setfit-model / README.md

Add SetFit model

e6971db 6 months ago

No virus

29 kB

	---
	library_name: setfit
	tags:
	- setfit
	- sentence-transformers
	- text-classification
	- generated_from_setfit_trainer
	metrics:
	- accuracy
	widget:
	- text: tron miner contract game . yield diamond 50 times earnings '
	- text: infinite yield revolutionizing yield farming experience optimizing platform
	sustainability , capital efficiency safety . recognize def rapid shifting ecosystem
	. dedicated staying cutting edge creating strategies incentive , utilize compose
	latest secure tech . '
	- text: 'hmmcoin community driven def project . hmmcoin open source , decentralized
	, peer peer digital currency uses innovative coin distribution model ( reward
	level schedule ) . coin distribution model : reward level schedule contains 77
	levels . starts level 1 , request equal 1 pmc . one million coins requested level
	. million withdrawn coins amount received coins next level decrease 7 % . every
	level users need perform actions advance next level . level 1 : 1 pmc actions
	required : 1.000.000 actions level 2 : 0.93 pmc actions required : 1.075.268 actions
	etc . users get coin free pay transaction fee . '''
	- text: ' pump dump cryptocurrency trading game lets make real money buying selling
	coins virtual trading market . '''
	- text: platform daily dividend investments , long term contract 3 tier referral commissions
	. running tron blockchain verified smart contract . dividends starting 2.5 % unlimited
	booster contracts 5.5 % shorter periods . '
	pipeline_tag: text-classification
	inference: true
	base_model: sentence-transformers/paraphrase-mpnet-base-v2
	model-index:
	- name: SetFit with sentence-transformers/paraphrase-mpnet-base-v2
	results:
	- task:
	type: text-classification
	name: Text Classification
	dataset:
	name: Unknown
	type: unknown
	split: test
	metrics:
	- type: accuracy
	value: 0.5666990448437753
	name: Accuracy
	---

	# SetFit with sentence-transformers/paraphrase-mpnet-base-v2

	This is a [SetFit](https://github.com/huggingface/setfit) model 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.

	The model has been trained using an efficient few-shot learning technique that involves:

	1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning.
	2. Training a classification head with features from the fine-tuned Sentence Transformer.

	## Model Details

	### Model Description
	- Model Type: SetFit
	- Sentence Transformer body: [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2)
	- Classification head: a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance
	- Maximum Sequence Length: 512 tokens
	- Number of Classes: 9 classes
	<!-- - Training Dataset: [Unknown](https://huggingface.co/datasets/unknown) -->
	<!-- - Language: Unknown -->
	<!-- - License: Unknown -->

	### Model Sources

	- Repository: [SetFit on GitHub](https://github.com/huggingface/setfit)
	- Paper: [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055)
	- Blogpost: [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit)

	### Model Labels
	\| Label \| Examples \|
	\|:-------------\|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\|
	\| games \| <ul><li>"crypto football strategic live time free play play earn football ( soccer ) game . players play community members rank high order get rewarded $ ball tokens native token ( $ ball ) used house advertising system . ads displayed game accepting payments $ ball token users able buy formations teams app game specific amount $ ball . 50 80 % tokens collected players buys sent burn address . also , $ ball token going native cryptocurrency upcoming nft marketplace '"</li><li>"`` mint fee 1st 10000th nft 1 sol . mint fee 10001st 20000th nft 20000 fleet tokens . mint fee 20001st 30000th nft 30000 fleet tokens . mint fee 90001st 100000th nft 100000 fleet tokens , upper limit mint forever . 80 % chance mint sheep , 20 % chance mint wolf . wolves alpha property ranging 4 8 , value alpha equal chance mint . sheep stated , 10000 fleet tokens generated every day , 70 % claimed . stated wolves share 30 % total daily output fleet tokens according respective alpha values . example , total 8 sheep 2 wolves stated , total 80000 fleet tokens produced every day . sheep get 7000 fleet tokens . 2 wolves share 24000 fleet tokens . one wolf 's alpha 4 , wolf 's alpha 8 , former get 24000x4 ( 4+8 ) 8000 fleet tokens , latter gets 16000 fleet tokens . claim long time , sheep get fleet tokens last three days claiming , wolf get fleet tokens last day , wolf needs claim every day get maximum benefit . acquired fleet tokens traded dex , minted nft traded nft market . ''"</li><li>" hold.earth new app ethereal blockchain . whereby players buy plots virtual earth , creating designs colouring plots plot owners share fee income generated game . platform development shape games developers extend functionality layering game extensions hold earth app framework . hold.earth simplifies process buying using cryptocurrency , entry level users , simplifying process wallets , crypto purchases illustrating key elements blockchain technology . '"</li></ul> \|
	\| exchanges \| <ul><li>"klayswap complete chain instant swap protocol operates chain liquidity pool , liquidity guaranteed automated market making ( amp ) mechanisms . chain swap service allows anyone play act type token cryptocurrency become liquidity provider earn income transaction fee commissions . klayswap , ethereal based tokens ( eth , orc , dai , wbc ) transferred clayton ecosystem via orbit bridge , transparent ibc bridge , built orbit chain , offer yield farming asset pairings previously unconnected decentralized world '"</li><li>"excalibur exchange fandom based dex , built focus sustainability , capital efficiency supporting new projects . sustainability always issue def space , capital follows incentives , incentives need properly configured encourage longevity capital big focus excalibur , via dynamic incentivization long term taking yield generating governance token separated liquidity rewards , amongst others . excalibur also features highly variable swap fees variety mechanisms support new projects fandom . '"</li><li>"`` honeyswap decentralized exchange built dai chain , enables users experience fast secure transactions incredibly low fees . multiple tokens available swap add liquidity . 1hive honeyswap integrated , 0.05 % collected fees exchange used buy honey token deposit common pool , exchange 's volume increases , buy pressure honey increases well responsibility honey holders allocate tokens productive manner dao 's governance . ''"</li></ul> \|
	\| social \| <ul><li>"allows passively earn interest lending assets help liquidity . fast easy way earn money making secure safe investments active . novice used online decentralized trading , , put sale app website , ensuring monthly revenue allowing users purchase token well . novice escaping entire ecology users may participate trading digitally anywhere globe cheap transaction cost almost free , allowing liberating , strategic , low risk , frictionless experience fully decentralized manner using novice . community oriented money rewards holders distributing others . part decision making process , community members also included give sense ownership assets . burn fees , token burning involves permanently removing digital asset circulation reducing supply . '"</li><li>"violet garden better way connect earn . pass go collect 50 vio , every user , every day . unique identity todays social media plagued fake news , online harassment bots . fighting verifying one person , one account . get paid creating content bidding collect 10 % commission bid content creator previous bidder . instant payment utilising power eos , trades , transfers awards instant . voice voice , unfortunately , shut original social media platform abandoned community created , violet garden build stronger one living eos manner . '"</li><li>" would like introduce matic.tube , vod streaming service based polygon blockchain ! currently two movies available see potential project . '"</li></ul> \|
	\| defi \| <ul><li>" unlock nft potential intrinsic value nts finally unlocked used liquidity def operations permissionless strategy nts , rules , choose time rate loans : rules 100 % decentralized def innovation bridge nft world powerful def ecosystem , new era blockchain innovation permissionless anybody list nts without permission . anybody lend ada without ky . '"</li><li>"provide liquidity safe place . farm earn $ wave token . reviewed rugdoc , ky soon plain audit . able stake stablecoins avoid permanent loss stake lp tokens yield farm pools . community plays central role involved project . also referral program . refer earn 2 % purchased token . wave governance token future l2 . '"</li><li>"goldencake def app created ros.finance allow earn cake free ros holder '"</li></ul> \|
	\| marketplaces \| <ul><li>"seem.ninja offers variety services seem . example allows easy boarding seem card payments smart referral system users set prices . soon seem.ninja also allow everyone quickly purchase seem . '"</li><li>"metamarket aggressive finance smart chain ( bsc ) based nft marketplace . powered metareserve team group experienced knowledgeable blockchain developers , nft fanatics , fintech professionals metamarket aims collaborating brands celebrities create virtual real life utilities . designed offer plethora features users , metamarket nft marketplace , launched , reward potential buyers purchasing exclusive nts bsc based marketplace . first , holders shared private metamarket nft collections receive $ power governance tokens , offering chance vote significant events take place metareserve . additionally , receive rewards airdropped $ power periodically . metamarket users get whitelisted nft collections launch aggressive nft marketplace . '"</li><li>"burden powerful platform creators publish , sell , grow business around content web3 . '"</li></ul> \|
	\| collectibles \| <ul><li>"meta elephant nft series nts based mex mascot 4 grade rarities ( n r sr ssr ) total number 10,000. owners stake nts mine assets participate high quality games chains mex work future . first edition meta elephant nft themed space expedition includes cruise maintenance workers , cruise special forces , cruise navigators , cruise pilots . released certain marketplace form mystery box , chance get meta elephant nft different rarities ! '"</li><li>"`` replace crypto addresses name one time registration fee , renewal fees send crypto , need know recipient 's nft domain . send crypto one domain . worrying sending wrong address . nft domain leased . buy domain one time registration fee never worry renewal . domains stored wallet , like cryptocurrency . control . ''"</li><li>"this series 999 explores playful relationship color , shape , line harmony . artwork combination 9 zones ( 3 rows 3 ) . displayed within 9 zones 1 59 possible smaller pieces , put together makes one large artwork . zones programmed way see duplicated pieces within zones , thus making artwork unique one kind . utility offered 9 zone holders . info come . created toronto artist dan1el . '"</li></ul> \|
	\| gambling \| <ul><li>"bulls cows ( also known cows bulls pigs bulls bulls lots ) old code breaking mind paper pencil game two players , predating commercially marketed board game mastermind . game may date back century uses numbers words . played two opponents . '"</li><li>"casino , leading crypto casino thunder core network , 3,000 games world renowned game providers years experiences fully certified . also offers sports book even pp poker users create tournaments play friends . casino casino operated cybergalaxy b.v. , company registered established laws cacao . cybergalaxy b.v. licensed regulated antillephone n.v '"</li><li>"the highest dividends gambling app . need freeze , play get div ! 60 % house edge go dividends pool . currently , run test mode . 7 games available , auto bets limited max bet . find bug , please contact us telegram . start enjoy high div ! '"</li></ul> \|
	\| other \| <ul><li>"seed multichain , community participation token , encouraging participation creation distribution . sesameseed distributes seed daily reward participation delegated governance across blockchain represents . '"</li><li>"`` stackerdaos one stop shop bitcoin days . stackerdao protocol smart contracts operating system modular comparable days stacks . stackerdao , users form anything ranging completely decentralized days automatic proposal functionality subcats multisigs . stackerdaos provides code platform users launch manage dao based stackerdao operating system . stackerdaos , communities manage dao 's treasury , submit proposals , vote . stackerdaos also provides legal tech tools days avail legal benefits meet compliance obligations , legal entity formation , compliance forms filings , tax assistance. ``"</li><li>"name web3 marketing platform allowing users earn tokens , nts , chain chain rewards completing quests . name , aim onboard empower next billion users web3 providing everyone opportunities . name , see today , one step massive pursuit ambitions personal data ownership , data monetization opportunities , web3 accessibility , comparable , chain agnostic , decentralized social identity . since launch , offered 70,000+ verified users opportunity use data earn 220,000+ rewards quests name alone . users , new incredible quests name coolest companies projects . complete quests claim rewards . companies projects , name opportunity run campaigns accurate predictive results . run campaigns based chain chain verified actions . reward users directly data participation . name currently works going multi chain alongside introducing singular , cross chain , comparable web3 identity many exciting features users companies . '"</li></ul> \|
	\| high-risk \| <ul><li>"cross live lottery crc20 token rewards holders automatic mmf reward payments several times day . launched stealthily governed cream community , make win. ! main goal become largest reliable reward jackpot token cross network . '"</li><li>"get 50 % investments daily . '"</li><li>"`` proof easy invest forever ( poet ) world 's first three dimensional crypto currency generates eth holding tokens ! finally , working three level referral system drain contract ! 5 % easyinvestforever dev fees fees token holders multi level sticky referrals 10 % ( rugged ) buys invests . sell fee reduces time 48 % 8 % ( discourages early dumping ) function send external eth donating div ( directly sends eth token holders ) ability enable apps accept if tokens . zero fees transfers , enabling 3rd party trading . ''"</li></ul> \|

	## Evaluation

	### Metrics
	\| Label \| Accuracy \|
	\|:--------\|:---------\|
	\| all \| 0.5667 \|

	## Uses

	### Direct Use for Inference

	First install the SetFit library:

	```bash
	pip install setfit
	```

	Then you can load this model and run inference.

	```python
	from setfit import SetFitModel

	# Download from the 🤗 Hub
	model = SetFitModel.from_pretrained("dappradar/setfit-model")
	# Run inference
	preds = model("tron miner contract game . yield diamond 50 times earnings '")
	```

	<!--
	### Downstream Use

	List how someone could finetune this model on their own dataset.
	-->

	<!--
	### Out-of-Scope Use

	List how the model may foreseeably be misused and address what users ought not to do with the model.
	-->

	<!--
	## Bias, Risks and Limitations

	What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.
	-->

	<!--
	### Recommendations

	What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.
	-->

	## Training Details

	### Training Set Metrics
	\| Training set \| Min \| Median \| Max \|
	\|:-------------\|:----\|:--------\|:----\|
	\| Word count \| 7 \| 65.7639 \| 210 \|

	\| Label \| Training Sample Count \|
	\|:-------------\|:----------------------\|
	\| collectibles \| 8 \|
	\| defi \| 8 \|
	\| exchanges \| 8 \|
	\| gambling \| 8 \|
	\| games \| 8 \|
	\| high-risk \| 8 \|
	\| marketplaces \| 8 \|
	\| other \| 8 \|
	\| social \| 8 \|

	### Training Hyperparameters
	- batch_size: (8, 8)
	- num_epochs: (1, 1)
	- max_steps: -1
	- sampling_strategy: oversampling
	- num_iterations: 10
	- body_learning_rate: (2e-05, 1e-05)
	- head_learning_rate: 0.01
	- loss: CosineSimilarityLoss
	- distance_metric: cosine_distance
	- margin: 0.25
	- end_to_end: False
	- use_amp: False
	- warmup_proportion: 0.1
	- seed: 42
	- eval_max_steps: -1
	- load_best_model_at_end: True

	### Training Results
	\| Epoch \| Step \| Training Loss \| Validation Loss \|
	\|:-------:\|:-------:\|:-------------:\|:---------------:\|
	\| 0.0056 \| 1 \| 0.1857 \| - \|
	\| 0.2778 \| 50 \| 0.1235 \| - \|
	\| 0.5556 \| 100 \| 0.0434 \| - \|
	\| 0.8333 \| 150 \| 0.0148 \| - \|
	\| 1.0 \| 180 \| - \| 0.2355 \|

	* The bold row denotes the saved checkpoint.
	### Framework Versions
	- Python: 3.10.12
	- SetFit: 1.0.1
	- Sentence Transformers: 2.2.2
	- Transformers: 4.35.2
	- PyTorch: 2.1.0+cu121
	- Datasets: 2.15.0
	- Tokenizers: 0.15.0

	## Citation

	### BibTeX
	```bibtex
	@article{https://doi.org/10.48550/arxiv.2209.11055,
	doi = {10.48550/ARXIV.2209.11055},
	url = {https://arxiv.org/abs/2209.11055},
	author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren},
	keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences},
	title = {Efficient Few-Shot Learning Without Prompts},
	publisher = {arXiv},
	year = {2022},
	copyright = {Creative Commons Attribution 4.0 International}
	}
	```

	<!--
	## Glossary

	Clearly define terms in order to be accessible across audiences.
	-->

	<!--
	## Model Card Authors

	Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.
	-->

	<!--
	## Model Card Contact

	Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.
	-->