![phoenix](https://d1a29h5kxv3oc2.cloudfront.net/dist/img/phoenix.81914efc7207f6a059bd10f5d3fd5f89.png)
Nlp Projects
Looking for freelance Nlp jobs and project work? PeoplePerHour has you covered.
ML Engineer: Develop offline LLM
Project Objective: Develop an offline Large Language Model (LLM) for UK legal professionals with advanced NLP capabilities. Will also include a simple CRM that is updated automatically. Key Features: - Chat interface with q and a - Document generation (contracts, engagement letter's etc.) - Audio transcription for user input for chatbot (alternative to typing) and TTS ouput from chat bot. -Audio transcription for client meetings and update CRM automatically - Legal research with up-to-date UK law database and uk case law - Split-screen citation display Tech Stack: - Open-source, commercially licensable frameworks (e.g., AnythingLLM, Ollama) - MIT-licenced codebase for customisation - Retrieval-Augmented Generation (RAG) for enhanced context handling and/ or Grokking Additional Considerations: - Exploring grokking techniques for improved model comprehension. This would be preferred to RAG if the claimed 99%+ accuracy is achievable - Potential integration of unsloth for optimised training pipelines - software would have to run locally and entirely offline Project Background: - Established relationships with UK legal sector - Validated market demand through discussions with Law Society representatives - Clear path to commercialisation Ideal Candidate: - Strong ML/NLP expertise - Experience with LLM fine-tuning and deployment - Familiarity with legal domain (preferred) - Passion for developing AI solutions with real-world impact If you're excited about this project and have relevant expertise, I'd love to hear from you.
23 days ago30 proposalsRemoteopportunity
ML Engineer for Legal AI Project: Develop offline LLM
Project Objective: Develop an offline LLM for UK legal professionals with advanced NLP capabilities. Key Features: - Chat interface - Document generation (contracts, engagement letters etc.) - Audio transcription from; user input (instead of typing), client meetings - Legal research with up-to-date UK law database and uk case law - Split-screen citation display Tech Stack: - Open-source, commercially licensable frameworks (e.g., AnythingLLM, Ollama) - MIT-licenced codebase for customisation - Retrieval-Augmented Generation (RAG) for enhanced context handling Additional Considerations: - Exploring grokking techniques for improved model comprehension. This method is preferable over RAG vector database if the the grokked transformer can achieve 99% accuracy for retrieval. - Potential integration of unsloth for optimised training pipelines -The 'finished' software must run locally on consumer grade hardware and be completely offline Project Background: - Established relationships with UK legal sector - Validated market demand through discussions with Law Society representatives - Clear path to commercialisation Ideal Candidate: - Strong ML/NLP expertise - Experience with LLM fine-tuning and deployment - Familiarity with legal domain (preferred) - Passion for developing AI solutions with real-world impact If you're excited about this project and have relevant expertise, I'd love to hear from you. This project offers a unique opportunity to shape the future of legal technology in the UK.
23 days ago26 proposalsRemote
Past "Nlp" Projects
opportunitypre-funded
Human Value Index
I am an artist looking to create a visual work that appears to be a stock market style tika-tape display formed of a live data stream, scoring nations by numbers of violent deaths through conflict divided by MSM coverage. I believe that creating a live index to monitor and rank countries based on media coverage of violent deaths, due to conflict, involves several key steps, including data collection, natural language processing, sentiment analysis, and real-time data updating. I will need to gather articles and news content from multiple media sources. This can be achieved using web scraping, RSS feeds, or media APIs. As well as prepare the collected data by cleaning and processing it to extract relevant information about violent deaths and conflicts. Then, use NLP techniques to identify and extract information about violent deaths and conflicts from the articles. Analyze the sentiment and classify the extracted events to ensure they are related to violent deaths due to conflict. I will need to develop a scoring system to quantify the coverage and create a live index as well as implement a pipeline to update the index in real-time as new articles are published. I can contribute to some of this, as far as aesthetics and theory is concerned but where I need help is making the data then code into a live Wordpress page or similar - a kind of online countdown clock and something I can then use to display in tick-tape form across buildings or in galleries. I have already been playing with Chatgpt (obviously) and a lot of the information is there, but its taking the Python based code that can run these elements and turning it into a visual web page that Im looking for some one to assist on.
I need an AI chatbot with voice phone ability
We are seeking a talented AI Chatbot Developer or team. Indcluded in the responsibility is designing, developing, and implementing cutting-edge chatbot solutions powered by artificial intelligence. You will work closely with cross-functional teams to understand requirements, build conversational interfaces, integrate natural language processing (NLP) capabilities, and continuously improve chatbot performance. Responsibilities: Design and Development: Design, develop, and deploy AI-driven chatbots tailored to meet specific business objectives and user needs. Natural Language Processing (NLP): Integrate NLP algorithms and techniques to enable chatbots to understand and respond to user queries in natural language effectively. Conversation Design: Create conversational flows, intents, entities, and dialogues to ensure seamless interactions between users and chatbots across various platforms. Integration: Collaborate with backend developers to integrate chatbots with existing systems, databases, and APIs to fetch and update information dynamically. Machine Learning: Implement machine learning algorithms to enhance chatbot capabilities, such as sentiment analysis, user behavior prediction, and personalized recommendations. Testing and Optimization: Conduct thorough testing of chatbots to identify and fix issues related to functionality, performance, and user experience. Continuously optimize chatbots based on feedback and analytics data. Documentation: Create comprehensive documentation including technical specifications, user guides, and troubleshooting procedures for chatbot development and maintenance. Stay Updated: Stay updated with the latest trends, technologies, and best practices in AI, NLP, and chatbot development. Leverage new tools and frameworks to improve chatbot functionality and performance.
Wordpress website with NLU and NLP redaction from documents
Wordpress website with NLU and NLP redaction from uploaded documents
opportunity
Data extraction and ChatGPT integration
To include ChatGPT in an app created on an app designed to assist athletes by answering questions and providing personalised explosive workout plans, sourced from PDFs of public domain books Step 1: Data Extraction from PDFs Handle Non-text Content: If the PDFs contain important non-text elements (like images of exercises), consider using an OCR (Optical Character Recognition) tool like Tesseract to convert these images to text. Step 2: Data Cleaning and Preparation Process: Clean Extracted Data: Use Python to clean the data, including removing unwanted characters, standardizing terminology, and correcting OCR errors. Structure Data: Convert cleaned data into a structured format like JSON or CSV, categorizing content by topics such as "speed training", "strength workouts", etc. Natural Language Processing: Apply NLP techniques to refine the text, extract keywords, and prepare it for easy querying. Configure API calls to send user queries to ChatGPT and receive responses. Store structured data from PDFs in Bubble’s database for quick reference by the AI.
opportunity
HyperScene: AI-Powered High-Fidelity Video Generation
We are seeking a highly skilled and experienced freelancer to develop a cutting-edge software application named HyperScene. This software will utilize advanced artificial intelligence (AI) to generate incredibly realistic and high-fidelity video scenes based solely on text descriptions. Project Scope HyperScene will be a desktop application with the following core functionalities: Natural Language Processing (NLP) Engine: This engine will analyze user-provided text descriptions of a scene, including details on environment, objects, lighting, characters, and actions. AI Scene Generation: Utilizing a deep learning model, HyperScene will generate a realistic video scene based on the analyzed text description. The video should be indistinguishable from real footage in terms of detail, lighting, and motion. Customization Options: Users should be able to fine-tune the generated scene with various parameters like camera angles, weather conditions, and the level of detail. Output Formats: HyperScene will export the generated videos in industry-standard formats like MP4 and MOV for seamless integration into editing workflows. Desired Qualities High-Fidelity Video Generation: The primary focus is achieving the most realistic visuals possible. The generated videos should be indistinguishable from actual footage captured with high-end cameras. Intuitive User Interface (UI): The user interface should be user-friendly and facilitate a seamless workflow for creating text descriptions and customizing the generated video. Efficiency: HyperScene should be able to generate video scenes within a reasonable timeframe, balancing quality with processing speed. Scalability: The software should be designed to scale and handle increasingly complex scene descriptions in the future. Freelancer Budget Proposal Due to the highly technical nature of this project, we are open to budget proposals from qualified freelancers. Please consider the following factors when outlining your proposed budget: Your experience with AI development and deep learning models. Proposed development timeline and milestones. Technologies and frameworks you plan to use. Your approach to ensuring high-fidelity video generation. We are committed to providing a competitive compensation package for the right freelancer who can deliver on this groundbreaking project. Selection Process We will evaluate proposals based on the following criteria: Technical Expertise: Your demonstrable experience in AI development, particularly deep learning models for video generation. Portfolio & References: Strong examples of previous work, including projects showcasing your skills in AI and video editing. Budget Proposal: A comprehensive breakdown of your proposed costs and timeline for development. Communication Skills: The ability to clearly explain technical concepts and maintain consistent communication throughout the project.
Identity Matching in News Articles with Machine Learning and NER
Use Case: Our objective is to enhance our capability to accurately identify individuals mentioned in news articles and match them with corresponding entries in our database. Specifically, we aim to develop a machine learning model capable of discerning whether a person entity referenced in an article aligns with an individual stored within our database. This NER model will assess attributes associated with individuals in our database, and return us with a Identity Matching Scoring. Technical Requirements: This task involves crafting and refining a machine learning model capable of Named Entity Recognition and Linking, considering the contextual relevance of attributes associated with individuals mentioned in news articles. The model should prioritize certain attributes over others based on their relevance and significance for accurate identification and matching. We are seeking expertise in the following areas: Named Entity Recognition (NER) techniques Natural Language Processing (NLP) and Machine Learning (ML) algorithms Attribute weighting and relevance assessment
Complete Explanation For Data Science Resume Projects
I want help in explaining my data science projects in interview in a detailed manner . I am data science professional looking for a job switch of 3 years. Projects: 1) CONSUMER COMPLAINT CLASSIFICATION (NLP) Building an API and training model to classify future complaints based on its content for a banking firm. The dataset is of 2 million rows, 5+ years historical data. Skills/Technology: Python, SVM classification, Random Forest ,Flask, Glove ,Word2Vec 2) PREDICTIVE MAINTENANCE Built an end to end machine learning model for a heavy industry firm which records different features like power,temperature etc for machines and predicting whether failure will happen or not. Skill/Technology: Python, FastAPI, Digital Ocean.,Streamlit,Docker 3) RECOMMENDATION ENGINE FOR AUTOMATED TRADING PLATFORM Developed custom Technical Indicators Functions to analyse historical price data and market trends, triggering buying signals when specific conditions were met. Empowered traders clients with actionable insights by providing timely buy signals aligned with market trends. Skill/Technology: TA-Lib, TensorFlow, NumPy, Pandas, sklearn, statsmodels 4) INSURANCE POLICY CROSS SELLING A classification/ranking project aimed to detect health insurance customers most likely to buy a new type of insurance from the company - car insurance. To solve this problem a machine learning model was built to classify the customers by their probability of buying the insurance. The Heroku platform was used to deploy the ML model, which will respond to requests via API. Skill/Technology: XGBoost,LightGBM, NumPy, Flask,Heroku 5) CUSTOMER CHURN PREDICTION FOR A MALAYSIAN BANK Developed a model to analyse customer data and predict churn to boost customer retention.Employed statistical techniques on customer data using Pandas, Seaborn, and Sklearn. Reduced customer churn rate by 7%, leading to increased revenue, lower marketing costs, and enhanced customer loyalty.
AI/Prompt expert
Hello everyone, Job Description: We are seeking a talented AI/Prompt Expert to join our team. The ideal candidate will have a strong background in natural language processing (NLP), machine learning, and deep learning, with a focus on generative models. The AI/Prompt Expert will be responsible for developing and fine-tuning AI models, such as GPT-3 and DALL·E, to generate creative and contextually relevant content in response to user prompts. The role involves working closely with cross-functional teams to understand user requirements and develop innovative solutions using cutting-edge AI technologies. Responsibilities: Research, develop, and implement state-of-the-art AI models for natural language generation. Train and fine-tune generative models to produce high-quality and diverse outputs. Collaborate with product managers, engineers, and designers to integrate AI models into user-facing applications. Analyze user interaction data to improve model performance and user experience. Stay updated with the latest advancements in NLP, machine learning, and generative models. Requirements: Master's or Ph.D. in computer science, artificial intelligence, or a related field. Solid understanding of NLP, deep learning, and generative modeling techniques. Proficiency in Python and experience with deep learning frameworks such as TensorFlow or PyTorch. Strong problem-solving skills and the ability to work in a fast-paced, collaborative environment. Excellent communication skills and the ability to explain complex AI concepts to non-technical stakeholders. Preferred Qualifications: Experience with large-scale language models such as GPT-3 and DALL·E. Research publications or contributions to the field of NLP and generative modeling. Familiarity with cloud-based AI services and deployment. If you are passionate about AI and have a strong background in NLP and generative modeling, we encourage you to apply for this exciting opportunity. Thank you.
opportunity
Build a AI chatbot for our Instagram DMs
Project Brief: AI-Powered Instagram Chatbot for Singer-Songwriter Outreach Objective: To develop an AI-powered chatbot capable of handling Instagram Direct Messages (DMs) to engage with singer-songwriters, nurture them as leads, and encourage them to book consultations for production and recording services. The chatbot should simulate human-like interactions while streamlining the conversation process, reducing manual effort, and increasing conversion rates. Background: For the past 7 years, Instagram has been the primary platform for customer acquisition. However, managing DM conversations has become increasingly time-consuming. To address this, we aim to implement an AI chatbot that can engage with potential clients, maintaining the personalized touch of human interaction. Instagram: @theonlinerecordingstudio Website: onlinerecordingstudio.com Requirements: 1. AI-Powered Chatbot: Develop a chatbot capable of processing and responding to Instagram DMs using natural language processing (NLP) and machine learning techniques. Manychat seems the best option. 2. Human-like Interaction: Ensure that the chatbot's responses mimic human conversation patterns, tone, and empathy to establish a genuine connection with singer-songwriters. 3. Lead Nurturing: Implement lead nurturing strategies within the chatbot's dialogue flow to guide potential clients through the sales funnel effectively. 4. Consultation Booking: Design the chatbot to encourage singer-songwriters to book consultations for production and recording services by highlighting the benefits and offering convenient scheduling options. 5. Integration with Existing Outreach Script: Incorporate the existing outreach script and chatflow into the chatbot's architecture to maintain consistency and familiarity with the brand's messaging. 6. Learning from Successful Interactions: Utilize data from successful interactions and conversions over the past 6 years to train the chatbot, improving its effectiveness in engaging and converting leads. 7. Scalability and Flexibility: Build the chatbot with scalability and flexibility in mind to accommodate future updates, changes in messaging, and evolving customer needs. 8. Compliance: Ensure compliance with Instagram's terms of service and data privacy regulations throughout the development process. Budget: The budget allocation for this project will be discussed and finalized based on the proposed scope of work, resources required, and project timeline. Note: The success of this project hinges on the ability to create a chatbot that seamlessly integrates with the existing Instagram DM workflow while providing a personalized and efficient experience for singer-songwriters. Constant monitoring, feedback, and refinement will be essential to optimize the chatbot's performance and drive meaningful conversions.
opportunity
AI Project with in Python and Chat GPT
We seek an experienced AI developer to spearhead development of a conversational agent using Python and GPT technologies. This individual will lead a small team of three others in building an intelligent bot capable of natural language interactions. The ideal candidate will have extensive hands-on experience designing and implementing similar advanced NLP systems. The goal of this project is to create an engaging and helpful virtual assistant that can understand users' requests and reply with relevant, nuanced responses. The developer will be responsible for architecting the bot's underlying framework, ensuring its language models and dialog capabilities are properly structured to support human-like conversations across a variety of topics. Strong Python skills are a must to effectively utilize GPT's powerful language generation. The team will focus on iteratively improving the agent's response quality, semantics, and appropriateness. It should learn from interactions and refine its knowledge over time. The lead developer will oversee integrating relevant data sources and selecting the most appropriate algorithms and training approaches. Natural language understanding, personalized recommendations, and conversational flows also need consideration. Excellent communication and collaboration skills are essential as this individual will be the primary point of contact coordinating across the small team. Regular stand-ups and demos will keep stakeholders informed of progress. A strong portfolio demonstrating past successes developing conversational interfaces is preferred. Experience managing or mentoring others is a plus but not required. The right candidate will have an innovative
Complete Explanation For Data Science Resume Projects
I want help in explaining my data science projects in interview in a detailed manner . I am data science professional looking for a job switch of 3 years. Projects: 1) CONSUMER COMPLAINT CLASSIFICATION (NLP) Building an API and training model to classify future complaints based on its content for a banking firm. The dataset is of 2 million rows, 5+ years historical data. Skills/Technology: Python, SVM classification, Random Forest ,Flask, Glove ,Word2Vec 2) PREDICTIVE MAINTENANCE Built an end to end machine learning model for a heavy industry firm which records different features like power,temperature etc for machines and predicting whether failure will happen or not. Skill/Technology: Python, FastAPI, Digital Ocean.,Streamlit,Docker 3) RECOMMENDATION ENGINE FOR AUTOMATED TRADING PLATFORM Developed custom Technical Indicators Functions to analyse historical price data and market trends, triggering buying signals when specific conditions were met. Empowered traders clients with actionable insights by providing timely buy signals aligned with market trends. Skill/Technology: TA-Lib, TensorFlow, NumPy, Pandas, sklearn, statsmodels 4) INSURANCE POLICY CROSS SELLING A classification/ranking project aimed to detect health insurance customers most likely to buy a new type of insurance from the company - car insurance. To solve this problem a machine learning model was built to classify the customers by their probability of buying the insurance. The Heroku platform was used to deploy the ML model, which will respond to requests via API. Skill/Technology: XGBoost,LightGBM, NumPy, Flask,Heroku 5) CUSTOMER CHURN PREDICTION FOR A MALAYSIAN BANK Developed a model to analyse customer data and predict churn to boost customer retention.Employed statistical techniques on customer data using Pandas, Seaborn, and Sklearn. Reduced customer churn rate by 7%, leading to increased revenue, lower marketing costs, and enhanced customer loyalty.
opportunity
Automated Semantic Text Analysis Pipeline
Comprehensive Use Case Specification: Automated Semantic Text Analysis Pipeline Objective Develop an automated semantic text analysis pipeline that processes and analyses textual data extracted from documents. This pipeline enriches text with metadata for deeper insights and enables semantic search capabilities through a user-friendly interface. This stage of the project is for a MVP system should leverage AWS services such as Textract for text extraction, a text categorising stage with a simple to use GUI, all-mpnet-base-v2 for embedding, and Postgres with a vector extension. This job posting is for the MPV stage only, but we must be mindful of the stage two development and facilitate rapid and straightforward scalability in any stage one MPV processes. System Overview The solution encompasses AWS services for storage and processing, a custom interface for metadata enrichment, all-mpnet-base-v2 for generating text embeddings, Postgres and a vector extension for efficient storage and retrieval of vectors, and a custom-built web interface for user interaction. RAG will be implemented with a broad a context as possible to the model across a large document set. Phase 1: MVP Stage 1. Document Storage and Processing Trigger - Tool: Amazon S3. - Process: Upload documents (PDFs initially) to designated S3 buckets, documents will be remained in accordance with a set naming convention and key metadata relating to the document entered into the database for future reference. This triggers the subsequent text extraction process. For test purposes the uploads will be made manually, and at later stages a web scraper will be added that automatically places PDF documents into relevant S3 buckets. 2. Text Extraction - Tool: AWS Textract. - Process: Text is extracted from uploaded PDF documents and temporarily stored in A3 buckets to facilitate further processing. 3. Text Enrichment Developer to advise on best method of adding labels / categories to the text, via an easy to use interface. Labels to be added at a granular level to allow the return of text snippets, providing context to the LLM in formulating it's responses from a broad range of documents without exceeding the token limit. 4. Text Vectorization - Embedding tool: all-mpnet-base-v2 - LLM: Amazon SageMaker (using LLaMA 2). - Process: The text is processed with LLaMA 2 to generate vector embeddings, capturing semantic information for advanced analysis and search functionalities. 5. Vector Storage - Tool: Postgres with a vector extension - Process: Text vectors are stored in the database, allowing for efficient management and retrieval of vectorized data for semantic searches. 6. Front-end Web Application and Search Functionality - Front-end Technology: React.js. - Key Features: - Semantic search input and results display. - email input field for collecting contact information for marketing purposes, forwarding to the client's email address. - Homepage containing descriptive marketing text. - 3 pages total: home page, interaction page, contact page, plus a pop up with GDPR info. Graphics provided as template guidance. - Back-end Technology: Python with FastAPI. Phase 2: Full Automation and Scaling 1. Automated Document Ingestion - Process: A web scraping tool is implemented to automatically identify and upload new documents to the S3 bucket, facilitating a continuous flow of data into the pipeline without manual intervention. 2. Scalable Architecture - Deployment: The application components are containerized using Docker and managed with Kubernetes (Amazon EKS), ensuring the system can scale efficiently to accommodate increased data volumes and user queries. 3. Enhanced Processing Capabilities - Improvements: Integrate additional NLP and ML models for broader and more nuanced text analysis. Consider fine-tuning custom models for specific domain applications. 4. User registration and user management system integration. Please note the attached contract agreement that will be deemed agreed to upon acceptance of the project. Your price given on PPH will be deemed to be your full and final price, and you will be deemed to have fully understood the scope, brief, and specification. To provide context, the project business plan has been uploaded. This is for context only and does not form part of the brief.
opportunity
Python Developer for Sentiment Analysis of Financial Updates
Project Overview: We are looking for a skilled Python developer to create a system that automatically retrieves and analyzes Regulatory News Service (RNS) updates from the Alternative Investment Market (AIM). The key goal is to evaluate these updates for sentiment and linguistic content to assess their potential impact. The system will provide a percentage score reflecting the positivity or negativity of the news and allow for the integration and sentiment weighting of specific keywords. Key Responsibilities: * Develop a Python-based solution to monitor and retrieve AIM RNS updates in real-time. * Utilize natural language processing (NLP) for sentiment and linguistic analysis of RNS data. * Create an algorithm for calculating a sentiment score for each update, represented as a percentage. * Implement functionality for adding and weighting specific keywords for sentiment analysis. * Ensure the system is scalable, reliable, and efficient for processing large data volumes. * Integrate and deploy the system in an AWS environment, ensuring optimal performance and scalability. Required Skills: * Proficient in Python with experience in real-time data processing and web scraping. * Strong background in natural language processing (NLP) and sentiment analysis. * Familiarity with Python libraries like NLTK, Pandas, Requests, BeautifulSoup, Scikit-learn. * Experience with APIs and web-based data retrieval. * Competence in writing clean, efficient, and well-documented code. * Solid experience with AWS (Amazon Web Services) for deployment and scaling of applications. * Proficiency in using Git/GitHub for version control and collaborative development. * Ability to work independently and collaboratively as needed. Desirable Skills: * Knowledge of financial markets, especially AIM RNS updates. * Experience in developing and scaling data analysis systems. * Familiarity with integrating Python applications with various platforms or systems.
Chat GPT Integration
I would like to create an NLP/NLU chatbot for searching and booking flights, which is based on technology like Chat GPT, to be integrated via API, to be trained by our team. To certify your level of knowledge of the technology in question, and your ability to also recommend the best solution to adopt, you are required to answer as many questions as possible contained in the attached file.
pre-funded
I need to create a simple Chatbot style Decision Tree
I need to create a simple Chatbot Decision Tree. A bit like a simple Virtual Assistant. The chatbot will just take the user through a series of scripted questions, and provide them a final decision. It won't require any NLP, the user will be asked fixed questions, and with multiple answers to select. The flow will have branching logic that will determine the next question, dependent on the previous answer given. At the end of the full question flow, the user will need to have a pdf of the final outcome emailed to their inbox. I've provided a super simple image of the flow. Our version will need lots more questions, this is just an example. It'll need to look user friendly and professional. It'll need to be built to allow for further customisation in terms of changing questions or branching logic flow. I need someone who's done this before and is comfortable. Please respond with 'grey squirrel' in your response, so I know you've read this in full and you're not a BOT! :-)
Predictive Model
I want a Predictive Learning Analytics project using machine learning. It should predict student performance based on various input features such as gender, class, attended classes, total classes, and math scores. I prefer using jupyter notebook or google colab. More information to be given. Skills: Machine Learning, NLP, Data science, Data Visualization
Need specific Task for NLP/NMT
i need NLP /NMT developer that develop different task , i will tell you if you're agree because i need task for my education purpose
opportunity
Chatbot to automate quoting
I need a developer in NLP chatbot technology to design and build a chatbot that will sit on a website and automate back-of-rack netting quotes for shelving located within warehouses. It must be able to use simple algebra to take data from questions asked as well as connect to a third-party CRM using API. The chatbot must create a quote for the total cost of materials needed based on a series of questions asked. Attached is the process mapped out.