RomainDarous commited on
Commit
0933eb7
·
verified ·
1 Parent(s): af93f3b

Add new SentenceTransformer model

Browse files
1_MultiHeadGeneralizedPooling/config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "sentence_dim": 768,
3
+ "token_dim": 768,
4
+ "num_heads": 8,
5
+ "initialize": 2,
6
+ "pooling_type": 1
7
+ }
1_MultiHeadGeneralizedPooling/multihead_pooling_weights.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bc4dbbeca20916e5a03895e93e5e612c329268f94595c29ddc43dd940a5184f9
3
+ size 2367886
2_Dense/config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"in_features": 768, "out_features": 512, "bias": true, "activation_function": "torch.nn.modules.activation.Tanh"}
2_Dense/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:078900f5c5dac9c746fab24f68d4d24b8be46567a1f7bdd2a53b9d648e72a6bc
3
+ size 1575072
README.md ADDED
@@ -0,0 +1,446 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - sentence-transformers
4
+ - sentence-similarity
5
+ - feature-extraction
6
+ - generated_from_trainer
7
+ - dataset_size:4460010
8
+ - loss:CoSENTLoss
9
+ base_model: RomainDarous/directTwoEpoch_dotProductPooling_randomInit_mistranslationModel
10
+ widget:
11
+ - source_sentence: Malformed target specific variable definition
12
+ sentences:
13
+ - Hedefe özgü değişken tanımı bozuk
14
+ - Kan alle data in die gids lees
15
+ - "слава Украине! героям слава!\uFEFF"
16
+ - source_sentence: Can't write an inode bitmap
17
+ sentences:
18
+ - Skontrolujte stav aktualizácií alebo to skúste znova neskôr.
19
+ - Malsukcesis skribi i nodan bitmapon
20
+ - Zastępuje wersję GL obsługiwaną przez sterownik
21
+ - source_sentence: Optimize soft proofing color transformations
22
+ sentences:
23
+ - 'arkadaslar biz artik her an kirmizi kart yiyecek,bencil,pas yapamayan,isabetsiz
24
+ orta yapani istemiyoruz. sozde efsaneniz bu sezon Besiktasa en cok zarar verenlerden
25
+ biriydi. kendini dusunmeden once Besiktasi dusunecek adam lazim bize. o yuzden
26
+ #GoHomeQuaresma'
27
+ - Yav bizim dedikodusunu yaptığımız insanın bile bi vizyonu var. Senin hakkında
28
+ neden oturup konuşalım?
29
+ - Ik ben een transgender.
30
+ - source_sentence: 'Pass 1: Checking @is, @bs, and sizes'
31
+ sentences:
32
+ - Bu adam cidden kurabiye gibi ben bunu çayın yanında yerim
33
+ - sagnat. errada. invisible. justificació. idioma
34
+ - Wilt u echt de primaire sleutel verplaatsen? (j N)
35
+ - source_sentence: Search for matching log entries
36
+ sentences:
37
+ - quem te lembra? caralho tô assustada aqui kkkkk
38
+ - sendotasunik gabeko\ egoera bistaratuko den ala ez adierazten du
39
+ - En aquest cas, hem d'incloure les imatges del contenidor )sr iov per a càrregues
40
+ de treball de telco (per exemple, com a referència, es podrien obtenir des de
41
+ valors de helm chart)
42
+ datasets:
43
+ - RomainDarous/corrupted_os_by_language
44
+ pipeline_tag: sentence-similarity
45
+ library_name: sentence-transformers
46
+ metrics:
47
+ - pearson_cosine
48
+ - spearman_cosine
49
+ model-index:
50
+ - name: SentenceTransformer based on RomainDarous/directTwoEpoch_dotProductPooling_randomInit_mistranslationModel
51
+ results:
52
+ - task:
53
+ type: semantic-similarity
54
+ name: Semantic Similarity
55
+ dataset:
56
+ name: sts eval
57
+ type: sts-eval
58
+ metrics:
59
+ - type: pearson_cosine
60
+ value: 0.9719312887381778
61
+ name: Pearson Cosine
62
+ - type: spearman_cosine
63
+ value: 0.8648678348337416
64
+ name: Spearman Cosine
65
+ - task:
66
+ type: semantic-similarity
67
+ name: Semantic Similarity
68
+ dataset:
69
+ name: sts test
70
+ type: sts-test
71
+ metrics:
72
+ - type: pearson_cosine
73
+ value: 0.971964567319277
74
+ name: Pearson Cosine
75
+ - type: spearman_cosine
76
+ value: 0.8648754707224892
77
+ name: Spearman Cosine
78
+ ---
79
+
80
+ # SentenceTransformer based on RomainDarous/directTwoEpoch_dotProductPooling_randomInit_mistranslationModel
81
+
82
+ This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [RomainDarous/directTwoEpoch_dotProductPooling_randomInit_mistranslationModel](https://huggingface.co/RomainDarous/directTwoEpoch_dotProductPooling_randomInit_mistranslationModel) on the [corrupted_open_os_by_language](https://huggingface.co/datasets/RomainDarous/corrupted_os_by_language) dataset. It maps sentences & paragraphs to a 512-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
83
+
84
+ ## Model Details
85
+
86
+ ### Model Description
87
+ - **Model Type:** Sentence Transformer
88
+ - **Base model:** [RomainDarous/directTwoEpoch_dotProductPooling_randomInit_mistranslationModel](https://huggingface.co/RomainDarous/directTwoEpoch_dotProductPooling_randomInit_mistranslationModel) <!-- at revision edb80d3114270f0a27e85bdf54df7d570212a283 -->
89
+ - **Maximum Sequence Length:** 128 tokens
90
+ - **Output Dimensionality:** 512 dimensions
91
+ - **Similarity Function:** Cosine Similarity
92
+ - **Training Dataset:**
93
+ - [corrupted_open_os_by_language](https://huggingface.co/datasets/RomainDarous/corrupted_os_by_language)
94
+ <!-- - **Language:** Unknown -->
95
+ <!-- - **License:** Unknown -->
96
+
97
+ ### Model Sources
98
+
99
+ - **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
100
+ - **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
101
+ - **HF中国镜像站:** [Sentence Transformers on HF中国镜像站](https://huggingface.co/models?library=sentence-transformers)
102
+
103
+ ### Full Model Architecture
104
+
105
+ ```
106
+ SentenceTransformer(
107
+ (0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: DistilBertModel
108
+ (1): MultiHeadGeneralizedPooling(
109
+ (Q): ModuleList(
110
+ (0-7): 8 x Linear(in_features=96, out_features=1, bias=True)
111
+ )
112
+ (P_K): ModuleList(
113
+ (0-7): 8 x Linear(in_features=768, out_features=96, bias=True)
114
+ )
115
+ )
116
+ (2): Dense({'in_features': 768, 'out_features': 512, 'bias': True, 'activation_function': 'torch.nn.modules.activation.Tanh'})
117
+ )
118
+ ```
119
+
120
+ ## Usage
121
+
122
+ ### Direct Usage (Sentence Transformers)
123
+
124
+ First install the Sentence Transformers library:
125
+
126
+ ```bash
127
+ pip install -U sentence-transformers
128
+ ```
129
+
130
+ Then you can load this model and run inference.
131
+ ```python
132
+ from sentence_transformers import SentenceTransformer
133
+
134
+ # Download from the 🤗 Hub
135
+ model = SentenceTransformer("RomainDarous/directThreeEpoch_dotProductPooling_randomInit_mistranslationModel")
136
+ # Run inference
137
+ sentences = [
138
+ 'Search for matching log entries',
139
+ 'quem te lembra? caralho tô assustada aqui kkkkk',
140
+ 'sendotasunik gabeko\\ egoera bistaratuko den ala ez adierazten du',
141
+ ]
142
+ embeddings = model.encode(sentences)
143
+ print(embeddings.shape)
144
+ # [3, 512]
145
+
146
+ # Get the similarity scores for the embeddings
147
+ similarities = model.similarity(embeddings, embeddings)
148
+ print(similarities.shape)
149
+ # [3, 3]
150
+ ```
151
+
152
+ <!--
153
+ ### Direct Usage (Transformers)
154
+
155
+ <details><summary>Click to see the direct usage in Transformers</summary>
156
+
157
+ </details>
158
+ -->
159
+
160
+ <!--
161
+ ### Downstream Usage (Sentence Transformers)
162
+
163
+ You can finetune this model on your own dataset.
164
+
165
+ <details><summary>Click to expand</summary>
166
+
167
+ </details>
168
+ -->
169
+
170
+ <!--
171
+ ### Out-of-Scope Use
172
+
173
+ *List how the model may foreseeably be misused and address what users ought not to do with the model.*
174
+ -->
175
+
176
+ ## Evaluation
177
+
178
+ ### Metrics
179
+
180
+ #### Semantic Similarity
181
+
182
+ * Datasets: `sts-eval` and `sts-test`
183
+ * Evaluated with [<code>EmbeddingSimilarityEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.EmbeddingSimilarityEvaluator)
184
+
185
+ | Metric | sts-eval | sts-test |
186
+ |:--------------------|:-----------|:-----------|
187
+ | pearson_cosine | 0.9719 | 0.972 |
188
+ | **spearman_cosine** | **0.8649** | **0.8649** |
189
+
190
+ <!--
191
+ ## Bias, Risks and Limitations
192
+
193
+ *What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
194
+ -->
195
+
196
+ <!--
197
+ ### Recommendations
198
+
199
+ *What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
200
+ -->
201
+
202
+ ## Training Details
203
+
204
+ ### Training Dataset
205
+
206
+ #### corrupted_open_os_by_language
207
+
208
+ * Dataset: [corrupted_open_os_by_language](https://huggingface.co/datasets/RomainDarous/corrupted_os_by_language) at [9d25780](https://huggingface.co/datasets/RomainDarous/corrupted_os_by_language/tree/9d25780e2032b1e8f06af6a4ff55124d7a930c3c)
209
+ * Size: 4,460,010 training samples
210
+ * Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>score</code>
211
+ * Approximate statistics based on the first 1000 samples:
212
+ | | sentence1 | sentence2 | score |
213
+ |:--------|:-----------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------|:------------------------------------------------|
214
+ | type | string | string | int |
215
+ | details | <ul><li>min: 6 tokens</li><li>mean: 18.49 tokens</li><li>max: 128 tokens</li></ul> | <ul><li>min: 3 tokens</li><li>mean: 30.77 tokens</li><li>max: 128 tokens</li></ul> | <ul><li>0: ~50.60%</li><li>1: ~49.40%</li></ul> |
216
+ * Samples:
217
+ | sentence1 | sentence2 | score |
218
+ |:--------------------------------------------------------------------------------------------|:-----------------------------------------------------------------------|:---------------|
219
+ | <code>Check spelling. Print the document. Show completion window. General. Show help</code> | <code>Kontrolli õigekirja. присоединяюсь. </code> | <code>0</code> |
220
+ | <code>EXIF not supported for this file format.</code> | <code>Šiam failo formatui EXIF nepalaikomas.</code> | <code>1</code> |
221
+ | <code>This package includes the documentation for texlive everyhook</code> | <code>Paket ini menyertakan dokumentasi untuk texlive everyhook</code> | <code>1</code> |
222
+ * Loss: [<code>CoSENTLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#cosentloss) with these parameters:
223
+ ```json
224
+ {
225
+ "scale": 20.0,
226
+ "similarity_fct": "pairwise_cos_sim"
227
+ }
228
+ ```
229
+
230
+ ### Evaluation Dataset
231
+
232
+ #### corrupted_open_os_by_language
233
+
234
+ * Dataset: [corrupted_open_os_by_language](https://huggingface.co/datasets/RomainDarous/corrupted_os_by_language) at [9d25780](https://huggingface.co/datasets/RomainDarous/corrupted_os_by_language/tree/9d25780e2032b1e8f06af6a4ff55124d7a930c3c)
235
+ * Size: 4,460,010 evaluation samples
236
+ * Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>score</code>
237
+ * Approximate statistics based on the first 1000 samples:
238
+ | | sentence1 | sentence2 | score |
239
+ |:--------|:-----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:------------------------------------------------|
240
+ | type | string | string | int |
241
+ | details | <ul><li>min: 5 tokens</li><li>mean: 17.92 tokens</li><li>max: 128 tokens</li></ul> | <ul><li>min: 4 tokens</li><li>mean: 31.1 tokens</li><li>max: 128 tokens</li></ul> | <ul><li>0: ~50.60%</li><li>1: ~49.40%</li></ul> |
242
+ * Samples:
243
+ | sentence1 | sentence2 | score |
244
+ |:----------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------|
245
+ | <code>Could not identify the current seat.</code> | <code> 天天花着男人的钱还这这创造新词汇男权你可真牛批,你也就这一出了一问男权,就说是我是吧,到现在我也没听到你给我们讲的男权,你也就是在网上喷喷,现实走道都不敢探头自卑,你现实要把你女权的劲拿出来总低啥头,您老应该去国家教育局把男权加上是吧,你们女权天天说自己生活不好没地位,给你们地位了你们能干啥?用你们的女权打到全世界男性是吧,能相出男权这一词您老也是人才呀,是不是庆幸自己是个女的,活在自己想想的世界里不觉得孤单吗,假象有男权是吧,自己假象和男权还说自己不是田园女权,田园女权能连自己都骂说自己妈是驴爸是大鼎的也是奇葩呀,那我们国家大肆宣扬过你们这么田园女权吗,国家要的是女性人群自主自理,你们可好看看你们女权干的啥事,给你们女权地位高了,看看你们女权干的事n绿地集团高管怎么都不说呀,人家可是有钱有地位,也不是我们说三从四德洗衣做饭你们女权会吗?,那我问问你们女权干过啥惊天大事,还甩锅给孔子,还封建社会,那我问问你们女权在福利面前为啥说自己是女性呀不是社会主义社会吗不应该男女平等吗,天天自己也不知道是不是抱个手机天天欧巴欧巴,你家那位要是不陪你看一会就会问你是不是不爱我了是吧大姐,您老也就赚这白菜钱操心国家事,中国五千年的历史被您老一句否决,还嘲讽人家日本女性,好意思说自己不是女权,三从四德流传这么久到您这变成日本文化了,我就想问问男权您老是怎么想的,那你问孔子老人家呗为什么女人要三从四德,我说的是女权你干嘛自己对号入座,连中华人民传承的东西都不认跟我这谈男权,还男权您老给我举个例子呗,让我们男权听听都是h啥,这些不都是你们女权的标准吗?,还男权,您老醒醒吧这里是现实,不是你的公主世界,总觉得自己多么多么重要,地球没你是不能转了还是人类要灭亡呀,我真的想问一句你给我找一条男权的新闻,咋了我们男人不能提女权呗你老授权了呗,那我们谈论田园女权你老对号入座干嘛,天天过节要礼物,还嫌弃自己男朋友没有钱,我寻思你找个有钱人包养你呗,对了有钱人怎么可能看上你这种女权的呢,还要孩子跟女方姓我也没看见你没跟你妈姓呀,年年过节男人给你们送礼物你们女人给男人送过礼物吗?,一问我不是陪着他吗我对他说我爱你了这不是最好的礼物吗?,男人只要不送礼物就是不爱你们了呗,人家国际女权讲的男人能做的我们女人也能做,田园女权男人能做的我们女人为啥要做,还男权我笑了,以前结婚几头牛换个衣服原装的,现在几十万彩...</code> | <code>0</code> |
246
+ | <code>Undoing Date and Time Adjustment</code> | <code>正在取消日期和时间调整</code> | <code>1</code> |
247
+ | <code>Dependency package for gsl_2_6 gnu hpc</code> | <code>Pacotes de desenvolvimento do KDE</code> | <code>1</code> |
248
+ * Loss: [<code>CoSENTLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#cosentloss) with these parameters:
249
+ ```json
250
+ {
251
+ "scale": 20.0,
252
+ "similarity_fct": "pairwise_cos_sim"
253
+ }
254
+ ```
255
+
256
+ ### Training Hyperparameters
257
+ #### Non-Default Hyperparameters
258
+
259
+ - `eval_strategy`: steps
260
+ - `per_device_train_batch_size`: 64
261
+ - `per_device_eval_batch_size`: 64
262
+ - `num_train_epochs`: 1
263
+ - `warmup_ratio`: 0.1
264
+
265
+ #### All Hyperparameters
266
+ <details><summary>Click to expand</summary>
267
+
268
+ - `overwrite_output_dir`: False
269
+ - `do_predict`: False
270
+ - `eval_strategy`: steps
271
+ - `prediction_loss_only`: True
272
+ - `per_device_train_batch_size`: 64
273
+ - `per_device_eval_batch_size`: 64
274
+ - `per_gpu_train_batch_size`: None
275
+ - `per_gpu_eval_batch_size`: None
276
+ - `gradient_accumulation_steps`: 1
277
+ - `eval_accumulation_steps`: None
278
+ - `torch_empty_cache_steps`: None
279
+ - `learning_rate`: 5e-05
280
+ - `weight_decay`: 0.0
281
+ - `adam_beta1`: 0.9
282
+ - `adam_beta2`: 0.999
283
+ - `adam_epsilon`: 1e-08
284
+ - `max_grad_norm`: 1.0
285
+ - `num_train_epochs`: 1
286
+ - `max_steps`: -1
287
+ - `lr_scheduler_type`: linear
288
+ - `lr_scheduler_kwargs`: {}
289
+ - `warmup_ratio`: 0.1
290
+ - `warmup_steps`: 0
291
+ - `log_level`: passive
292
+ - `log_level_replica`: warning
293
+ - `log_on_each_node`: True
294
+ - `logging_nan_inf_filter`: True
295
+ - `save_safetensors`: True
296
+ - `save_on_each_node`: False
297
+ - `save_only_model`: False
298
+ - `restore_callback_states_from_checkpoint`: False
299
+ - `no_cuda`: False
300
+ - `use_cpu`: False
301
+ - `use_mps_device`: False
302
+ - `seed`: 42
303
+ - `data_seed`: None
304
+ - `jit_mode_eval`: False
305
+ - `use_ipex`: False
306
+ - `bf16`: False
307
+ - `fp16`: False
308
+ - `fp16_opt_level`: O1
309
+ - `half_precision_backend`: auto
310
+ - `bf16_full_eval`: False
311
+ - `fp16_full_eval`: False
312
+ - `tf32`: None
313
+ - `local_rank`: 0
314
+ - `ddp_backend`: None
315
+ - `tpu_num_cores`: None
316
+ - `tpu_metrics_debug`: False
317
+ - `debug`: []
318
+ - `dataloader_drop_last`: False
319
+ - `dataloader_num_workers`: 0
320
+ - `dataloader_prefetch_factor`: None
321
+ - `past_index`: -1
322
+ - `disable_tqdm`: False
323
+ - `remove_unused_columns`: True
324
+ - `label_names`: None
325
+ - `load_best_model_at_end`: False
326
+ - `ignore_data_skip`: False
327
+ - `fsdp`: []
328
+ - `fsdp_min_num_params`: 0
329
+ - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
330
+ - `fsdp_transformer_layer_cls_to_wrap`: None
331
+ - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
332
+ - `deepspeed`: None
333
+ - `label_smoothing_factor`: 0.0
334
+ - `optim`: adamw_torch
335
+ - `optim_args`: None
336
+ - `adafactor`: False
337
+ - `group_by_length`: False
338
+ - `length_column_name`: length
339
+ - `ddp_find_unused_parameters`: None
340
+ - `ddp_bucket_cap_mb`: None
341
+ - `ddp_broadcast_buffers`: False
342
+ - `dataloader_pin_memory`: True
343
+ - `dataloader_persistent_workers`: False
344
+ - `skip_memory_metrics`: True
345
+ - `use_legacy_prediction_loop`: False
346
+ - `push_to_hub`: False
347
+ - `resume_from_checkpoint`: None
348
+ - `hub_model_id`: None
349
+ - `hub_strategy`: every_save
350
+ - `hub_private_repo`: None
351
+ - `hub_always_push`: False
352
+ - `gradient_checkpointing`: False
353
+ - `gradient_checkpointing_kwargs`: None
354
+ - `include_inputs_for_metrics`: False
355
+ - `include_for_metrics`: []
356
+ - `eval_do_concat_batches`: True
357
+ - `fp16_backend`: auto
358
+ - `push_to_hub_model_id`: None
359
+ - `push_to_hub_organization`: None
360
+ - `mp_parameters`:
361
+ - `auto_find_batch_size`: False
362
+ - `full_determinism`: False
363
+ - `torchdynamo`: None
364
+ - `ray_scope`: last
365
+ - `ddp_timeout`: 1800
366
+ - `torch_compile`: False
367
+ - `torch_compile_backend`: None
368
+ - `torch_compile_mode`: None
369
+ - `dispatch_batches`: None
370
+ - `split_batches`: None
371
+ - `include_tokens_per_second`: False
372
+ - `include_num_input_tokens_seen`: False
373
+ - `neftune_noise_alpha`: None
374
+ - `optim_target_modules`: None
375
+ - `batch_eval_metrics`: False
376
+ - `eval_on_start`: False
377
+ - `use_liger_kernel`: False
378
+ - `eval_use_gather_object`: False
379
+ - `average_tokens_across_devices`: False
380
+ - `prompts`: None
381
+ - `batch_sampler`: batch_sampler
382
+ - `multi_dataset_batch_sampler`: proportional
383
+
384
+ </details>
385
+
386
+ ### Training Logs
387
+ | Epoch | Step | Training Loss | corrupted open os by language loss | sts-eval_spearman_cosine | sts-test_spearman_cosine |
388
+ |:-----:|:-----:|:-------------:|:----------------------------------:|:------------------------:|:------------------------:|
389
+ | 1.0 | 55751 | 0.4292 | 0.6757 | 0.8649 | - |
390
+ | -1 | -1 | - | - | - | 0.8649 |
391
+
392
+
393
+ ### Framework Versions
394
+ - Python: 3.10.13
395
+ - Sentence Transformers: 3.4.1
396
+ - Transformers: 4.48.2
397
+ - PyTorch: 2.1.2+cu121
398
+ - Accelerate: 1.3.0
399
+ - Datasets: 2.16.1
400
+ - Tokenizers: 0.21.0
401
+
402
+ ## Citation
403
+
404
+ ### BibTeX
405
+
406
+ #### Sentence Transformers
407
+ ```bibtex
408
+ @inproceedings{reimers-2019-sentence-bert,
409
+ title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
410
+ author = "Reimers, Nils and Gurevych, Iryna",
411
+ booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
412
+ month = "11",
413
+ year = "2019",
414
+ publisher = "Association for Computational Linguistics",
415
+ url = "https://arxiv.org/abs/1908.10084",
416
+ }
417
+ ```
418
+
419
+ #### CoSENTLoss
420
+ ```bibtex
421
+ @online{kexuefm-8847,
422
+ title={CoSENT: A more efficient sentence vector scheme than Sentence-BERT},
423
+ author={Su Jianlin},
424
+ year={2022},
425
+ month={Jan},
426
+ url={https://kexue.fm/archives/8847},
427
+ }
428
+ ```
429
+
430
+ <!--
431
+ ## Glossary
432
+
433
+ *Clearly define terms in order to be accessible across audiences.*
434
+ -->
435
+
436
+ <!--
437
+ ## Model Card Authors
438
+
439
+ *Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
440
+ -->
441
+
442
+ <!--
443
+ ## Model Card Contact
444
+
445
+ *Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
446
+ -->
config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "RomainDarous/directTwoEpoch_dotProductPooling_randomInit_mistranslationModel",
3
+ "activation": "gelu",
4
+ "architectures": [
5
+ "DistilBertModel"
6
+ ],
7
+ "attention_dropout": 0.1,
8
+ "dim": 768,
9
+ "dropout": 0.1,
10
+ "hidden_dim": 3072,
11
+ "initializer_range": 0.02,
12
+ "max_position_embeddings": 512,
13
+ "model_type": "distilbert",
14
+ "n_heads": 12,
15
+ "n_layers": 6,
16
+ "output_hidden_states": true,
17
+ "output_past": true,
18
+ "pad_token_id": 0,
19
+ "qa_dropout": 0.1,
20
+ "seq_classif_dropout": 0.2,
21
+ "sinusoidal_pos_embds": false,
22
+ "tie_weights_": true,
23
+ "torch_dtype": "float32",
24
+ "transformers_version": "4.48.2",
25
+ "vocab_size": 119547
26
+ }
config_sentence_transformers.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "__version__": {
3
+ "sentence_transformers": "3.4.1",
4
+ "transformers": "4.48.2",
5
+ "pytorch": "2.1.2+cu121"
6
+ },
7
+ "prompts": {},
8
+ "default_prompt_name": null,
9
+ "similarity_fn_name": "cosine"
10
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:78374707fbec261856d4beaf932cd3ebd056e841dc6c9fe53d048ade0a241ab3
3
+ size 538947416
modules.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "idx": 0,
4
+ "name": "0",
5
+ "path": "",
6
+ "type": "sentence_transformers.models.Transformer"
7
+ },
8
+ {
9
+ "idx": 1,
10
+ "name": "1",
11
+ "path": "1_MultiHeadGeneralizedPooling",
12
+ "type": "sentence_generalized_pooling.multihead_generalized_pooling.MultiHeadGeneralizedPooling"
13
+ },
14
+ {
15
+ "idx": 2,
16
+ "name": "2",
17
+ "path": "2_Dense",
18
+ "type": "sentence_transformers.models.Dense"
19
+ }
20
+ ]
sentence_bert_config.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "max_seq_length": 128,
3
+ "do_lower_case": false
4
+ }
special_tokens_map.json ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cls_token": {
3
+ "content": "[CLS]",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "mask_token": {
10
+ "content": "[MASK]",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "[PAD]",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "sep_token": {
24
+ "content": "[SEP]",
25
+ "lstrip": false,
26
+ "normalized": false,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "unk_token": {
31
+ "content": "[UNK]",
32
+ "lstrip": false,
33
+ "normalized": false,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ }
37
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,67 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "[PAD]",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "100": {
12
+ "content": "[UNK]",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "101": {
20
+ "content": "[CLS]",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "102": {
28
+ "content": "[SEP]",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "103": {
36
+ "content": "[MASK]",
37
+ "lstrip": false,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "clean_up_tokenization_spaces": false,
45
+ "cls_token": "[CLS]",
46
+ "do_basic_tokenize": true,
47
+ "do_lower_case": false,
48
+ "extra_special_tokens": {},
49
+ "full_tokenizer_file": null,
50
+ "mask_token": "[MASK]",
51
+ "max_len": 512,
52
+ "max_length": 128,
53
+ "model_max_length": 128,
54
+ "never_split": null,
55
+ "pad_to_multiple_of": null,
56
+ "pad_token": "[PAD]",
57
+ "pad_token_type_id": 0,
58
+ "padding_side": "right",
59
+ "sep_token": "[SEP]",
60
+ "stride": 0,
61
+ "strip_accents": null,
62
+ "tokenize_chinese_chars": true,
63
+ "tokenizer_class": "DistilBertTokenizer",
64
+ "truncation_side": "right",
65
+ "truncation_strategy": "longest_first",
66
+ "unk_token": "[UNK]"
67
+ }
vocab.txt ADDED
The diff for this file is too large to render. See raw diff