_manhattan | 0.7732 | -| pearson_dot | 0.6023 | -| spearman_dot | 0.5958 | -| pearson_max | 0.7947 | -| **spearman_max** | **0.8008** | +| pearson_cosine | 0.8352 | +| spearman_cosine | 0.8406 | +| pearson_euclidean | 0.8257 | +| spearman_euclidean | 0.8336 | +| pearson_manhattan | 0.8261 | +| spearman_manhattan | 0.8341 | +| pearson_dot | 0.7368 | +| spearman_dot | 0.7201 | +| pearson_max | 0.8352 | +| **spearman_max** | **0.8406** |