Add SetFit model
Browse files- 1_Pooling/config.json +7 -0
- README.md +344 -0
- config.json +24 -0
- config_sentence_transformers.json +7 -0
- config_setfit.json +4 -0
- model.safetensors +3 -0
- model_head.pkl +3 -0
- modules.json +14 -0
- sentence_bert_config.json +4 -0
- special_tokens_map.json +51 -0
- tokenizer.json +0 -0
- tokenizer_config.json +59 -0
- vocab.txt +0 -0
1_Pooling/config.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
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 |
+
}
|
README.md
ADDED
@@ -0,0 +1,344 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
library_name: setfit
|
3 |
+
tags:
|
4 |
+
- setfit
|
5 |
+
- sentence-transformers
|
6 |
+
- text-classification
|
7 |
+
- generated_from_setfit_trainer
|
8 |
+
datasets:
|
9 |
+
- SetFit/bbc-news
|
10 |
+
metrics:
|
11 |
+
- accuracy
|
12 |
+
widget:
|
13 |
+
- text: 'versace art portfolio up for sale the art collection of murdered fashion
|
14 |
+
designer gianni versace could fetch up to £9m ($17m) when it is auctioned in new
|
15 |
+
york and london later this year. among the pictures for sale are works by roy
|
16 |
+
lichtenstein andy warhol and henri matisse. the collection was housed at versace
|
17 |
+
s six-storey new york townhouse. the 51-year-old designer was shot outside his
|
18 |
+
florida home in 1997 by suspected serial killer andrew cunanan who later killed
|
19 |
+
himself. the auction at sotheby s will feature 45 contemporary impressionist
|
20 |
+
and 19th century paintings. one of the highlights of the sale is roy lichtenstein
|
21 |
+
s blue nude which has been given an estimate of £1.8m ($3.4m). tobias meyer sotheby
|
22 |
+
s worldwide head of contemporary art said: this collection reflects mr versace
|
23 |
+
s wide-ranging taste and impeccable eye and many of the works were commissioned
|
24 |
+
directly from the artists. outstanding later examples from champions of the pop
|
25 |
+
movement such as roy lichtenstein are juxtaposed with masterpieces from the
|
26 |
+
most visible artists of the 1980 s including jean-michel basquiat and the collaborative
|
27 |
+
genius of basquiat and warhol as well as francesco clemente. much of the collection
|
28 |
+
will be offered for sale at three auctions in new york in june with smaller contemporary
|
29 |
+
paintings going under the hammer in london on 22 and 23 june. a sale of versace
|
30 |
+
s furniture and artworks sold in 2001fetched £5.5m ($10.3m).'
|
31 |
+
- text: 'councils prepare to set tax rises council tax in scotland is set to rise
|
32 |
+
by an average of about 4% in the coming year bbc scotland has learned. authorities
|
33 |
+
will decide final figures on thursday when projected increases will be more than
|
34 |
+
twice the rate of inflation which is currently 1.6%. the finance minister has
|
35 |
+
urged councils to limit increases but they have warned that they will struggle
|
36 |
+
to maintain services unless funding is increased. they say much additional government
|
37 |
+
money is for new initiatives. scottish finance minister tom mccabe msp said: last
|
38 |
+
week in parliament i announced an additional £419m for core expenditure to local
|
39 |
+
government in scotland. that s a 5.5% increase and sits against an inflation
|
40 |
+
rate of 1.6% so i think we have quite rightly said to councils this year that
|
41 |
+
we would at the very least ask them to exercise restraint. mr mccabe is also
|
42 |
+
looking for local authorities to become more efficient and save money in coming
|
43 |
+
years. he told bbc radio scotland s sunday live programme: here in scotland we
|
44 |
+
have 32 councils who all have their own individual collection systems for council
|
45 |
+
tax they have their own payroll systems and their own human resource systems. we
|
46 |
+
think there has to be opportunities there for rationalisation and using the money
|
47 |
+
saved to reinvest in frontline services. the councils umbrella organisation
|
48 |
+
cosla which provided bbc scotland with the indicative figures for next year warned
|
49 |
+
that councils would face a continuous struggle to maintain services. mr mccabe
|
50 |
+
has promised them about £8.1bn next year. however most of the increase is targeted
|
51 |
+
to new initiatives and councils will experience difficulties in maintaining core
|
52 |
+
services a cosla spokesman said. cosla says that it is willing to work with
|
53 |
+
the executive on finding efficiency savings but that these will not be enough
|
54 |
+
to maintain services. they say the funding plans for the next three years will
|
55 |
+
see councils lose more of the share of public spending. the conservatives accuse
|
56 |
+
the scottish executive of using the council tax to raise funds because it is too
|
57 |
+
afraid to raise income tax. the tory finance spokesman brian monteith msp said: its
|
58 |
+
a form of disguise... yet again we see that council tax is being used as a way
|
59 |
+
of passing on costs. scared of actually using its three pence income tax that
|
60 |
+
it could put up what we ve seen over the years is more and more burdens being
|
61 |
+
put onto local authorities and the council tax payer having to pick up the bill. there
|
62 |
+
are also warnings that unless funding to councils is increased in the next few
|
63 |
+
years then services may have to be reduced. linda knox director of the scottish
|
64 |
+
local authority management centre at strathclyde university said: with this
|
65 |
+
current settlement the increase is slowing. at the same time the burdens on councils
|
66 |
+
are greater than they were. the settlement figures don t include pay increases
|
67 |
+
and the executive is also requiring a substantial figure - in the area of £325m
|
68 |
+
- in efficiency savings across the settlement period. education will be protected
|
69 |
+
from any cuts but linda knox says this will mean other services will suffer. she
|
70 |
+
said: in practice that will mean a 4-5% cut for other services. on the face
|
71 |
+
of it the settlement looks like an increase of about 9.7% but by the time you
|
72 |
+
take into account other factors its probably only about 1% in real terms.'
|
73 |
+
- text: gadget show heralds mp3 christmas partners of those who love their hi-tech
|
74 |
+
gear may want to get their presents in early as experts predict a gadget shortage
|
75 |
+
this christmas. with apple s ipod topping wish lists again there may not be
|
76 |
+
enough ipod minis to go round predicts oliver irish editor of gadget magazine
|
77 |
+
stuff. the ipod mini is likely to be this year s tracey island said mr irish.
|
78 |
+
stuff has compiled a list of the top 10 gadgets for 2004 and the ipod is at number
|
79 |
+
one. for anyone bewildered by the choice of gadgets on the market stuff and
|
80 |
+
what hi-fi are hosting a best-of gadget show in london this weekend. star of
|
81 |
+
the show will be sony s qrio robot an all-singing all-dancing football-playing
|
82 |
+
man-machine who can even hold intelligent conversations. but he is not for sale
|
83 |
+
and sony has no commercial plans for the robot. he will greet visitors and is
|
84 |
+
flying in from japan. he probably has his own airplane seat that is how highly
|
85 |
+
sony prize him said mr irish. also on display will be a virtual keyboard which
|
86 |
+
projects itself onto any flat surface. the event will play host to a large collection
|
87 |
+
of digital music players from companies such as creative sony and philips as
|
88 |
+
well as the ubiquitously fashionable ipod from apple. suggestions that it could
|
89 |
+
be a gaming or wireless christmas are unlikely to come true as mp3 players remain
|
90 |
+
the most popular stocking filler said mr irish. demand is huge and apple has
|
91 |
+
promised that it can supply enough but people might struggle to get their hands
|
92 |
+
on ipod minis said mr irish. for those who like their gadgets to be multi-talented the
|
93 |
+
gizmondo a powerful gaming console with gps and gprs that also doubles up as
|
94 |
+
an mp3 player movie player and camera could be a must-have. what is impressive
|
95 |
+
is how much it can do and how well it can do them said mr irish. this christmas gadgets
|
96 |
+
will not be an all-male preserve. women will be getting gadgets from husbands
|
97 |
+
and boyfriends as well as buying them for themselves said mr irish. gadgets
|
98 |
+
nowadays are lifestyle products rather than just for geeks.
|
99 |
+
- text: 'virus poses as christmas e-mail security firms are warning about a windows
|
100 |
+
virus disguising itself as an electronic christmas card. the zafi.d virus translates
|
101 |
+
the christmas greeting on its subject line into the language of the person receiving
|
102 |
+
infected e-mail. anti-virus firms speculate that this multilingual ability is
|
103 |
+
helping the malicious program spread widely online. anti-virus firm sophos said
|
104 |
+
that 10% of the e-mail currently on the net was infected with the zafi virus. like
|
105 |
+
many other windows viruses zafi-d plunders microsoft outlook for e-mail addresses
|
106 |
+
and then uses mail-sending software to despatch itself across the web to new victims.
|
107 |
+
to be infected users must open up the attachment travelling with the message which
|
108 |
+
bears the code for the malicious bug. the attachment on the e-mail poses as an
|
109 |
+
electronic christmas card but anyone opening it will simply get a crude image
|
110 |
+
of two smiley faces. the virus subject line says merry christmas and translates
|
111 |
+
this into one of 15 languages depending of the final suffix of the e-mail address
|
112 |
+
the infected message has been sent to. the message in the body of the e-mail reads: happy
|
113 |
+
holidays and this too is translated. on infected machines the virus tries to
|
114 |
+
disable anti-virus and firewall software and opens up a backdoor on the pc to
|
115 |
+
hand over control to the writer of the virus. the virus is thought to have spread
|
116 |
+
most widely in south america italy spain bulgaria and hungary. the original
|
117 |
+
zafi virus appeared in april this year. we have seen these hoaxes for several
|
118 |
+
christmases already and personally i prefer traditional pen and paper cards and
|
119 |
+
we recommend this to all our clients too said mikko hypponen who heads f-secure
|
120 |
+
s anti-virus team.'
|
121 |
+
- text: desailly backs blues revenge trip marcel desailly insists there is no chance
|
122 |
+
of history repeating itself when chelsea take on barcelona on wednesday. the
|
123 |
+
french star was part of the chelsea side crushed 5-1 at the nou camp in the champions
|
124 |
+
league quarter-final second leg in 2000. things will be totally different this
|
125 |
+
time he told bbc sport. now everyone knows about chelsea and is a little bit
|
126 |
+
afraid of them. they are one of the major clubs in europe and the pressure will
|
127 |
+
be on barcelona. chelsea have not played barcelona since that quarter-final tie
|
128 |
+
five years ago. the blues had looked destined to progress after winning the first
|
129 |
+
leg at stamford bridge 3-1 courtesy of two goals from tore andre flo and one
|
130 |
+
by gianfranco zola. but they collapsed in the second leg going down to strikes
|
131 |
+
from rivaldo (2) luis figo dani and patrick kluivert. former chelsea captain
|
132 |
+
desailly who is now playing for al-gharafa in qatar says there is no comparison
|
133 |
+
between that side and the current blues team who are top of the premiership. mentally
|
134 |
+
they are much stronger even though a lot of their players are young the 36-year-old
|
135 |
+
said. we made some mistakes at the nou camp in 2000 - a lot of them were individual
|
136 |
+
mistakes. it would not happen now. this team has a new motivation and a different
|
137 |
+
mentality. world cup winner desailly saw huge changes during his time at stamford
|
138 |
+
bridge. he was signed for £4.6m from ac milan in 1998 by ruud gullit and went
|
139 |
+
on to play under gianluca vialli and claudio ranieri. but the biggest change occurred
|
140 |
+
when billionaire roman abramovich bought the club in 2003. desailly says the russian
|
141 |
+
s arrival helped to instil a winning mentality at the club as well as a demand
|
142 |
+
for success. the whole of chelsea is different now - the chairman the manager
|
143 |
+
and all the players he said. everything is new and there is a huge determination
|
144 |
+
to win. since that game in 2000 chelsea have gained more experience in europe
|
145 |
+
and were very close to reaching the champions league final last season. desailly
|
146 |
+
is one of the most decorated players in the history of football. he won the 1998
|
147 |
+
world cup and 2000 european championship with france the champions league in
|
148 |
+
1993 with marseilles and 1994 with ac milan two serie a titles and the fa cup
|
149 |
+
in 2000 with chelsea. he is now winding down his career in qatar alongside the
|
150 |
+
likes of frank lebeouf josep guardiola titi camara gabriel batistuta and christophe
|
151 |
+
dugarry. so he is full of admiration for two of his colleagues from the great
|
152 |
+
milan side of the mid-90s who are likely to line up against manchester united
|
153 |
+
on wednesday - paolo maldini and alessandro costacurta. i m happy that they have
|
154 |
+
managed to play so long at a high level he said. i made a vow to costacurta
|
155 |
+
that as long as he plays i will continue to play. and it s amazing that paolo
|
156 |
+
has managed to play at such a high level for such a long time.
|
157 |
+
pipeline_tag: text-classification
|
158 |
+
inference: true
|
159 |
+
base_model: sentence-transformers/paraphrase-mpnet-base-v2
|
160 |
+
model-index:
|
161 |
+
- name: SetFit with sentence-transformers/paraphrase-mpnet-base-v2
|
162 |
+
results:
|
163 |
+
- task:
|
164 |
+
type: text-classification
|
165 |
+
name: Text Classification
|
166 |
+
dataset:
|
167 |
+
name: SetFit/bbc-news
|
168 |
+
type: SetFit/bbc-news
|
169 |
+
split: test
|
170 |
+
metrics:
|
171 |
+
- type: accuracy
|
172 |
+
value: 0.879
|
173 |
+
name: Accuracy
|
174 |
+
---
|
175 |
+
|
176 |
+
# SetFit with sentence-transformers/paraphrase-mpnet-base-v2
|
177 |
+
|
178 |
+
This is a [SetFit](https://github.com/huggingface/setfit) model trained on the [SetFit/bbc-news](https://huggingface.co/datasets/SetFit/bbc-news) dataset that can be used for Text Classification. This SetFit model uses [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) as the Sentence Transformer embedding model. A [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance is used for classification.
|
179 |
+
|
180 |
+
The model has been trained using an efficient few-shot learning technique that involves:
|
181 |
+
|
182 |
+
1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning.
|
183 |
+
2. Training a classification head with features from the fine-tuned Sentence Transformer.
|
184 |
+
|
185 |
+
## Model Details
|
186 |
+
|
187 |
+
### Model Description
|
188 |
+
- **Model Type:** SetFit
|
189 |
+
- **Sentence Transformer body:** [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2)
|
190 |
+
- **Classification head:** a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance
|
191 |
+
- **Maximum Sequence Length:** 512 tokens
|
192 |
+
- **Number of Classes:** 5 classes
|
193 |
+
- **Training Dataset:** [SetFit/bbc-news](https://huggingface.co/datasets/SetFit/bbc-news)
|
194 |
+
<!-- - **Language:** Unknown -->
|
195 |
+
<!-- - **License:** Unknown -->
|
196 |
+
|
197 |
+
### Model Sources
|
198 |
+
|
199 |
+
- **Repository:** [SetFit on GitHub](https://github.com/huggingface/setfit)
|
200 |
+
- **Paper:** [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055)
|
201 |
+
- **Blogpost:** [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit)
|
202 |
+
|
203 |
+
### Model Labels
|
204 |
+
| Label | Examples |
|
205 |
+
|:------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
|
206 |
+
| 3 | <ul><li>'prince crowned top music earner prince earned more than any other pop star in 2004 beating artists such madonna and elton john in us magazine rolling stone s annual list. the singer banked $56.5m (£30.4m) from concerts album and publishing sales with his musicology tour and album. he kept madonna in second place as she earned $54.9m (£29.5m) while embarking on her global re-invention tour. veterans simon and garfunkel were in 10th place their comeback tour helping them earn $24.9m (£13.4m) last year. prince returned to centre stage after a decade in the commercial wilderness the magazine reported. the singer s 2004 tour took $90.3m (£48.5m) in ticket sales and he sold 1.9 million copies of his latest album musicology. although she grossed more than prince last year madonna remained in second place because of the monumental production costs of her tour. heavy metal band metallica s madly in anger with the world tour helped push their 2004 earnings up to $43.1m (£23.1m). they were ahead of sir elton john who took fourth place and almost $42.7m (£23m) from performances including a debut on the las vegas strip. other seasoned performers in the list included rod stewart whose sold-out shows and third volume of the great american songbook covers album helped net him £35m (£19m). the highest-ranking rap act in the list was 50 cent who at number 19 took $24m (£13m) to the bank.'</li><li>'bening makes awards breakthrough film actress annette bening is up for an oscar for her starring role in the award-winning film being julia. bening who was born in texas in 1958 has gained prominence for a string of key roles. although an oscar has so far eluded her her status as one of hollywood s favourite actresses remains solid. one of the biggest oscars buzzes in recent years was for bening s role as troubled carolyn burnham in 1999 s american beauty. but her deliciously neurotic portrayal of surburban life turning sour was overlooked in favour of hilary swank s leading role in boys don t cry. after opening her career in the theatre - and gaining a tony award nomination in 1986 - bening had a low-key spell in television. she then made her film debut as a sex-starved wife in 1988 s the great outdoors opposite comics dan akroyd and john candy. following a cameo in postcards from the edge bening s breakthrough role came in 1990 playing seductive con artist myra langtry in the grifters - a role that won her an oscar nomination. despite some rave reviews bening did not win the best supporting actress academy award. however the high-profile performance enabled bening to capture roles in a number of big budget hollywood productions co-starring with some of hollywood s greatest players including robert de niro and harrison ford. but it was her role opposite warren beatty in 1991 gangster flick bugsy which had the greatest impact. bening played virginia hill another role which won her high critical acclaim but the film only picked up two of a staggering 12 oscar nominations. bugsy was the start of a very significant liaison for bening - she married beatty with whom she had a child as the couple promoted the film in europe. the couple now have four children together. a pair of contrasting yet successful performances peppered bening s career in the mid-1990s. she played michael douglas opposite number in romantic comedy the american president and as a fiery queen elizabeth i in a contemporary adaptation of shakespeare s richard iii. bening was the first choice to play the female lead in american beauty director sam mendes film debut which was a critical and commercial success - but the actress had to be content with a bafta award for her strong performance. but current film being julia has made bening an award-winning actress rather than the perennial nominee with success at the golden globes. the witty tale of revenge and love set in 1930s london sees the eponymous theatrical diva - played by bening - grow tired of her success and fall for a young american - and could be the role which finally wins her the coveted academy award.'</li></ul> |
|
207 |
+
| 1 | <ul><li>'china continues rapid growth china s economy has expanded by a breakneck 9.5% during 2004 faster than predicted and well above 2003 s 9.1%. the news may mean more limits on investment and lending as beijing tries to take the economy off the boil. china has sucked in raw materials and energy to feed its expansion which could have knock-on effects on the rest of the world if it overheats. but officials pointed out that industrial growth had slowed with services providing much of the impetus. growth in industrial output - the main target of government efforts to impose curbs on credit and investments - was 11.5% in 2004 down from 17% the previous year. still consumer prices - at 2.4% - rose faster than in 2004 adding to concern that a sharp rise in producer prices of 7.1% could stoke inflation. and overall investment in fixed assets was still high up 21.3% from the previous year - although some way off the peak of 43% seen in the first quarter of 2004. the result could be higher interest rates. china raised rates by 0.27 percentage points to 5.8% - its first hike in nine years - in october 2004. despite the apparent rebalancing of the economy the overall growth picture remains strong economists said. there is no sign of a slowdown in 2005 said tim congdon economist at ing barings. china s economy is not only gathering speed thanks to domestic demand but also from soaring sales overseas. figures released earlier this year showed exports at a six-year high in 2004 up 35%. part of the impetus comes from the relative cheapness of the yuan china s currency. the government keeps it pegged close to a rate of 8.28 to the us dollar - much to the chagrin of many us lawmakers who blame china for lost jobs and competitiveness. despite urging to ease the peg officials insist they are a long way from ready to make a shift to a more market-set rate. we need a good and feasible plan and formulating such a plan also needs time national bureau of statistics chief li deshui told reuters. those who hope to make a fortune by speculating on a renminbi revaluation will not succeed in making a profit.'</li><li>'retail sales show festive fervour uk retail sales were better than expected in november as christmas shoppers began their seasonal flock to the high street figures show. the office for national statistics (ons) said retail sales rose 0.6% on the month and 6.1% on the year. but the figures along with this week s inflation report could trigger another interest rate rise in the new year. however recent data from the british retail consortium showed a 0.2% slip in high street sales during november. the ons data confounded analyst expectations. many had expected sales to fall slightly in november as shoppers put off buying christmas presents until december. however retailers attempts to draw in the crowds may be behind november s unexpected rise in sales they say. aggressive tactics such as one-day discount sales adopted by stores such as marks & spencer appear to have paid off. price discounting has certainly accounted for much of this because the value of retail sales hasn t grown as much as volumes said investec economist david page. the figures sparked a rally for sterling as the data supported the view that it is too early to assume that base rates have peaked.'</li><li>'uk firm faces venezuelan land row venezuelan authorities have said they will seize land owned by a british company as part of president chavez s agrarian reform programme. officials in cojedes state said on friday that farmland owned by a subsidiary of the vestey group would be taken and used to settle poor farmers. the government is cracking down on so-called latifundios or large rural estates which it says are lying idle. the vestey group said it had not been informed of any planned seizure. the firm whose agroflora subsidiary operates 13 farms in venezuela insisted that it had complied fully with venezuelan law. prosecutors in the south of the country have targeted hato el charcote a beef cattle ranch owned by agroflora. according to reuters they plan to seize 12 900 acres (5 200 hectares) from the 32 000 acre (13 000 hectare) farm. officials claim that agroflora does not possess valid documents proving its ownership of the land in question. they also allege that areas of the ranch are not being used for any form of active production. the legal boundaries did not match up with the actual boundaries and there is surplus state prosecutor alexis ortiz told reuters. as a consequence the government has taken action. controversial reforms passed in 2001 give the government the right to take control of private property if it is declared idle or ownership cannot be traced back to the 19th century. critics say the powers - which president chavez argues are needed to help the country s poorest citizens and develop the venezuelan economy - trample all over private property rights. the vestey group said it had owned the land since 1920 and would co-operate fully with the authorities. but a spokesman added: agroflora is absolutely confident that what it has submitted will demonstrate the legality of its title to the land. the company pointed out that the farm which employs 300 workers provides meat solely for the venezuelan market. last month the government said it had identified more than 500 idle farms and had yet to consider the status of a further 40 000. the authorities said landowners whose titles were in order and whose farms were productive had nothing to fear . under president chavez the venezuelan government has steadily expanded the state s involvement in the country s economy. it recently said all mining contracts involving foreign firms would be examined to ensure they provided sufficient economic benefits to the state.'</li></ul> |
|
208 |
+
| 4 | <ul><li>'tories would cut number of mps the conservative party would cut the number of mps by about one-fifth if they were elected tory leader michael howard has said. the plan forms part of the party s smaller government bill to be unveiled later this week. mr howard told the sunday times the party would also reduce the number of government special advisers. and he said a referendum would be held in wales to decide whether or not to scrap the welsh assembly. the changes would all take place within five years of the conservatives winning a general election mr howard told the paper. the precise number of mps to go would depend on the result of the welsh referendum but it would probably mean a reduction of around 120 from the current total of 659. if wales decided to keep its assembly it would stand to lose more mps. mr howard said as both parties planned to cut the number of civil servants at whitehall - labour by more than 80 000 and the tories by almost 100 000 - they should accept a similar drop in their own numbers. it is all very well saying government departments should be reduced but what about ministers parliament and special advisers he said. shadow leader of the commons oliver heald said: this will be part of our aim to reduce unnecessary and costly interference in people s lives by reducing the size and role of the state . mr howard said the plan would also even out the great unfairness of there being proportionately more scottish and welsh mps at westminster than those from england. the average size of an english constituency was 70 000 people mr howard said. in northern ireland it was just over 66 000 in wales just over 59 000 and in scotland 53 000. the number of welsh mps would be set at an amount that was consistent and fair in terms of representation with the rest of the united kingdom if the assembly was scrapped. mr howard said the changes should be carried out quickly and could even be implemented by the election after next. you have got to have a big bang. we don t want this like the house of lords reform getting to one stage and then not having the next stage.'</li><li>'short attacks us over tsunami aid former cabinet minister clare short has criticised the us-led tsunami aid coalition saying the un should be leading efforts. president bush has announced that an alliance of the us india australia and japan will co-ordinate a humanitarian drive. but ms short said the effect of the parallel coalition would be to undermine the un. she said only the un had the moral authority to lead the relief work. ms short resigned as international development secretary over the iraq war. i think this initiative from america to set up four countries claiming to co-ordinate sounds like yet another attempt to undermine the un when it is the best system we have got and the one that needs building up she said. only really the un can do that job she told bbc radio four s pm programme. it is the only body that has the moral authority. but it can only do it well if it is backed up by the authority of the great powers. ms short said the countries involved could not boast good records on their response to major disasters. the us was very bad at coordinating with anyone and india had its own problems ms short said. i don t know what that is about but it sounds very much i am afraid like the us trying to have a separate operation and not work with the rest of the world through the un system she added.'</li><li>'ministers deny care sums wrong ministers have insisted they are committed to free personal care for the elderly despite research suggesting the cost of the policy was under-estimated. a report by the fraser of allander institute says the decision to push ahead with the flagship policy was based on flawed research. deputy health minister rhona brankin has pledged to study the research. snp holyrood leader nicola sturgeon said the public needed reassurance that the care programme was secure. the rise in costs stems from a series of mistakes in the research used by the care development group of scottish executive experts who prepared the original costings according to findings published in the quarterly economic commentary of strathclyde university s fraser of allander institute. dubious assumptions about improving health expectancy could drive the cost of the policy up by another £130m by 2022 the report warned. it was carried out by husband and wife economist team jim and margaret cuthbert. but ms brankin told bbc radio scotland: we don t think we got our sums wrong. obviously we will examine the findings of this new report along with figures from our own research that we have already commissioned. we will look in great detail at any contribution to this because we need to be sure we can provide free personal care and nursing care for our older people into the future. and we are absolutely committed to doing that. but the scottish national party called on ministers to reassure people that enough funding is in place to support the free personal care policy. ms sturgeon said that while she had no reason to doubt the executive s support for the policy there were questions which needed to be answered and if necessary sums redone. she said: serious concerns have been raised and there are questions which need to be answered by the scottish executive. we need to know that the money is there not just for this year or next year but into the future so that older people and those who are looking forward to older age can rest assured that their personal care needs will be met.'</li></ul> |
|
209 |
+
| 2 | <ul><li>'vickery out of six nations england tight-head prop phil vickery has been ruled out of the rest of the 2005 rbs six nations after breaking a bone in his right forearm. vickery was injured as his club side gloucester beat bath 17-16 in the west country derby on saturday. he could be joined on the sidelines by bath centre olly barkley who sat out the derby due to a leg injury. barkley will have a scan on sunday and might miss england s trip to six nations leaders ireland next weekend. the news is just the latest blow for coach andy robinson who has seen his side lose their opening two matches in the 2005 six nations. robinson is already without world cup winners jonny wilkinson will greenwood mike tindall richard hill and trevor woodman through injury. vickery has broken the radius a large bone in his forearm. he only returned to the england side last weekend after a long-term back injury which was followed by a fractured eye socket. and the gloucester prop was only recalled after leicester tight-head julian white suffered a neck injury which has already seen him ruled out of the ireland game. bath prop matt stevens is the only remaining tight-head in england s training squad and could be involved against ireland. but he has to play second fiddle at club level to duncan bell who excelled for england a against france and may now be called into the squad. the extent of barkley s injury is not yet clear but bath boss john connolly rates him no better than 50-50 to face ireland. barkley played at inside cente in england s defeat by france and if he is unable to play england s constantly-changing midfield will once again have to be altered. robinson could choose to recall mathew tait or henry paul although tait endured a nightmare for newcastle against leicester on saturday and paul limped off with an ankle injury against bath. in-form leicester centre ollie smith is the other outstanding candidate and two tries against newcastle will have boosted his chances. fly-half andy goode is also a strong contender for the match-day 22 after an immaculate kicking display on saturday. england fourth in the six nations table with zero points play ireland top of the table in dublin on 27 february kick-off 1500 gmt.'</li><li>'robinson answers critics england captain jason robinson has rubbished suggestions that the world champions are a team in decline. england were beaten 11-9 by wales in their six nations opener in cardiff last week and face current champions france at twickenham on sunday. robinson said: we are certainly not on the decline. you lose one game and it doesn t make you a bad team. i have no doubt in the players we ve got. we have still got the team to go out and beat anyone on our day. england find themselves striving to avoid a third successive championship defeat for the first time since 1987. but full-back robinson believes the new-look england team can stop the rot against france. last weekend we should have won the game he said. but if we can under-perform and lose by only two points then i am sure if we play well this week we will get the win we need. we proved that in the autumn - when we put in some excellent performances - and we just need to build on that. it was a disappointing start against wales and we might be down on that. but we are certainly not out. we will come out fighting this week. robinson also had words of comfort for 18-year-old newcastle centre mathew tait who made his international debut against wales but has been demoted from the squad to face france. i have had a word with mathew said robinson. i still believe in him. he is an outstanding player but we have gone for olly (barkley) because of the kicking. mathew has just got to take it on the chin keep working hard like he is doing and i m sure he will feature in some of the games.'</li><li>'tevez - an argentine in brazil some 65 years ago bernardo gandulla left boca juniors in his native argentina and headed for brazil. he arrived at the vasco da gama club with considerable prestige - which he failed to live up to. he left an impression but not for reasons he would have wanted. the only thing of note he did was to throw the ball back when it went out of play. even today gandula is the term used in brazil for ballboy. brazilians delighted in the failure of gandulla then and many will delight now if history is repeated. another boca juniors player is coming to brazil. the new maradona - carlos tevez - is joining corinthians for nearly $20m in a strong candidate for the strangest move of the year. corinthians - or the london-based company who have just entered into partnership with them - appear to have paid well above the odds for tevez. it is particularly strange as - until last week - tevez was not even an especially big name in brazil where the argentine championship is not shown. presumably there is some financial logic behind the move. what is harder to work out is why tevez should be interested in heading north rather than across the atlantic. argentina s brightest new star hit top form in august as he spearheaded the charge towards the olympic gold medal. but the subsequent months have not been easy for tevez. boca re-signed martin palermo forcing tevez to vacate his favoured centre forward role. and off the pitch he gave the press plenty to report - atletico madrid gave up on him after declaring his private life to be disorganized. hounded by the press tevez grew tired of his life in buenos aires. so now he says that he is heading for corinthians in search of tranquillity. anyone acquainted with the city of são paulo will relish the nonsense of this declaration. brazil s biggest city is a sprawling ugly restless metropolis. once it shared with buenos aires the same melancholy air of a city of italian immigrants. but that world now lies buried under skyscrapers. the most turbulent place in this frenetic city is sport club corinthians. it is the most passionately supported club in brazil. pelé scored reams of goals against them reasoning that unless he did something special the corinthians fans were capable of winning the game on their own. that passion can also be turned against the team - especially nowadays when the level of urban violence gives a hysterical edge to life in brazil s big cities. corinthians fans have invaded the dressing room or ambushed the team coach to attack the players. tevez then can kiss tranquillity goodbye. he can also say farewell to the copa libertadores south america s champions league. corinthians have not qualified. indeed he will have a long wait for a worthwhile match. next year s brazilian championship only gets underway at the end of april. until then he will have to make do with the foolish são paulo state championship an appalling waste of his talent. there will be matches in small towns on poor pitches with plenty of defenders keen to kick him all the way back to argentina. like so many of the top south american players tevez has the gift of being able to surprise - to feint right and go left to pass when he was expected to dribble to shoot and catch the goalkeeper unawares. but his choice of destination is the biggest surprise of all.'</li></ul> |
|
210 |
+
| 0 | <ul><li>'video phone help for deaf people deaf people who prefer to communicate using british sign language (bsl) could soon be having their phone conversations relayed using webcams or videophones and an interpreter. the video relay service is being piloted by the royal national institute for deaf people (rnid) but the organisation says unless the service is provided at the same rate as voice calls it will be beyond most people s pockets. the rnid is urging telecoms regulator ofcom to reduce the cost of the service from the current £7.00 per minute and make it the same as ordinary phone calls. the service works by putting a deaf person in visual contact with a bsl interpreter via a webcam or video phone and the interpreter then relays the deaf person s conversation using a telephone and translates the other person s response into sign language. for many deaf people especially those born deaf bsl is a first and preferred means of communication. until now the only alternative has been to use textphones which means having to type a message and have it relayed via an operator. in the past i ve used textphones but they have problems said robert currington who is taking part in the pilot. i communicate in bsl; my written english is not very good and it takes me longer to think in english and type my message. i sometimes find it difficult to understand the reply. the rnid says the uk is lagging behind other countries which are already making relay services available at the cost of an ordinary phone call. there are no technical or economic reasons for not providing equivalent access to services for deaf people said rnid technology director guido gybels. in the us and australia sign language relay services have already been made universally available at the same cost as a voice call. by failing to provide and fund the video relay service for sign language users the telecommunications sector is effectively discriminating against an already disenfranchised group. ofcom says it has plans to review the services that telecoms companies are obliged to provide early next year. and new technology including the video relay service will be discussed with interested parties in the near future. but a spokesman said its powers were limited by legislation. any proposals to extend existing arrangements to cover new services would be for government to consider he said. mr currington like many of the uk s 70 000 bsl users will be hoping that a way can be found to make a cost-effective service available. the relay service makes phone conversations a pleasure he said. i can show my emotions more easily in bsl in the same way hearing people express emotions through voice calls.'</li><li>'concern over rfid tags consumers are very concerned about the use of radio frequency id (rfid) tags in shops a survey says. more than half of 2 000 people surveyed said they had privacy worries about the tags which can be used to monitor stock on shelves or in warehouses. some consumer groups have expressed concern that the tags could be used to monitor shoppers once they had left shops with their purchases. the survey showed that awareness of tags among consumers in europe was low. the survey of consumers in the uk france germany and the netherlands was carried out by consultancy group capgemini. the firm works on behalf of more than 30 firms who are seeking to promote the growth of rfid technology. the tags are a combination of computer chip and antenna which can be read by a scanner - each item contains a unique identification number. more than half (55%) of the respondents said they were either concerned or very concerned that rfid tags would allow businesses to track consumers via product purchases. fifty nine percent of people said they were worried that rfid tags would allow data to be used more freely by third parties. ard jan vetham capgemini s principal consultant on rfid said the survey showed that retailers needed to inform and educate people about rfid before it would become accepted technology. acceptance of new technologies always has a tipping point at which consumers believe that benefits outweigh concerns. with the right rfid approach and ongoing communication with consumers the industry can reach this point. he said that the survey also showed people would accept rfid if they felt that the technology could mean a reduction in car theft or faster recovery of stolen items. the tags are currently being used at one tesco distribution centre in the uk - the tags allow the rapid inventory of bulk items. they are also in use as a passcard for the m6 toll in the midlands in the uk. mr vetham said the majority of people surveyed (52%) believed that rfid tags could be read from a distance. he said that was a misconception based on a lack of awareness of the technology. at least once consumer group - consumers against supermarket privacy invasion and numbering (caspian) - has claimed that rfid chips could be used to secretly identify people and the things they are carrying or wearing. all kinds of personal belongings including clothes could constantly broadcast messages about their whereabouts and their owners it warned.'</li></ul> |
|
211 |
+
|
212 |
+
## Evaluation
|
213 |
+
|
214 |
+
### Metrics
|
215 |
+
| Label | Accuracy |
|
216 |
+
|:--------|:---------|
|
217 |
+
| **all** | 0.879 |
|
218 |
+
|
219 |
+
## Uses
|
220 |
+
|
221 |
+
### Direct Use for Inference
|
222 |
+
|
223 |
+
First install the SetFit library:
|
224 |
+
|
225 |
+
```bash
|
226 |
+
pip install setfit
|
227 |
+
```
|
228 |
+
|
229 |
+
Then you can load this model and run inference.
|
230 |
+
|
231 |
+
```python
|
232 |
+
from setfit import SetFitModel
|
233 |
+
|
234 |
+
# Download from the 🤗 Hub
|
235 |
+
model = SetFitModel.from_pretrained("vidhi0206/setfit-paraphrase-mpnet-base-v2")
|
236 |
+
# Run inference
|
237 |
+
preds = model("versace art portfolio up for sale the art collection of murdered fashion designer gianni versace could fetch up to £9m ($17m) when it is auctioned in new york and london later this year. among the pictures for sale are works by roy lichtenstein andy warhol and henri matisse. the collection was housed at versace s six-storey new york townhouse. the 51-year-old designer was shot outside his florida home in 1997 by suspected serial killer andrew cunanan who later killed himself. the auction at sotheby s will feature 45 contemporary impressionist and 19th century paintings. one of the highlights of the sale is roy lichtenstein s blue nude which has been given an estimate of £1.8m ($3.4m). tobias meyer sotheby s worldwide head of contemporary art said: this collection reflects mr versace s wide-ranging taste and impeccable eye and many of the works were commissioned directly from the artists. outstanding later examples from champions of the pop movement such as roy lichtenstein are juxtaposed with masterpieces from the most visible artists of the 1980 s including jean-michel basquiat and the collaborative genius of basquiat and warhol as well as francesco clemente. much of the collection will be offered for sale at three auctions in new york in june with smaller contemporary paintings going under the hammer in london on 22 and 23 june. a sale of versace s furniture and artworks sold in 2001fetched £5.5m ($10.3m).")
|
238 |
+
```
|
239 |
+
|
240 |
+
<!--
|
241 |
+
### Downstream Use
|
242 |
+
|
243 |
+
*List how someone could finetune this model on their own dataset.*
|
244 |
+
-->
|
245 |
+
|
246 |
+
<!--
|
247 |
+
### Out-of-Scope Use
|
248 |
+
|
249 |
+
*List how the model may foreseeably be misused and address what users ought not to do with the model.*
|
250 |
+
-->
|
251 |
+
|
252 |
+
<!--
|
253 |
+
## Bias, Risks and Limitations
|
254 |
+
|
255 |
+
*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
|
256 |
+
-->
|
257 |
+
|
258 |
+
<!--
|
259 |
+
### Recommendations
|
260 |
+
|
261 |
+
*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
|
262 |
+
-->
|
263 |
+
|
264 |
+
## Training Details
|
265 |
+
|
266 |
+
### Training Set Metrics
|
267 |
+
| Training set | Min | Median | Max |
|
268 |
+
|:-------------|:----|:-------|:----|
|
269 |
+
| Word count | 213 | 390.5 | 625 |
|
270 |
+
|
271 |
+
| Label | Training Sample Count |
|
272 |
+
|:------|:----------------------|
|
273 |
+
| 0 | 2 |
|
274 |
+
| 1 | 6 |
|
275 |
+
| 2 | 3 |
|
276 |
+
| 3 | 2 |
|
277 |
+
| 4 | 3 |
|
278 |
+
|
279 |
+
### Training Hyperparameters
|
280 |
+
- batch_size: (8, 8)
|
281 |
+
- num_epochs: (1, 1)
|
282 |
+
- max_steps: -1
|
283 |
+
- sampling_strategy: oversampling
|
284 |
+
- num_iterations: 20
|
285 |
+
- body_learning_rate: (2e-05, 2e-05)
|
286 |
+
- head_learning_rate: 2e-05
|
287 |
+
- loss: CosineSimilarityLoss
|
288 |
+
- distance_metric: cosine_distance
|
289 |
+
- margin: 0.25
|
290 |
+
- end_to_end: False
|
291 |
+
- use_amp: False
|
292 |
+
- warmup_proportion: 0.1
|
293 |
+
- seed: 42
|
294 |
+
- eval_max_steps: -1
|
295 |
+
- load_best_model_at_end: False
|
296 |
+
|
297 |
+
### Training Results
|
298 |
+
| Epoch | Step | Training Loss | Validation Loss |
|
299 |
+
|:------:|:----:|:-------------:|:---------------:|
|
300 |
+
| 0.0125 | 1 | 0.1976 | - |
|
301 |
+
| 0.625 | 50 | 0.0015 | - |
|
302 |
+
|
303 |
+
### Framework Versions
|
304 |
+
- Python: 3.10.12
|
305 |
+
- SetFit: 1.0.2
|
306 |
+
- Sentence Transformers: 2.2.2
|
307 |
+
- Transformers: 4.35.2
|
308 |
+
- PyTorch: 2.1.0+cu121
|
309 |
+
- Datasets: 2.16.1
|
310 |
+
- Tokenizers: 0.15.0
|
311 |
+
|
312 |
+
## Citation
|
313 |
+
|
314 |
+
### BibTeX
|
315 |
+
```bibtex
|
316 |
+
@article{https://doi.org/10.48550/arxiv.2209.11055,
|
317 |
+
doi = {10.48550/ARXIV.2209.11055},
|
318 |
+
url = {https://arxiv.org/abs/2209.11055},
|
319 |
+
author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren},
|
320 |
+
keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences},
|
321 |
+
title = {Efficient Few-Shot Learning Without Prompts},
|
322 |
+
publisher = {arXiv},
|
323 |
+
year = {2022},
|
324 |
+
copyright = {Creative Commons Attribution 4.0 International}
|
325 |
+
}
|
326 |
+
```
|
327 |
+
|
328 |
+
<!--
|
329 |
+
## Glossary
|
330 |
+
|
331 |
+
*Clearly define terms in order to be accessible across audiences.*
|
332 |
+
-->
|
333 |
+
|
334 |
+
<!--
|
335 |
+
## Model Card Authors
|
336 |
+
|
337 |
+
*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
|
338 |
+
-->
|
339 |
+
|
340 |
+
<!--
|
341 |
+
## Model Card Contact
|
342 |
+
|
343 |
+
*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
|
344 |
+
-->
|
config.json
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "/root/.cache/torch/sentence_transformers/sentence-transformers_paraphrase-mpnet-base-v2/",
|
3 |
+
"architectures": [
|
4 |
+
"MPNetModel"
|
5 |
+
],
|
6 |
+
"attention_probs_dropout_prob": 0.1,
|
7 |
+
"bos_token_id": 0,
|
8 |
+
"eos_token_id": 2,
|
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-05,
|
15 |
+
"max_position_embeddings": 514,
|
16 |
+
"model_type": "mpnet",
|
17 |
+
"num_attention_heads": 12,
|
18 |
+
"num_hidden_layers": 12,
|
19 |
+
"pad_token_id": 1,
|
20 |
+
"relative_attention_num_buckets": 32,
|
21 |
+
"torch_dtype": "float32",
|
22 |
+
"transformers_version": "4.35.2",
|
23 |
+
"vocab_size": 30527
|
24 |
+
}
|
config_sentence_transformers.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"__version__": {
|
3 |
+
"sentence_transformers": "2.0.0",
|
4 |
+
"transformers": "4.7.0",
|
5 |
+
"pytorch": "1.9.0+cu102"
|
6 |
+
}
|
7 |
+
}
|
config_setfit.json
ADDED
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"labels": null,
|
3 |
+
"normalize_embeddings": false
|
4 |
+
}
|
model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4de473d5c032d2124f68cab6cb63b10f972fd20be0686054f01b14f1db6e80b3
|
3 |
+
size 437967672
|
model_head.pkl
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fbde7e4d9fbf05e07e9ecf69609fefe25e7940647ef584a3fccfee7408f4e42b
|
3 |
+
size 31631
|
modules.json
ADDED
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
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 |
+
]
|
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 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": {
|
3 |
+
"content": "<s>",
|
4 |
+
"lstrip": false,
|
5 |
+
"normalized": false,
|
6 |
+
"rstrip": false,
|
7 |
+
"single_word": false
|
8 |
+
},
|
9 |
+
"cls_token": {
|
10 |
+
"content": "<s>",
|
11 |
+
"lstrip": false,
|
12 |
+
"normalized": true,
|
13 |
+
"rstrip": false,
|
14 |
+
"single_word": false
|
15 |
+
},
|
16 |
+
"eos_token": {
|
17 |
+
"content": "</s>",
|
18 |
+
"lstrip": false,
|
19 |
+
"normalized": false,
|
20 |
+
"rstrip": false,
|
21 |
+
"single_word": false
|
22 |
+
},
|
23 |
+
"mask_token": {
|
24 |
+
"content": "<mask>",
|
25 |
+
"lstrip": true,
|
26 |
+
"normalized": false,
|
27 |
+
"rstrip": false,
|
28 |
+
"single_word": false
|
29 |
+
},
|
30 |
+
"pad_token": {
|
31 |
+
"content": "<pad>",
|
32 |
+
"lstrip": false,
|
33 |
+
"normalized": false,
|
34 |
+
"rstrip": false,
|
35 |
+
"single_word": false
|
36 |
+
},
|
37 |
+
"sep_token": {
|
38 |
+
"content": "</s>",
|
39 |
+
"lstrip": false,
|
40 |
+
"normalized": true,
|
41 |
+
"rstrip": false,
|
42 |
+
"single_word": false
|
43 |
+
},
|
44 |
+
"unk_token": {
|
45 |
+
"content": "[UNK]",
|
46 |
+
"lstrip": false,
|
47 |
+
"normalized": false,
|
48 |
+
"rstrip": false,
|
49 |
+
"single_word": false
|
50 |
+
}
|
51 |
+
}
|
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
ADDED
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"added_tokens_decoder": {
|
3 |
+
"0": {
|
4 |
+
"content": "<s>",
|
5 |
+
"lstrip": false,
|
6 |
+
"normalized": false,
|
7 |
+
"rstrip": false,
|
8 |
+
"single_word": false,
|
9 |
+
"special": true
|
10 |
+
},
|
11 |
+
"1": {
|
12 |
+
"content": "<pad>",
|
13 |
+
"lstrip": false,
|
14 |
+
"normalized": false,
|
15 |
+
"rstrip": false,
|
16 |
+
"single_word": false,
|
17 |
+
"special": true
|
18 |
+
},
|
19 |
+
"2": {
|
20 |
+
"content": "</s>",
|
21 |
+
"lstrip": false,
|
22 |
+
"normalized": false,
|
23 |
+
"rstrip": false,
|
24 |
+
"single_word": false,
|
25 |
+
"special": true
|
26 |
+
},
|
27 |
+
"104": {
|
28 |
+
"content": "[UNK]",
|
29 |
+
"lstrip": false,
|
30 |
+
"normalized": false,
|
31 |
+
"rstrip": false,
|
32 |
+
"single_word": false,
|
33 |
+
"special": true
|
34 |
+
},
|
35 |
+
"30526": {
|
36 |
+
"content": "<mask>",
|
37 |
+
"lstrip": true,
|
38 |
+
"normalized": false,
|
39 |
+
"rstrip": false,
|
40 |
+
"single_word": false,
|
41 |
+
"special": true
|
42 |
+
}
|
43 |
+
},
|
44 |
+
"bos_token": "<s>",
|
45 |
+
"clean_up_tokenization_spaces": true,
|
46 |
+
"cls_token": "<s>",
|
47 |
+
"do_basic_tokenize": true,
|
48 |
+
"do_lower_case": true,
|
49 |
+
"eos_token": "</s>",
|
50 |
+
"mask_token": "<mask>",
|
51 |
+
"model_max_length": 512,
|
52 |
+
"never_split": null,
|
53 |
+
"pad_token": "<pad>",
|
54 |
+
"sep_token": "</s>",
|
55 |
+
"strip_accents": null,
|
56 |
+
"tokenize_chinese_chars": true,
|
57 |
+
"tokenizer_class": "MPNetTokenizer",
|
58 |
+
"unk_token": "[UNK]"
|
59 |
+
}
|
vocab.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|