--- language: en library_name: setfit tags: - setfit - sentence-transformers - text-classification - generated_from_setfit_trainer base_model: sentence-transformers/multi-qa-mpnet-base-cos-v1 metrics: - f1 widget: - text: in durankulak near varna is another important example other signs of early metals are found from the third millennium bc in palmela portugal los millares spain and stonehenge united kingdom the precise beginnings however have not be clearly ascertained and new discoveries are both continuous and ongoing in tamilnadu in approximately 1900 bc ancient iron smelting sites were functioning in tamil nadu in the near east about 3500 bc it was discovered that by combining copper and tin a superior metal could be made an alloy called bronze this represented a major technological shift known as the bronze age the extraction of iron from its ore into a workable metal is much more difficult than for copper or tin the process appears to have been invented by the hittites in about 1200 bc beginning the iron age the secret of extracting and working iron was a key factor in the success of the philistineshistorical developments in ferrous metallurgy can be found in a wide variety of past cultures and civilizations this includes the ancient and medieval kingdoms and empires of the middle east and near east ancient iran ancient egypt ancient nubia and anatolia in presentday turkey ancient nok carthage the greeks and romans of ancient europe medieval europe ancient and medieval china ancient and medieval india ancient and medieval japan amongst others many applications practices and devices associated or involved in metallurgy were established in ancient china such as the innovation of the blast furnace cast iron hydraulicpowered trip hammers and double acting piston bellowsa 16th century book by georg agricola de re metallica describes the highly developed and complex processes of mining metal ores metal extraction and metallurgy of the time agricola has been described as the father of metallurgy extractive metallurgy is the practice of removing valuable metals from an ore and refining the extracted raw metals into a purer form in order to convert a metal oxide or sulphide to a purer metal the ore must be reduced physically chemically or electrolytically extractive metallurgists are interested in three primary streams feed concentrate metal oxidesulphide and tailings waste after mining large pieces of the ore feed are broken through crushing or grinding in order to obtain particles small enough where each particle is either mostly valuable or mostly waste concentrating the particles of value in a form supporting separation enables the desired metal to be removed from waste products mining may not be necessary if the ore body and physical environment are conducive to leaching leaching dissolves minerals in an ore body and results in an enriched solution the solution is collected and processed to extract valuable metals ore - text: '##rch procedure that evaluates the objective function p x displaystyle pmathbf x on a grid of candidate source locations g displaystyle mathcal g to estimate the spatial location of the sound source x s displaystyle textbf xs as the point of the grid that provides the maximum srp modifications of the classical srpphat algorithm have been proposed to reduce the computational cost of the gridsearch step of the algorithm and to increase the robustness of the method in the classical srpphat for each microphone pair and for each point of the grid a unique integer tdoa value is selected to be the acoustic delay corresponding to that grid point this procedure does not guarantee that all tdoas are associated to points on the grid nor that the spatial grid is consistent since some of the points may not correspond to an intersection of hyperboloids this issue becomes more problematic with coarse grids since when the number of points is reduced part of the tdoa information gets lost because most delays are not anymore associated to any point in the grid the modified srpphat collects and uses the tdoa information related to the volume surrounding each spatial point of the search grid by considering a modified objective function where l m 1 m 2 l x displaystyle lm1m2lmathbf x and l m 1 m 2 u x displaystyle lm1m2umathbf x are the lower and upper accumulation limits of gcc delays which depend on the spatial location x displaystyle mathbf x the accumulation limits can be calculated beforehand in an exact way by exploring the boundaries separating the regions corresponding to the points of the grid alternatively they can be selected by considering the spatial gradient of the tdoa ∇ τ m 1 m 2 x ∇ x τ m 1 m 2 x ∇ y τ m 1 m 2 x ∇ z τ m 1 m 2 x t displaystyle nabla tau m1m2mathbf x nabla xtau m1m2mathbf x nabla ytau m1m2mathbf x nabla ztau m1m2mathbf x t where each component γ ∈ x y z displaystyle gamma in leftxyzright of the gradient is for a rectangular grid where neighboring points are separated a distance r displaystyle r the lower and upper accumulation limits are given by where d r 2 min 1 sin θ cos [UNK] 1 sin θ sin [UNK] 1 cos θ displaystyle dr2min leftfrac 1vert sintheta cosphi vert frac 1vert sintheta sinphi vert frac 1vert' - text: authority to select projects and mandated new metropolitan planning initiatives for the first time state transportation officials were required to consult seriously with local representatives on mpo governing boards regarding matters of project prioritization and decisionmaking these changes had their roots in the need to address increasingly difficult transportation problems — in particular the more complicated patterns of traffic congestion that arose with the suburban development boom in the previous decades many recognized that the problems could only be addressed effectively through a stronger federal commitment to regional planning the legislation that emerged the intermodal surface transportation efficiency act istea was signed into federal law by president george h w bush in december 1991 it focused on improving transportation not as an end in itself but as the means to achieve important national goals including economic progress cleaner air energy conservation and social equity istea promoted a transportation system in which different modes and facilities — highway transit pedestrian bicycle aviation and marine — were integrated to allow a seamless movement of both goods and people new funding programs provided greater flexibility in the use of funds particularly regarding using previously restricted highway funds for transit development improved intermodal connections and emphasized upgrades to existing facilities over building new capacity — particularly roadway capacity to accomplish more serious metropolitan planning istea doubled federal funding for mpo operations and required the agencies to evaluate a variety of multimodal solutions to roadway congestion and other transportation problems mpos also were required to broaden public participation in the planning process and to see that investment decisions contributed to meeting the air quality standards of the clean air act amendments in addition istea placed a new requirement on mpos to conduct fiscally constrained planning and ensure that longrange transportation plans and shortterm transportation improvement programs were fiscally constrained in other words adopted plans and programs can not include more projects than reasonably can be expected to be funded through existing or projected sources of revenues this new requirement represented a major conceptual shift for many mpos and others in the planning community since the imposition of fiscal discipline on plans now required not only understanding how much money might be available but how to prioritize investment needs and make difficult choices among competing needs adding to this complexity is the need to plan across transportation modes and develop approaches for multimodal investment prioritization and decision making it is in this context of greater prominence funding and requirements that mpos function today an annual element is composed of transportation improvement projects contained in an areas transportation improvement program tip which is proposed for implementation during the current year the annual element is submitted to the us department of transportation as part of the required planning process the passage of safe accountable flexible efficient transportation equity act a legacy for users safetealu - text: '##pignygiroux served as an assistant professor from 1997 2003 associate professor from 2003 2014 chair of the department of geography from 2015 2018 and professor beginning in 2014 with secondary appointments in department of geology the college of education social services and rubenstein school of environment natural resources she teaches courses in meteorology climatology physical geography remote sensing and landsurface processes in her work as state climatologist for vermont dupignygiroux uses her expertise hydrology and extreme weather such as floods droughts and storms to keep the residents of vermont informed on how climate change will affect their homes health and livelihoods she assists other state agencies in preparing for and adapting to current and future impacts of climate change on vermonts transportation system emergency management planning and agriculture and forestry industries for example she has published analyses of the impacts of climate change on the health of vermonts sugar maples a hardwood species of key economic and cultural importance to the state as cochair of vermonts state ’ s drought task force she played a key role in developing the 2018 vermont state hazard mitigation plandupignygiroux served as secretary for the american association of state climatologists from 20102011 and president elect from 20192020 in june 2020 she was elected as president of the american association of state climatologists which is a twoyear term in addition to her research on climate change dupignygiroux is known for her efforts to research and promote climate literacy climate literacy is an understanding of the influences of and influences on the climate system including how people change the climate how climate metrics are observed and modelled and how climate change affects society “ being climate literate is more critical than ever before ” lesleyann dupignygiroux stated for a 2020 article on climate literacy “ if we do not understand weather climate and climate change as intricate and interconnected systems then our appreciation of the big picture is lost ” dupignygiroux is known for her climate literacy work with elementary and high school teachers and students she cofounded the satellites weather and climate swac project in 2008 which is a professional development program for k12 teachers designed to promote climate literacy and interest in the stem science technology engineering and mathematics careers dupignygiroux is also a founding member of the climate literacy and energy awareness network clean formerly climate literacy network a communitybased effort to support climate literacy and communication in a 2016 interview dupignygiroux stated “ sharing knowledge and giving back to my community are my two axioms in life watching students mature and flourish in' - text: no solutions to x n y n z n displaystyle xnynzn for all n ≥ 3 displaystyle ngeq 3 this claim appears in his annotations in the margins of his copy of diophantus euler the interest of leonhard euler 1707 – 1783 in number theory was first spurred in 1729 when a friend of his the amateur goldbach pointed him towards some of fermats work on the subject this has been called the rebirth of modern number theory after fermats relative lack of success in getting his contemporaries attention for the subject eulers work on number theory includes the following proofs for fermats statements this includes fermats little theorem generalised by euler to nonprime moduli the fact that p x 2 y 2 displaystyle px2y2 if and only if p ≡ 1 mod 4 displaystyle pequiv 1bmod 4 initial work towards a proof that every integer is the sum of four squares the first complete proof is by josephlouis lagrange 1770 soon improved by euler himself the lack of nonzero integer solutions to x 4 y 4 z 2 displaystyle x4y4z2 implying the case n4 of fermats last theorem the case n3 of which euler also proved by a related method pells equation first misnamed by euler he wrote on the link between continued fractions and pells equation first steps towards analytic number theory in his work of sums of four squares partitions pentagonal numbers and the distribution of prime numbers euler pioneered the use of what can be seen as analysis in particular infinite series in number theory since he lived before the development of complex analysis most of his work is restricted to the formal manipulation of power series he did however do some very notable though not fully rigorous early work on what would later be called the riemann zeta function quadratic forms following fermats lead euler did further research on the question of which primes can be expressed in the form x 2 n y 2 displaystyle x2ny2 some of it prefiguring quadratic reciprocity diophantine equations euler worked on some diophantine equations of genus 0 and 1 in particular he studied diophantuss work he tried to systematise it but the time was not yet ripe for such an endeavour — algebraic geometry was still in its infancy he did notice there was a connection between diophantine problems and elliptic integrals whose study he had himself initiated lagrange legendre and gauss josephlouis pipeline_tag: text-classification inference: true model-index: - name: SetFit with sentence-transformers/multi-qa-mpnet-base-cos-v1 on Wiki Labeled Articles results: - task: type: text-classification name: Text Classification dataset: name: Wiki Labeled Articles type: unknown split: test metrics: - type: f1 value: 0.7832245351020043 name: F1 --- # SetFit with sentence-transformers/multi-qa-mpnet-base-cos-v1 on Wiki Labeled Articles This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification. This SetFit model uses [sentence-transformers/multi-qa-mpnet-base-cos-v1](https://huggingface.co/sentence-transformers/multi-qa-mpnet-base-cos-v1) as the Sentence Transformer embedding model. A [SetFitHead](huggingface.co/docs/setfit/reference/main#setfit.SetFitHead) 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/multi-qa-mpnet-base-cos-v1](https://huggingface.co/sentence-transformers/multi-qa-mpnet-base-cos-v1) - **Classification head:** a [SetFitHead](huggingface.co/docs/setfit/reference/main#setfit.SetFitHead) instance - **Maximum Sequence Length:** 512 tokens - **Number of Classes:** 43 classes - **Language:** en ### 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 | |:------|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | 12 | | | 30 | | | 42 | | | 25 | | | 38 | | | 32 | | | 14 | | | 5 | | | 16 | | | 28 | | | 41 | | | 29 | | | 4 | | | 19 | | | 26 | | | 20 | | | 7 | | | 10 | | | 8 | | | 0 | | | 11 | | | 3 | | | 1 | | | 6 | | | 23 | | | 18 | | | 39 | | | 2 | | | 40 | | | 17 | | | 37 | | | 27 | | | 9 | | | 34 | | | 33 | | | 15 | | | 31 | | | 35 | | | 36 | | | 21 | | | 13 | | | 22 | | | 24 | | ## Evaluation ### Metrics | Label | F1 | |:--------|:-------| | **all** | 0.7832 | ## 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("udrearobert999/multi-qa-mpnet-base-cos-v1-contrastive-logistic-500s") # Run inference preds = model("##rch procedure that evaluates the objective function p x displaystyle pmathbf x on a grid of candidate source locations g displaystyle mathcal g to estimate the spatial location of the sound source x s displaystyle textbf xs as the point of the grid that provides the maximum srp modifications of the classical srpphat algorithm have been proposed to reduce the computational cost of the gridsearch step of the algorithm and to increase the robustness of the method in the classical srpphat for each microphone pair and for each point of the grid a unique integer tdoa value is selected to be the acoustic delay corresponding to that grid point this procedure does not guarantee that all tdoas are associated to points on the grid nor that the spatial grid is consistent since some of the points may not correspond to an intersection of hyperboloids this issue becomes more problematic with coarse grids since when the number of points is reduced part of the tdoa information gets lost because most delays are not anymore associated to any point in the grid the modified srpphat collects and uses the tdoa information related to the volume surrounding each spatial point of the search grid by considering a modified objective function where l m 1 m 2 l x displaystyle lm1m2lmathbf x and l m 1 m 2 u x displaystyle lm1m2umathbf x are the lower and upper accumulation limits of gcc delays which depend on the spatial location x displaystyle mathbf x the accumulation limits can be calculated beforehand in an exact way by exploring the boundaries separating the regions corresponding to the points of the grid alternatively they can be selected by considering the spatial gradient of the tdoa ∇ τ m 1 m 2 x ∇ x τ m 1 m 2 x ∇ y τ m 1 m 2 x ∇ z τ m 1 m 2 x t displaystyle nabla tau m1m2mathbf x nabla xtau m1m2mathbf x nabla ytau m1m2mathbf x nabla ztau m1m2mathbf x t where each component γ ∈ x y z displaystyle gamma in leftxyzright of the gradient is for a rectangular grid where neighboring points are separated a distance r displaystyle r the lower and upper accumulation limits are given by where d r 2 min 1 sin θ cos [UNK] 1 sin θ sin [UNK] 1 cos θ displaystyle dr2min leftfrac 1vert sintheta cosphi vert frac 1vert sintheta sinphi vert frac 1vert") ``` ## Training Details ### Training Set Metrics | Training set | Min | Median | Max | |:-------------|:----|:---------|:----| | Word count | 1 | 369.0421 | 509 | | Label | Training Sample Count | |:------|:----------------------| | 0 | 500 | | 1 | 500 | | 2 | 420 | | 3 | 500 | | 4 | 356 | | 5 | 374 | | 6 | 500 | | 7 | 364 | | 8 | 422 | | 9 | 372 | | 10 | 494 | | 11 | 295 | | 12 | 500 | | 13 | 278 | | 14 | 314 | | 15 | 500 | | 16 | 417 | | 17 | 379 | | 18 | 357 | | 19 | 370 | | 20 | 337 | | 21 | 373 | | 22 | 500 | | 23 | 500 | | 24 | 312 | | 25 | 481 | | 26 | 386 | | 27 | 500 | | 28 | 500 | | 29 | 500 | | 30 | 500 | | 31 | 470 | | 32 | 284 | | 33 | 311 | | 34 | 500 | | 35 | 318 | | 36 | 500 | | 37 | 500 | | 38 | 500 | | 39 | 500 | | 40 | 500 | | 41 | 500 | | 42 | 336 | ### Training Hyperparameters - batch_size: (32, 32) - num_epochs: (4, 8) - max_steps: -1 - sampling_strategy: oversampling - num_iterations: 20 - body_learning_rate: (3e-05, 0.01) - head_learning_rate: 0.01 - loss: SupConLoss - distance_metric: cosine_distance - margin: 0.25 - end_to_end: False - use_amp: False - warmup_proportion: 0.1 - max_length: 512 - seed: 42 - eval_max_steps: -1 - load_best_model_at_end: True ### Training Results | Epoch | Step | Training Loss | Validation Loss | |:----------:|:-------:|:-------------:|:---------------:| | 0.0015 | 1 | 2.182 | - | | 0.3671 | 250 | 1.0321 | - | | **0.7342** | **500** | **1.01** | **0.9291** | | 1.1013 | 750 | 0.7586 | - | | 1.4684 | 1000 | 0.2408 | 0.9875 | | 1.8355 | 1250 | 0.8995 | - | | 2.2026 | 1500 | 0.3702 | 0.9411 | | 2.5698 | 1750 | 0.669 | - | | 2.9369 | 2000 | 0.2361 | 0.9538 | | 3.3040 | 2250 | 0.1108 | - | | 3.6711 | 2500 | 0.5895 | 0.9276 | | 0.0017 | 1 | 0.0591 | - | | 0.4371 | 250 | 0.3805 | - | | **0.8741** | **500** | **0.5506** | **0.9742** | | 1.3112 | 750 | 0.5571 | - | | 1.7483 | 1000 | 0.1259 | 1.1268 | | 2.1853 | 1250 | 0.7435 | - | | 2.6224 | 1500 | 0.7133 | 1.1094 | | 3.0594 | 1750 | 0.0812 | - | | 3.4965 | 2000 | 0.3421 | 1.2851 | | 3.9336 | 2250 | 0.0722 | - | * The bold row denotes the saved checkpoint. ### Framework Versions - Python: 3.10.12 - SetFit: 1.0.3 - Sentence Transformers: 2.7.0 - Transformers: 4.40.1 - PyTorch: 2.2.1+cu121 - Datasets: 2.19.1 - Tokenizers: 0.19.1 ## 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} } ```