wjunwei commited on
Commit
efd0a1a
1 Parent(s): 648d167

Add new SentenceTransformer model.

Browse files
1_Pooling/config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "word_embedding_dimension": 768,
3
+ "pooling_mode_cls_token": false,
4
+ "pooling_mode_mean_tokens": true,
5
+ "pooling_mode_max_tokens": false,
6
+ "pooling_mode_mean_sqrt_len_tokens": false,
7
+ "pooling_mode_weightedmean_tokens": false,
8
+ "pooling_mode_lasttoken": false,
9
+ "include_prompt": true
10
+ }
README.md ADDED
@@ -0,0 +1,533 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: intfloat/e5-base-v2
3
+ datasets: []
4
+ language: []
5
+ library_name: sentence-transformers
6
+ pipeline_tag: sentence-similarity
7
+ tags:
8
+ - sentence-transformers
9
+ - sentence-similarity
10
+ - feature-extraction
11
+ - generated_from_trainer
12
+ - dataset_size:5864
13
+ - loss:ContrastiveTensionLoss
14
+ widget:
15
+ - source_sentence: 'hp inch touchscreen laptop th generation intel core ig intel iris
16
+ xe graphics gb ram gb ssd windows home natural silver '
17
+ sentences:
18
+ - 'suposeu baby playpen play pen for kids activity center large baby playard for
19
+ indoor and outdoor sturdy safety baby fence with soft breathable mesh for toddler
20
+ grey suposeu baby playpen the x x inch playpen can accommodate a large number
21
+ of children toys and pets in addition to providing plenty of room for play it
22
+ provides a scientifically safe height for babies aged months to months to train
23
+ stand and walk parents helper our baby playpens are useful for allowing kids to
24
+ play learn how to move independently and explore the world of perception it protects
25
+ them from getting lost or exposed to danger the playpen can transform any space
26
+ into a play area so parents can use their unoccupied hands to take care of other
27
+ things full vision design the playpens side walls are composed of mesh that
28
+ is both visible and breathablethis seethrough mesh enables you to keep an eye
29
+ on your boy or girl and allows the baby to see you fostering comfort and safety
30
+ the external gate features a zipper design that allows for interaction with your
31
+ baby at any time thus instilling a sense of security in your baby safety play
32
+ area constructed with oxford fabric and robust steel pipe this square infant playpen
33
+ comes with sturdy suction cups at the bottom making it difficult to tip over
34
+ or move additionally our fence is entirely covered in soft cloth leaving no exposed
35
+ gaps which prevents any pinching dangers easy to assemble super sturdy the baby
36
+ fences pipes are made of rustproof alloy with abs joints for easy assembly and
37
+ disassembly the bottom is made of nonslip breathable and quickdrying fabric which
38
+ can be cleaned by simply wiping it down with a wet cloth and soap '
39
+ - ' in charging station for apple devices mag safe charger standw fast magnetic
40
+ charger wireless compatible for iphone promaxplusminiairpods proiwatch se mag
41
+ safe charger stand features three charging spots designed for your iphone apple
42
+ watch and airpods keeping your nightstand and desk at home or at work clutterfree
43
+ after a long day your devices find a cozy home here getting charged up and ready
44
+ for the morning compatible with iphone series iwatch ultra se and airpods pro with
45
+ magnetic precision and strength enjoy handsfree convenience in both landscape
46
+ and portrait mode this mag safe charger can power up your iphone pro max from to within
47
+ the span of to hours as confirmed by lab testing mag safe charger is certified
48
+ by rohs ce and fcc providing you the confidence to purchase worryfree the builtin
49
+ improved intelligent chipset rest you assured with overcurrent overvoltage and
50
+ overtemperature protection give you a safe and relible charging exeprience led
51
+ light ring is designed for checking charging status making sure no more dead phone
52
+ in the morning it also features a soft touch button to turn it off if you find
53
+ the light annoying at night equipped with an ergonomic design and a sturdy base
54
+ this magnetic phone charger creates a comfortable viewing angle for scrolling
55
+ and bingewatching and maintains stability while charging compatible with iphone series
56
+ iwatch ultra se and airpods pro treat your loved ones
57
+ to a gift that combines functionality with style this sleek accessory that enhances
58
+ any workspace or bedside table bringing a new level of convenience to your loved
59
+ ones daily routine its a perfect christmas gift and stocking stuffers for men
60
+ and women you enjoy months of warranty and feel free to reach out to us with
61
+ any inquiries regarding our product were here to assist you promptly and effectively
62
+ important note this charger is compatible solely with mag safe approved phone
63
+ cases alternatively you can also charge your device without a phone case magnetic
64
+ charger stand w qc adapter usb type c cablemft user manual'
65
+ - uscce alarm clock bluetooth fm radio w stereo sound speaker fast wireless charging
66
+ for iphone samsung dimmable clock radio for bedroom versatile multifunctional
67
+ device bluetooth speaker alarm clock with w fast wireless charger station builtin
68
+ fm radio and color night light exceptional audio quality enjoy superior sound
69
+ with the w stereo bluetooth speaker for a rich and immersive audio experience
70
+ whether youre streaming music listening to radio or sleeping with it powering
71
+ devices while you sleep convenient wireless charging and an extra usb charging
72
+ port ensure your devices are fully charged while you rest display dimmer get
73
+ your personal comfort brightness with full range display dimmer slider without
74
+ disturbing your sleep at night color night light with adjustable brightness choose
75
+ from seven vibrant colors and adjust the brightness to create the perfect ambiance
76
+ for any occasion stylish modern design a combination of fabric and wooden finish
77
+ offers a contemporary and fashionable appearance adding a touch of modern elegance
78
+ to your surroundings
79
+ - source_sentence: baby hair brush cradle cap brush baby hair comb baby hair brush
80
+ and comb set for newborns toddlers baby brush soft bristles ideal for cradle
81
+ cap perfect baby registry gift pink baby hair brush with the finest bristles
82
+ made from made of soft pp bristles bpa free you can softly brush your babys hair
83
+ and massage his scalp both for grooming and relaxation perfect brush for a healthy
84
+ cheerful baby baby hair comb the baby hair comb has spherical teeth that wont
85
+ pull or damage your babys hair or scalp provides safe easy grooming experience prevent
86
+ craddle cap by using the soft bristle cradle cap brush daily to massage your
87
+ babys scalp you can prevent cradle cap from building up while also protecting
88
+ their soft spot extra safe for a super happy baby baby shower registry gift this
89
+ baby brush will cover every parents needs for grooming baby hair choose it for
90
+ your registry or buy one for your friends with babies everyone loves this thoughtful
91
+ practical gift satisfaction guarantee you and your baby will love our product
92
+ or you will receive a complete refund of your money take advantage of our premium
93
+ risk free offer now and buy your baby the best grooming tools
94
+ sentences:
95
+ - pyle wireless bluetooth home audio amplifier systemupgraded channel watt sound
96
+ power stereo receiver wusb micro sd headphone microphone input wecho talkover
97
+ for pa ptabt powerful watt the pyle channel bluetooth home audio amplifier gives
98
+ you w peak power which can be used for multi speakers w ohms impedance allowing
99
+ you to enjoy high quality audio perfect for your pa and home theater entertainment
100
+ bluetooth capable the personal portable digital amp box works with todays latest
101
+ devices including smartphones tablets laptops and computers as it is equipped
102
+ with bluetooth wireless music streaming with hasslefree receiver pairing multiple
103
+ inputs the integrated indoor stereo receiver supports ipod or mp usb micro sd
104
+ aux playback builtin fm radio w digital lcd display audio input and rec subwoofer
105
+ output rca lr connectors also features microphone inputs mic talkover function
106
+ the professional sound amplifying device features a mic talkover function which
107
+ when activated the current music or audio being played will be temporarily halted
108
+ this function is used in paging voiceover and announcements integrated equalizer
109
+ the compact amplifier has crisp buttons for the audio sources and selectors rotary
110
+ knob for equalization mic and master volume adjustments a remote control is also
111
+ included in the package for distant audio adjustments
112
+ - hair thinning scissors ulg professional barbers texturizing teeth shears for hairdressing
113
+ salon and home use thinning shears for hair cutting made of japanese stainless
114
+ steel inch premium material made from japanese stainless steel hair thinning
115
+ scissor with precise thinning blades and handsharpened thinning edges to evenly
116
+ trim hair with ease high end design luxurious blue diamond stones in the twoleaf
117
+ intersection and mirror polish finish contains teeth on one blade and a smooth
118
+ razor edge on second multifunctional this thinning shear measures inch with
119
+ cutting blades that measure inch the professional barber hair thinning scissor
120
+ is safe and easy to use on men women the elderly offset grip thumb handle is
121
+ shorter than finger handle reduces over extension of thumb and wrist motion relaxes
122
+ your elbow and wrist this type of handle has the best ergonomics what you get teeth
123
+ scissors and removable rubber handle inserts to fit any thumb and finger with
124
+ comfort grip month warranty for qualityrelated issues
125
+ - handepo pcs bridal wedding gifts bridal bridesmaids robe sets with sleeping eye
126
+ mask slippers party tumbler tote gifts bridesmaid fluffy fabric silk made in the
127
+ china bridal party set package includes pair pieces of pink bridesmaid slippers
128
+ with bridesmaid and piece of bridesmaid robe pc eye mask cup tote bag an ideal
129
+ choice for bridal shower bridesmaid night party bridal party sleeping bathing
130
+ and lounging let you shine at any time reliable material these bridesmaid slippers
131
+ are made of fuzzy plush fluffy fabric thickness and comfortable can bring you
132
+ a nice wearing experience the bridal party robes are made of silk and are silky
133
+ soft and smooth providing you with reliable comfort after the bath or before going
134
+ to bed thoughtful design these bridesmaid slippers are embroidered to give you
135
+ a soft dry touch without freezing your feet even when outdoors the silk robe has
136
+ a belt which makes it fit well for a romantic wedding occasion meaningful ways
137
+ to apply our womens robes are suitable for wedding gifts bridal showers party
138
+ gifts its also a nice match for honeymoons trips and vacations its very memorable
139
+ and brings back precious memories for you one size the bridesmaid robes are suit
140
+ for most people and these bride slippers fit most us women tote bag measures about x inches x cm
141
+ the proper size is suitable for most ladies and women to wear
142
+ - source_sentence: senston badminton rackets set of graphite shaft badminton racquets
143
+ including badminton bag badminton shuttlecock racquet grip complete set get
144
+ everything you need to play badminton with family this badminton set includes highquality
145
+ badminton rackets nylon shuttlecocks random color grips and a convenient carrying
146
+ bag highquality material our badminton racket is made of a carbonfiber alloy frame
147
+ and graphite shaft making it both lightweight and extremely durable youll enjoy
148
+ the strength and stability it provides during your games solid construction with
149
+ a onepiece design outside and a builtin tjoint inside our racket offers improved
150
+ stability and highprecision control performance youll feel confident and in control
151
+ on the court perfect for beginners whether youre new to the game or just looking
152
+ for some casual fun these rackets are perfect for you for more advanced players
153
+ check out our store for professionalgrade rackets excellent customer service if
154
+ you have any issues with your order dont hesitate to contact us at any time
155
+ sentences:
156
+ - hot wheels car rolling storage case with retractable handle model conveniently
157
+ store and carry up to hot wheels cars makes collecting and clean up easy and
158
+ fun measures x x kids ages and up
159
+ - audiotechnica atlpxbtbk fully automatic bluetooth beltdrive stereo turntable black
160
+ hifi speed dust cover antiresonance diecast aluminum platter fully automatic
161
+ operation with speeds rpm and rpmwow and flutter wtd at khz antiresonance
162
+ diecast aluminum platter redesigned tonearm base and headshell for improved tracking
163
+ and reduced resonance integral dual moving magnet phono cartridge with replaceable
164
+ diamond stylus atnl highfidelity audio with bluetooth wireless technology compatible
165
+ bluetooth profile adp
166
+ - 'a pair of childrens bicycle riding flash silent training wheels suitable for inch
167
+ single speed bicycles huffy '
168
+ - source_sentence: 'squishmallows pajamas girls short pajamas set pink '
169
+ sentences:
170
+ - blush pink table runner hemstitch detail cotton ideal for spring memorial
171
+ day and summer solid hemstitch x inch perfect as ft table linen cotton cotton
172
+ hemstitch collection table runner made of cotton fabric handwoven in a weight
173
+ and texture perfectly suited for all seasons measures x in size our dining table
174
+ runner is inches wide and inches long when laid centered and lengthwise our runner
175
+ is long enough to fit any standard size table comfortably and yet have enough
176
+ cloth or material hanging over the edges highquality table linen to enhance your
177
+ table setting perfect for rustic vintage or country look coordinating placemats
178
+ pair with our matching placemats to create the foundation for a fun casual dining
179
+ experience this table runner and placemat set can be used for table dcor dressers
180
+ sofa tables and coffee tables it is perfect for everyday dinners parties holidays
181
+ picnics brunches and barbeques machine wash on gentle cycle and lay flat to dry
182
+ for stress free care sometimes the fabric doesnt lay flat after shipping use a
183
+ warm iron with steam and the wrinkles will come right out leaving you with a perfectly
184
+ flat and crisp table runner our offers an assortment of quality home and kitchen
185
+ products including tablecloths table runners cloth napkins placemats dish towels
186
+ visit our store to explore more
187
+ - 'kidscool space boys shortslittle big boys denim ripped summer shortalls '
188
+ - williams lectric shave electric razor preshave ounce stands up whiskers for a
189
+ closer shave lets your razor glide for less friction and skin irritation gives
190
+ you a closer more comfortable shave and smoother skin with soothing green tea
191
+ complex blade close electric smooth
192
+ - source_sentence: yeti yonder chug cap a bottle is only as good as its cap which
193
+ is why we brought the best parts of our rambler chug cap to the yonder cap leakproof leakproof
194
+ so you can carry it with confidence clippable slip it through a backpack strap
195
+ or clip it onto a carabiner to take water just about anywhere dishwasher safe because
196
+ no one needs more work to do spin the top off when you need a drink from the controlled
197
+ spout twist off the bottom when youre ready to refill or wash it
198
+ sentences:
199
+ - 'strapless prom dresses for women mermaid long sweetheart satin pleated with
200
+ high slit beaded evening gowns formal '
201
+ - 'lid for hydro flask oz wide mouth bottle replacement lid for thermoflaskiron
202
+ flasktakeya and more wide mouth bottles pack compatibilitysuitable for hydro
203
+ flaskshydroflaskthermoflaskiron flasktakeyaklean kanteensimple modern hydro cellkoodeebjpkpk
204
+ and more brands wide mouth water bottlesplease confirm the mouth inner diameter
205
+ and thread height of the water bottle before purchase inner diameter thread height important
206
+ note this lid does not fit tal hydro flask growler series hydropeak manna yeti
207
+ nalgene ozark trail water bottles or standard and narrow mouth water bottles when
208
+ you are not sure please feel free to contact us by email we will reply you in minutes
209
+ during working hours meanwhile we offer zerorisk purchase with a promise of full
210
+ refund or exchange soft handle the soft silicone handle and flexible rotation
211
+ design make it easy for you to carry a water bottle even when filled with water
212
+ simple and easy to replenish at any time safe and leak proof bpa free healthy
213
+ and safe eliminating leaks whether you are undergoing safety checks or traveling
214
+ keep your bag and clothes dry classic style simple and atmospheric appearance
215
+ design increases the charm of your water bottle the simpler the more classic it
216
+ is you will love your water bottle more because of this replacement lid '
217
+ - hydro flask standard mouth lids accessory for standard mouth water bottle standard
218
+ mouth flex straw cap fits all hydro flask standard mouth bottles straw is easy
219
+ to trim to fit your favorite hydro flask flex strap is easy to transport and comfortable
220
+ to carry honeycomb insulated cap for maximum temperature retention leakproof when
221
+ closed so you can reliably sip and transport your refreshment without worry bpafree toxinfree
222
+ removable components for easy cleaning dishwasher safe flex straw cap not intended
223
+ for use with hot liquids show more
224
+ ---
225
+
226
+ # SentenceTransformer based on intfloat/e5-base-v2
227
+
228
+ This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [intfloat/e5-base-v2](https://huggingface.co/intfloat/e5-base-v2). 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.
229
+
230
+ ## Model Details
231
+
232
+ ### Model Description
233
+ - **Model Type:** Sentence Transformer
234
+ - **Base model:** [intfloat/e5-base-v2](https://huggingface.co/intfloat/e5-base-v2) <!-- at revision 1c644c92ad3ba1efdad3f1451a637716616a20e8 -->
235
+ - **Maximum Sequence Length:** 512 tokens
236
+ - **Output Dimensionality:** 768 tokens
237
+ - **Similarity Function:** Cosine Similarity
238
+ <!-- - **Training Dataset:** Unknown -->
239
+ <!-- - **Language:** Unknown -->
240
+ <!-- - **License:** Unknown -->
241
+
242
+ ### Model Sources
243
+
244
+ - **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
245
+ - **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
246
+ - **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers)
247
+
248
+ ### Full Model Architecture
249
+
250
+ ```
251
+ SentenceTransformer(
252
+ (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel
253
+ (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
254
+ (2): Normalize()
255
+ )
256
+ ```
257
+
258
+ ## Usage
259
+
260
+ ### Direct Usage (Sentence Transformers)
261
+
262
+ First install the Sentence Transformers library:
263
+
264
+ ```bash
265
+ pip install -U sentence-transformers
266
+ ```
267
+
268
+ Then you can load this model and run inference.
269
+ ```python
270
+ from sentence_transformers import SentenceTransformer
271
+
272
+ # Download from the 🤗 Hub
273
+ model = SentenceTransformer("wjunwei/ecommerce_text_embedding_retrieval_v2")
274
+ # Run inference
275
+ sentences = [
276
+ 'yeti yonder chug cap a bottle is only as good as its cap which is why we brought the best parts of our rambler chug cap to the yonder cap leakproof leakproof so you can carry it with confidence clippable slip it through a backpack strap or clip it onto a carabiner to take water just about anywhere dishwasher safe because no one needs more work to do spin the top off when you need a drink from the controlled spout twist off the bottom when youre ready to refill or wash it',
277
+ 'lid for hydro flask oz wide mouth bottle replacement lid for thermoflaskiron flasktakeya and more wide mouth bottles pack compatibilitysuitable for hydro flaskshydroflaskthermoflaskiron flasktakeyaklean kanteensimple modern hydro cellkoodeebjpkpk and more brands wide mouth water bottlesplease confirm the mouth inner diameter and thread height of the water bottle before purchase inner diameter thread height important note this lid does not fit tal hydro flask growler series hydropeak manna yeti nalgene ozark trail water bottles or standard and narrow mouth water bottles when you are not sure please feel free to contact us by email we will reply you in minutes during working hours meanwhile we offer zerorisk purchase with a promise of full refund or exchange soft handle the soft silicone handle and flexible rotation design make it easy for you to carry a water bottle even when filled with water simple and easy to replenish at any time safe and leak proof bpa free healthy and safe eliminating leaks whether you are undergoing safety checks or traveling keep your bag and clothes dry classic style simple and atmospheric appearance design increases the charm of your water bottle the simpler the more classic it is you will love your water bottle more because of this replacement lid ',
278
+ 'hydro flask standard mouth lids accessory for standard mouth water bottle standard mouth flex straw cap fits all hydro flask standard mouth bottles straw is easy to trim to fit your favorite hydro flask flex strap is easy to transport and comfortable to carry honeycomb insulated cap for maximum temperature retention leakproof when closed so you can reliably sip and transport your refreshment without worry bpafree toxinfree removable components for easy cleaning dishwasher safe flex straw cap not intended for use with hot liquids show more',
279
+ ]
280
+ embeddings = model.encode(sentences)
281
+ print(embeddings.shape)
282
+ # [3, 768]
283
+
284
+ # Get the similarity scores for the embeddings
285
+ similarities = model.similarity(embeddings, embeddings)
286
+ print(similarities.shape)
287
+ # [3, 3]
288
+ ```
289
+
290
+ <!--
291
+ ### Direct Usage (Transformers)
292
+
293
+ <details><summary>Click to see the direct usage in Transformers</summary>
294
+
295
+ </details>
296
+ -->
297
+
298
+ <!--
299
+ ### Downstream Usage (Sentence Transformers)
300
+
301
+ You can finetune this model on your own dataset.
302
+
303
+ <details><summary>Click to expand</summary>
304
+
305
+ </details>
306
+ -->
307
+
308
+ <!--
309
+ ### Out-of-Scope Use
310
+
311
+ *List how the model may foreseeably be misused and address what users ought not to do with the model.*
312
+ -->
313
+
314
+ <!--
315
+ ## Bias, Risks and Limitations
316
+
317
+ *What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
318
+ -->
319
+
320
+ <!--
321
+ ### Recommendations
322
+
323
+ *What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
324
+ -->
325
+
326
+ ## Training Details
327
+
328
+ ### Training Dataset
329
+
330
+ #### Unnamed Dataset
331
+
332
+
333
+ * Size: 5,864 training samples
334
+ * Columns: <code>sentence_0</code>, <code>sentence_1</code>, and <code>label</code>
335
+ * Approximate statistics based on the first 1000 samples:
336
+ | | sentence_0 | sentence_1 | label |
337
+ |:--------|:------------------------------------------------------------------------------------|:------------------------------------------------------------------------------------|:------------------------------------------------|
338
+ | type | string | string | int |
339
+ | details | <ul><li>min: 5 tokens</li><li>mean: 145.16 tokens</li><li>max: 512 tokens</li></ul> | <ul><li>min: 3 tokens</li><li>mean: 147.19 tokens</li><li>max: 512 tokens</li></ul> | <ul><li>0: ~87.50%</li><li>1: ~12.50%</li></ul> |
340
+ * Samples:
341
+ | sentence_0 | sentence_1 | label |
342
+ |:-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------|
343
+ | <code>squishmallows original inch bluey hugmees mediumsized ultrasoft official jazwares plush squad up grow your squishmallows squad with bluey a supersoft collectible mediumsized hugmees plush musthave bring the fun home with this squishmallows made with ultrasoft highquality materials hugmees squishmallows hugmees have extended arms and are always ready for a hug collectible look for other squishmallows extensions including flipamallows fuzzamallows mystery squad and stackables only by original squishmallows officially licensed product this inch plush is officially licensed by the bbc</code> | <code>squishmallows original inch bluey hugmees mediumsized ultrasoft official jazwares plush squad up grow your squishmallows squad with bluey a supersoft collectible mediumsized hugmees plush musthave bring the fun home with this squishmallows made with ultrasoft highquality materials hugmees squishmallows hugmees have extended arms and are always ready for a hug collectible look for other squishmallows extensions including flipamallows fuzzamallows mystery squad and stackables only by original squishmallows officially licensed product this inch plush is officially licensed by the bbc</code> | <code>1</code> |
344
+ | <code>rechargeable headlamp high lumen bright led head lamp with red white light ipx waterproof headlight mode head flashlight for outdoor running hunting fishing hiking camping gear illuminate your world in all directions designed in the usa mioisy head lamp features powerful xpgled bulbs that provide up to lumens max ensuring that you can see everything around you clearly perfect for exploring caves night runningcyclingfishingcamping construction work and other outdoor adventure activities the red safety warning light switch is located on the back battery compartment to ensure all direction safety and emergency response usb rechargeable and long battery life do not use unsafe cylindrical batteries our rechargeable headlamp usa builtin rechargeable batteries to ensure your safety first our head lamps support typec usb charging making it convenient for everyday use the headlamp rechargeable can provide hours of longlasting power in different lighting modes so you can adventure without worrying about running out of juice long press and motion sensor in any mode press the on switch button for seconds the rechargeable headlamp flashlight will turn off directly no need to cycle through all the modes the headlights for head is also equipped with the smart motion sensor which easily controls the headlamps for adults on and off with a wave of your hand more convenient for your work ipx waterproof and modes for any situation our headlamp flashlight is built to withstand splashes of water from all angles so you can take it on any weather rain or shine the head light has modes controlled by buttons one button switch key modes the other button switch sensor modes our led headlamp is the ultimate adaptable tool for any situation ensuring you have the right light for any adventure adjustable angle and comfortable headband to ensure flexible lighting our head lights for forehead can be adjusted and the handsfree headlamp provides bright and steady lighting while you work the headlights for head use a soft and comfortable elastic headband that can be adjusted to fit different head size perfect headlamps for adults and kids only weight oz its comfortable to wear for long time ensuring you can explore with easethe band can be taken off to wash perfect gift for any occasion whether its fathers day thanksgiving christmas valentines day easter halloween or any special festival our rechargeable head flashlight is the perfect gift for anyone who loves the outdoor adventure give your father mother husband son or boyfriend the great gift with our powerful and reliable led headlamps if you have any questions please reach out to us to get professional solutions </code> | <code>ocyclone tablet stand ipad stand for desk adjustable height and angle foldable tablet holder stand compatible with portable monitor ipad pro air mini black wide compability ocyclone tablet holder stand works with all inches smartphones and most tablets with cases such as ipad pro ipad air ipad ipad mini samsung galaxy tabs surface surface pro kindle fire hd portable monitor drawing tablet height angle adjustable the height of the tablet stand holder can be simply adjusted the angle can be adjusted from to by hand with this ocyclone tablet holder you can enjoy your movies cooking reading studying playing games watching youtube without any worries providing you comfortable viewing angle which helps to fix your posture and reduce neck back ache hands free portable the foldable design of the ipad stand makes you easy to carry your phone and ipad everywhere you can put the stand in the bag or on the body undoubltly it is a great ideal accessories for you take it any place of course it is also a great ideal gift for your family or your friends they will definitely be satisfied with the portable tablet stand super sturdy fully protective silicone pad ocyclone desk tablet ipad stand with premium aluminum abs material makes it more durable than others quality nonskid rubber covered on the front and the bottom can mamximum protect your phone from slide and scratches you can easily tap the screen without worrying the devices will tip over or fall off friendly user design the reserved charging hole makes it more convenient to charge your devices while using this tablet phone holder in addtion the silicone hook pad will not cover the subtitle when you watching movies ocyclone always aims at providing our customers the best happy shopping experience if you have any confusing please get in touch with us we will answer you within hours</code> | <code>0</code> |
345
+ | <code>colgate extra soft toothbrush for sensitive teeth and gums with tongue and cheek cleaner pack extra soft toothbrush for sensitive teeth softer bristles protect tooth enamel and gums vs an ordinary soft manual toothbrush polishing cups gently remove teeth stains to whiten teeth our unique tongue and cheek cleaner remove bad breath bacteria raised cleaning tip helps get into hard to reach areas</code> | <code>water bottle stickers pcs cool neon stickers sticker pack for kids adults teens waterproof vinyl stickers stickers for laptop skateboard journal notesbook computer phone cup guitar luggage etc great variety sticker pack contains pieces mix neon stickers designed to be friendly healthy and nonrepetitive cool neon and fun patterns add a unique eyecatching flair to your bland items make your life more colorful funny gifts neon stickers have a unique visual effect injecting brilliant and cool colors and trendy vitality into life stickers for adults teens kids and stickers lovers stickers can be used as birthday gifts party favors home or classroom behavior rewards etc good quality beautiful vinyl stickers size in waterproof design bright colors and high resolution good sticking power no fading no unhealthy motifs not easy to tear safe and nontoxic even outdoors it can easily handle inclement weather widely used fun stickers can not only cover or embellish items so you can feel the fun and delightful emotions that come with decoration stickers for water bottle laptop journal scrapbook computer skateboard phone case macbook ipad planner cups suitcase luggage notebook scooter bike etc simple to use reusable stickers made with nonmarking adhesive can be randomly pasted or torn off without hurting the surface no residue is left behind when replacing or peel light up life all it takes is a cool and fun neon stickers brand stickers airnogo every product is carefully checked to ensure perfection if you have any questions we will take care of it immediately until you are satisfied</code> | <code>0</code> |
346
+ * Loss: [<code>ContrastiveTensionLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#contrastivetensionloss)
347
+
348
+ ### Training Hyperparameters
349
+ #### Non-Default Hyperparameters
350
+
351
+ - `num_train_epochs`: 5
352
+ - `multi_dataset_batch_sampler`: round_robin
353
+
354
+ #### All Hyperparameters
355
+ <details><summary>Click to expand</summary>
356
+
357
+ - `overwrite_output_dir`: False
358
+ - `do_predict`: False
359
+ - `eval_strategy`: no
360
+ - `prediction_loss_only`: True
361
+ - `per_device_train_batch_size`: 8
362
+ - `per_device_eval_batch_size`: 8
363
+ - `per_gpu_train_batch_size`: None
364
+ - `per_gpu_eval_batch_size`: None
365
+ - `gradient_accumulation_steps`: 1
366
+ - `eval_accumulation_steps`: None
367
+ - `learning_rate`: 5e-05
368
+ - `weight_decay`: 0.0
369
+ - `adam_beta1`: 0.9
370
+ - `adam_beta2`: 0.999
371
+ - `adam_epsilon`: 1e-08
372
+ - `max_grad_norm`: 1
373
+ - `num_train_epochs`: 5
374
+ - `max_steps`: -1
375
+ - `lr_scheduler_type`: linear
376
+ - `lr_scheduler_kwargs`: {}
377
+ - `warmup_ratio`: 0.0
378
+ - `warmup_steps`: 0
379
+ - `log_level`: passive
380
+ - `log_level_replica`: warning
381
+ - `log_on_each_node`: True
382
+ - `logging_nan_inf_filter`: True
383
+ - `save_safetensors`: True
384
+ - `save_on_each_node`: False
385
+ - `save_only_model`: False
386
+ - `restore_callback_states_from_checkpoint`: False
387
+ - `no_cuda`: False
388
+ - `use_cpu`: False
389
+ - `use_mps_device`: False
390
+ - `seed`: 42
391
+ - `data_seed`: None
392
+ - `jit_mode_eval`: False
393
+ - `use_ipex`: False
394
+ - `bf16`: False
395
+ - `fp16`: False
396
+ - `fp16_opt_level`: O1
397
+ - `half_precision_backend`: auto
398
+ - `bf16_full_eval`: False
399
+ - `fp16_full_eval`: False
400
+ - `tf32`: None
401
+ - `local_rank`: 0
402
+ - `ddp_backend`: None
403
+ - `tpu_num_cores`: None
404
+ - `tpu_metrics_debug`: False
405
+ - `debug`: []
406
+ - `dataloader_drop_last`: False
407
+ - `dataloader_num_workers`: 0
408
+ - `dataloader_prefetch_factor`: None
409
+ - `past_index`: -1
410
+ - `disable_tqdm`: False
411
+ - `remove_unused_columns`: True
412
+ - `label_names`: None
413
+ - `load_best_model_at_end`: False
414
+ - `ignore_data_skip`: False
415
+ - `fsdp`: []
416
+ - `fsdp_min_num_params`: 0
417
+ - `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
418
+ - `fsdp_transformer_layer_cls_to_wrap`: None
419
+ - `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
420
+ - `deepspeed`: None
421
+ - `label_smoothing_factor`: 0.0
422
+ - `optim`: adamw_torch
423
+ - `optim_args`: None
424
+ - `adafactor`: False
425
+ - `group_by_length`: False
426
+ - `length_column_name`: length
427
+ - `ddp_find_unused_parameters`: None
428
+ - `ddp_bucket_cap_mb`: None
429
+ - `ddp_broadcast_buffers`: False
430
+ - `dataloader_pin_memory`: True
431
+ - `dataloader_persistent_workers`: False
432
+ - `skip_memory_metrics`: True
433
+ - `use_legacy_prediction_loop`: False
434
+ - `push_to_hub`: False
435
+ - `resume_from_checkpoint`: None
436
+ - `hub_model_id`: None
437
+ - `hub_strategy`: every_save
438
+ - `hub_private_repo`: False
439
+ - `hub_always_push`: False
440
+ - `gradient_checkpointing`: False
441
+ - `gradient_checkpointing_kwargs`: None
442
+ - `include_inputs_for_metrics`: False
443
+ - `eval_do_concat_batches`: True
444
+ - `fp16_backend`: auto
445
+ - `push_to_hub_model_id`: None
446
+ - `push_to_hub_organization`: None
447
+ - `mp_parameters`:
448
+ - `auto_find_batch_size`: False
449
+ - `full_determinism`: False
450
+ - `torchdynamo`: None
451
+ - `ray_scope`: last
452
+ - `ddp_timeout`: 1800
453
+ - `torch_compile`: False
454
+ - `torch_compile_backend`: None
455
+ - `torch_compile_mode`: None
456
+ - `dispatch_batches`: None
457
+ - `split_batches`: None
458
+ - `include_tokens_per_second`: False
459
+ - `include_num_input_tokens_seen`: False
460
+ - `neftune_noise_alpha`: None
461
+ - `optim_target_modules`: None
462
+ - `batch_eval_metrics`: False
463
+ - `batch_sampler`: batch_sampler
464
+ - `multi_dataset_batch_sampler`: round_robin
465
+
466
+ </details>
467
+
468
+ ### Training Logs
469
+ | Epoch | Step | Training Loss |
470
+ |:------:|:----:|:-------------:|
471
+ | 0.6821 | 500 | 7.9052 |
472
+ | 1.3643 | 1000 | 4.3803 |
473
+ | 2.0464 | 1500 | 3.6253 |
474
+ | 2.7285 | 2000 | 3.6853 |
475
+ | 3.4106 | 2500 | 3.6878 |
476
+ | 4.0928 | 3000 | 3.602 |
477
+ | 4.7749 | 3500 | 3.6512 |
478
+
479
+
480
+ ### Framework Versions
481
+ - Python: 3.10.12
482
+ - Sentence Transformers: 3.0.1
483
+ - Transformers: 4.41.2
484
+ - PyTorch: 2.3.0+cu121
485
+ - Accelerate: 0.31.0
486
+ - Datasets: 2.20.0
487
+ - Tokenizers: 0.19.1
488
+
489
+ ## Citation
490
+
491
+ ### BibTeX
492
+
493
+ #### Sentence Transformers
494
+ ```bibtex
495
+ @inproceedings{reimers-2019-sentence-bert,
496
+ title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
497
+ author = "Reimers, Nils and Gurevych, Iryna",
498
+ booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
499
+ month = "11",
500
+ year = "2019",
501
+ publisher = "Association for Computational Linguistics",
502
+ url = "https://arxiv.org/abs/1908.10084",
503
+ }
504
+ ```
505
+
506
+ #### ContrastiveTensionLoss
507
+ ```bibtex
508
+ @inproceedings{carlsson2021semantic,
509
+ title={Semantic Re-tuning with Contrastive Tension},
510
+ author={Fredrik Carlsson and Amaru Cuba Gyllensten and Evangelia Gogoulou and Erik Ylip{"a}{"a} Hellqvist and Magnus Sahlgren},
511
+ booktitle={International Conference on Learning Representations},
512
+ year={2021},
513
+ url={https://openreview.net/forum?id=Ov_sMNau-PF}
514
+ }
515
+ ```
516
+
517
+ <!--
518
+ ## Glossary
519
+
520
+ *Clearly define terms in order to be accessible across audiences.*
521
+ -->
522
+
523
+ <!--
524
+ ## Model Card Authors
525
+
526
+ *Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
527
+ -->
528
+
529
+ <!--
530
+ ## Model Card Contact
531
+
532
+ *Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
533
+ -->
config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "intfloat/e5-base-v2",
3
+ "architectures": [
4
+ "BertModel"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "classifier_dropout": null,
8
+ "gradient_checkpointing": false,
9
+ "hidden_act": "gelu",
10
+ "hidden_dropout_prob": 0.1,
11
+ "hidden_size": 768,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 3072,
14
+ "layer_norm_eps": 1e-12,
15
+ "max_position_embeddings": 512,
16
+ "model_type": "bert",
17
+ "num_attention_heads": 12,
18
+ "num_hidden_layers": 12,
19
+ "pad_token_id": 0,
20
+ "position_embedding_type": "absolute",
21
+ "torch_dtype": "float32",
22
+ "transformers_version": "4.41.2",
23
+ "type_vocab_size": 2,
24
+ "use_cache": true,
25
+ "vocab_size": 30522
26
+ }
config_sentence_transformers.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "__version__": {
3
+ "sentence_transformers": "3.0.1",
4
+ "transformers": "4.41.2",
5
+ "pytorch": "2.3.0+cu121"
6
+ },
7
+ "prompts": {},
8
+ "default_prompt_name": null,
9
+ "similarity_fn_name": null
10
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:41c7ad57254d533fcab3b982d6ac720380d4ae3e6503931ae8a2f95ef9f3d8af
3
+ size 437951328
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_Pooling",
12
+ "type": "sentence_transformers.models.Pooling"
13
+ },
14
+ {
15
+ "idx": 2,
16
+ "name": "2",
17
+ "path": "2_Normalize",
18
+ "type": "sentence_transformers.models.Normalize"
19
+ }
20
+ ]
sentence_bert_config.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "max_seq_length": 512,
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 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": true,
45
+ "cls_token": "[CLS]",
46
+ "do_lower_case": true,
47
+ "mask_token": "[MASK]",
48
+ "model_max_length": 512,
49
+ "pad_token": "[PAD]",
50
+ "sep_token": "[SEP]",
51
+ "strip_accents": null,
52
+ "tokenize_chinese_chars": true,
53
+ "tokenizer_class": "BertTokenizer",
54
+ "unk_token": "[UNK]"
55
+ }
vocab.txt ADDED
The diff for this file is too large to render. See raw diff