Upload folder using huggingface_hub

README.md

@@ -4,37 +4,39 @@ tags:
 - sentence-similarity
 - feature-extraction
 - generated_from_trainer
-- dataset_size:5749
+- dataset_size:16729
 - loss:CosineSimilarityLoss
 base_model: hon9kon9ize/bert-large-cantonese-nli
 widget:
-- source_sentence: 有個男人彈緊豎琴。
+- source_sentence: 啲狗喺雪入面玩緊。
   sentences:
-  - 我鍾意將一心多用諗做快速轉換工作。
-  - 有個女人彈緊結他。
-  - 正如你原先個問題所講：維基百科嗰段嘢無話約瑟夫·P·甘迺迪買咗約翰·F·甘迺迪嘅勝選。
-- source_sentence: 一艘大型白色郵輪浮喺水上。
+  - 呢個係我成日覺得對一年級學生好有幫助嘅例子。
+  - 兩隻狗喺沙灘到玩緊。
+  - 喺Linux系統，我用Bibble，雖然有啲缺點，但係依家得呢個係比較專業嘅選擇。
+- source_sentence: 個女人整緊蛋。
   sentences:
-  - 有個女人喺台上唱歌。
-  - 有個男人吹緊笛。
-  - 一艘大型郵輪浮喺水上。
-- source_sentence: 個男人送緊蛋糕。
+  - 一班老人家圍住張飯枱影相。
+  - 有個男人向個女人唱歌。
+  - 個女人係度食嘢。
+- source_sentence: 一架電單車泊喺一幅畫滿城市景觀塗鴉嘅牆邊。
   sentences:
-  - 有個女人剝緊蝦。
-  - 有個男人係度調味鵪鶉。
-  - 一架紅色雙層巴士喺一條擠迫嘅街道上。
-- source_sentence: 我哋相對宇宙共同靜止參考系嘅速度係……每秒大約 371 公里，方向係朝住獅子座。」
+  - 夜晚，一架電單車泊喺一幅城市壁畫隔離。
+  - 一隻黑白相間嘅狗喺藍色嘅水到游水。
+  - 個細路仔頭髮豎晒起，係咁碌落藍色滑梯。
+- source_sentence: 有個男人孭住隻狗同埋一艘獨木舟。
   sentences:
-  - 間廳擺咗啲啡色嘅傢俬同埋一台平面電視。
-  - 冇一樣嘢係「靜止」嘅，除非係相對某啲其他物體先至係。
-  - 一個人喺水邊蹓狗。
-- source_sentence: 恆星喺恆星形成區形成，而恆星形成區本身係由分子雲演變而來。
+  - 隻狗孭住個男人喺獨木舟到。
+  - 我見我對孖仔就係咁：細路仔學說話嗰陣，都會自己發明啲獨特嘅方言。
+  - 「出汗就係出汗，你真係控制唔到。」
+- source_sentence: 一個細路女同一個細路仔喺度睇書。
   sentences:
-  - 一個金毛小朋友喺屋企前面吹緊喇叭表演，佢細佬喺隔離睇緊。
-  - 一架四驅車泥濘路面度行緊。
-  - 「可能喺銀河系以外都存在好似我哋呢個太陽系嘅星系。」
+  - 個女人孭住個BB。
+  - 有個男人彈緊結他。
+  - 一個大啲嘅小朋友玩緊公仔，望住窗外。
 datasets:
-- hon9kon9ize/yue-stsb
+  - hon9kon9ize/yue-stsb
+  - sentence-transformers/stsb
+  - C-MTEB/STSB
 pipeline_tag: sentence-similarity
 library_name: sentence-transformers
 metrics:
@@ -51,10 +53,10 @@ model-index:
       type: sts-dev
     metrics:
     - type: pearson_cosine
-      value: 0.825800689249711
+      value: 0.7983233550249502
       name: Pearson Cosine
     - type: spearman_cosine
-      value: 0.8262408727980405
+      value: 0.7996394101125816
       name: Spearman Cosine
   - task:
       type: semantic-similarity
@@ -64,16 +66,17 @@ model-index:
       type: sts-test
     metrics:
     - type: pearson_cosine
-      value: 0.7926469146205422
+      value: 0.7637579307526682
       name: Pearson Cosine
     - type: spearman_cosine
-      value: 0.7911184368054363
+      value: 0.7604840209490058
       name: Spearman Cosine
 ---
 
 # SentenceTransformer based on hon9kon9ize/bert-large-cantonese-nli
 
-This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [hon9kon9ize/bert-large-cantonese-nli](https://huggingface.co/hon9kon9ize/bert-large-cantonese-nli) on the [yue-stsb](https://huggingface.co/datasets/hon9kon9ize/yue-stsb) dataset. It maps sentences & paragraphs to a 1024-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 [hon9kon9ize/bert-large-cantonese-nli](https://huggingface.co/hon9kon9ize/bert-large-cantonese-nli) on the [yue-stsb](https://huggingface.co/datasets/hon9kon9ize/yue-stsb), [stsb](https://huggingface.co/datasets/sentence-transformers/stsb) and [C-MTEB/STSB](https://huggingface.co/datasets/C-MTEB/STSB) dataset. It maps sentences & paragraphs to a 1024-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
+
 
 ## Model Details
 
@@ -83,8 +86,7 @@ This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [h
 - **Maximum Sequence Length:** 512 tokens
 - **Output Dimensionality:** 1024 dimensions
 - **Similarity Function:** Cosine Similarity
-- **Training Dataset:**
-    - [yue-stsb](https://huggingface.co/datasets/hon9kon9ize/yue-stsb)
+<!-- - **Training Dataset:** Unknown -->
 <!-- - **Language:** Unknown -->
 <!-- - **License:** Unknown -->
 
@@ -121,9 +123,9 @@ from sentence_transformers import SentenceTransformer
 model = SentenceTransformer("sentence_transformers_model_id")
 # Run inference
 sentences = [
-    '恆星喺恆星形成區形成，而恆星形成區本身係由分子雲演變而來。',
-    '「可能喺銀河系以外都存在好似我哋呢個太陽系嘅星系。」',
-    '一架四驅車泥濘路面度行緊。',
+    '一個細路女同一個細路仔喺度睇書。',
+    '一個大啲嘅小朋友玩緊公仔，望住窗外。',
+    '有個男人彈緊結他。',
 ]
 embeddings = model.encode(sentences)
 print(embeddings.shape)
@@ -170,8 +172,8 @@ You can finetune this model on your own dataset.
 
 | Metric              | sts-dev    | sts-test   |
 |:--------------------|:-----------|:-----------|
-| pearson_cosine      | 0.8258     | 0.7926     |
-| **spearman_cosine** | **0.8262** | **0.7911** |
+| pearson_cosine      | 0.7983     | 0.7638     |
+| **spearman_cosine** | **0.7996** | **0.7605** |
 
 <!--
 ## Bias, Risks and Limitations
@@ -212,12 +214,32 @@ You can finetune this model on your own dataset.
   }
   ```
 
+* Size: 16,729 training samples
+* Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>score</code>
+* Approximate statistics based on the first 1000 samples:
+  |         | sentence1                                                                         | sentence2                                                                         | score                                                          |
+  |:--------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:---------------------------------------------------------------|
+  | type    | string                                                                            | string                                                                            | float                                                          |
+  | details | <ul><li>min: 5 tokens</li><li>mean: 20.29 tokens</li><li>max: 74 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 20.36 tokens</li><li>max: 76 tokens</li></ul> | <ul><li>min: 0.0</li><li>mean: 0.52</li><li>max: 1.0</li></ul> |
+* Samples:
+  | sentence1                                                       | sentence2                                                | score             |
+  |:----------------------------------------------------------------|:---------------------------------------------------------|:------------------|
+  | <code>奧巴馬登記咗參加奧巴馬醫保。 </code>                                    | <code>美國人爭住喺限期前登記參加奧巴馬醫保計劃，</code>                       | <code>0.24</code> |
+  | <code>Search ends for missing asylum-seekers</code>             | <code>Search narrowed for missing man</code>             | <code>0.28</code> |
+  | <code>檢察官喺五月突然轉軚，要求公開驗屍報告，因為有利於辯方嘅康納·彼得森驗屍報告部分內容已經洩露畀媒體。</code> | <code>佢哋要求公開驗屍報告，因為彼得森腹中胎兒嘅驗屍報告中，對辯方有利嘅部分已經洩露俾傳媒。</code> | <code>0.8</code>  |
+* Loss: [<code>CosineSimilarityLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#cosinesimilarityloss) with these parameters:
+  ```json
+  {
+      "loss_fct": "torch.nn.modules.loss.MSELoss"
+  }
+  ```
+
 ### Evaluation Dataset
 
-#### yue-stsb
+#### Unnamed Dataset
 
-* Dataset: [yue-stsb](https://huggingface.co/datasets/hon9kon9ize/yue-stsb) at [40cea5d](https://huggingface.co/datasets/hon9kon9ize/yue-stsb/tree/40cea5d8e9d1aeb1498816d90d1e417bafcc96a8)
-* Size: 1,500 evaluation samples
+
+* Size: 4,458 evaluation samples
 * Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>score</code>
 * Approximate statistics based on the first 1000 samples:
   |         | sentence1                                                                         | sentence2                                                                         | score                                                          |
@@ -370,8 +392,12 @@ You can finetune this model on your own dataset.
 ### Training Logs
 | Epoch  | Step | Training Loss | Validation Loss | sts-dev_spearman_cosine | sts-test_spearman_cosine |
 |:------:|:----:|:-------------:|:---------------:|:-----------------------:|:------------------------:|
-| 2.2222 | 100  | 0.0299        | 0.0312          | 0.8262                  | -                        |
-| 4.0    | 180  | -             | -               | -                       | 0.7911                   |
+| 0.7634 | 100  | 0.0549        | 0.0403          | 0.7895                  | -                        |
+| 1.5267 | 200  | 0.027         | 0.0368          | 0.7941                  | -                        |
+| 2.2901 | 300  | 0.0187        | 0.0349          | 0.7968                  | -                        |
+| 3.0534 | 400  | 0.0119        | 0.0354          | 0.8004                  | -                        |
+| 3.8168 | 500  | 0.0076        | 0.0359          | 0.7996                  | -                        |
+| 4.0    | 524  | -             | -               | -                       | 0.7605                   |
 
 
 ### Framework Versions

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9c1d5859a25e08b31120af6f1f2ababea81234d1ef546b566846773dc944d964
+oid sha256:2bb63b9fa78a3638d714e24acbe696ea02301ea7dfbba5a232a65f11cdaac6c1
 size 1304182568