#AzureAIStudio

[govindhtech]

Mistral Large 2: Setting New Standards In Code Generation

Mistral is pleased to present the next iteration of their flagship model, Mistral Large 2, today. Mistral Large 2 is far more proficient in mathematics, logic, and code production than its predecessor. It also offers sophisticated function calling capabilities and far better linguistic support.

The most recent generation is still pushing the limits of performance, speed, and cost effectiveness. Mistral Large 2 is made available on la Platform and has been enhanced with additional functionalities to make the development of creative AI apps easier.

Mistral Large 2

With a 128k context window, Mistral Large 2 is compatible with more than 80 coding languages, including Python, Java, C, C++, JavaScript, and Bash, and it supports dozens of languages, including Arabic, Hindi, French, German, Spanish, Italian, Portuguese, and Chinese.

Mistral Large 2’s size of 123 billion parameters allows it to run at high throughput on a single node; it is intended for single-node inference with long-context applications in mind. Mistral is making Mistral Large 2 available for use and modification for non-commercial and research purposes under the terms of the Mistral Research License. A Mistral Commercial License must be obtained by getting in touch with them in order to use Mistral Large 2 for commercial purposes that call for self-deployment.

General performance

In terms of performance / cost of serving on assessment parameters, Mistral Large 2 establishes new benchmarks. Specifically, on MMLU, the pretrained version attains an accuracy of 84.0% and establishes a new benchmark on the open models’ performance/cost Pareto front.

Code and Reasoning

After using Codestral 22B and Codestral Mamba, Mistral trained a significant amount of code on Mistral Large 2. Mistral Large 2 performs on par with top models like GPT-4o, Claude 3 Opus, and Llama 3 405B, and it significantly outperforms the preceding Mistral Large.

Also, a lot of work went into improving the model’s capacity for reasoning. Reducing the model’s propensity to “hallucinate” or produce information that sounds reasonable but is factually inaccurate or irrelevant was one of the main goals of training. This was accomplished by fine-tuning the model to respond with greater caution and discernment, resulting in outputs that are dependable and accurate.

The new Mistral Large 2 is also programmed to recognise situations in which it is unable to solve problems or lacks the knowledge necessary to give a definite response. This dedication to precision is seen in the better model performance on well-known mathematical benchmarks, showcasing its increased logic and problem-solving abilities:Image credit to Mistral Performance accuracy on code generation benchmarks (all models were benchmarked through the same evaluation pipeline) Image credit to Mistral Performance accuracy on MultiPL-E (all models were benchmarked through the same evaluation pipeline, except for the “paper” row)

Direction after & Alignment

Mistral Large 2’s ability to follow instructions and carry on a conversation was significantly enhanced. The new Mistral Large 2 excels at conducting lengthy multi-turn talks and paying close attention to directions.

Longer responses typically result in higher results on various standards. Conciseness is crucial in many business applications, though, as brief model development leads to faster interactions and more economical inference. This is the reason Mistral worked so hard to make sure that, if feasible, generations stay brief and direct.

Varieties in Language

Working with multilingual documents is a significant portion of today’s corporate use cases. A significant amount of multilingual data was used to train the new Mistral Large 2, despite the fact that most models are English-centric. It performs exceptionally well in Hindi, Arabic, Dutch, Russian, Chinese, Japanese, Korean, English, French, German, Spanish, Italian, Portuguese, and Dutch. The performance results of Mistral Large 2 on the multilingual MMLU benchmark are shown here, along with comparisons to Cohere’s Command R+ and the previous Mistral Large, Llama 3.1 models.Image credit to MistralImage credit to Mistral

Use of Tools and Function Calling

Mistral Large 2 can power complicated commercial applications since it has been trained to handle both sequential and parallel function calls with ease. It also has improved function calling and retrieval skills.

Check out Mistral Large 2 on the Platform

Today, you can test Mistral Large 2 on le Chat and utilise it via la Plateforme under the name mistral-large-2407. Mistral is using a YY.MM versioning scheme for all of their models, therefore version 24.07 is available, and the API name is mistral-large-2407. HuggingFace hosts and makes available weights for the teach model.

Two general-purpose models, Mistral Nemo and Mistral Large, and two specialised models, Codestral and Embed, are the focal points of Mistral’s consolidation of the offerings on la Plateforme. All Apache models (Mistral 7B, Mixtral 8x7B and 8x22B, Codestral Mamba, Mathstral) are still available for deployment and fine-tuning using Mistral SDK mistral-inference and mistral-finetune, even as they gradually phase out older models on la Plateforme.

Mistral is expanding the fine-tuning options on la Plateforme with effect from today on: Mistral Large, Mistral Nemo, and Codestral are now covered.

Use cloud service providers to access Mistral models

Mistral is excited to collaborate with top cloud service providers to introduce the new Mistral Large 2 to a worldwide customer base. Specifically, today they are growing the collaboration with Google Cloud Platform to enable the models from Mistral AI to be accessed on Vertex AI using a Managed API. Right now, Vertex AI, Azure AI Studio, Amazon Bedrock, and IBM Watsonx.ai are all offering the best models from Mistral AI.

Timeline for Mistral AI models’ availability

Read more on govindhtech.com

[govindhtech]

Utilize Azure AI Studio To Create Your Own Copilot

Microsoft Azure AI Studio

With Microsoft Azure AI Studio now broadly available, organisations may now construct their own AI copilots in the fast evolving field of AI technology. Organisations can design and create their own copilot using AI Studio to suit their specific requirements.

AI Studio speeds up the generative AI development process for all use cases, enabling businesses to leverage AI to create and influence the future.

An essential part of Microsoft’s copilot platform is Azure AI Studio. With Azure-grade security, privacy, and compliance, it is a pro-code platform that allows generative AI applications to be fully customised and configured. Utilising Azure AI services and tools, copilot creation is streamlined and accelerated with full control over infrastructure thanks to flexible and integrated visual and code-first tooling and pre-built quick-start templates.

With its simple setup, management, and API support, it eases the idea-to-production process and assists developers in addressing safety and quality concerns. The platform contains well-known Azure Machine Learning technology, such as prompt flow for guided experiences for speedy prototyping, and Azure AI services, such as Azure OpenAI Service and Azure AI Search. It is compatible with code-first SDKs and CLIs, and when demand increases, it can be scaled with the help of the AI Toolkit for Visual Studio Code and the Azure Developer (AZD) CLI.

AI Studios

Model Selection and API

Find the most appropriate AI models and services for your use case.

Developers can create intelligent multimodal, multilingual copilots with customisable models and APIs that include language, voice, content safety, and more, regardless of the use case.

More than 1600 models from vendors such as Meta, Mistral, Microsoft, and OpenAI are available with the model catalogue. These models include GPT 4 Turbo with Vision, Microsoft’s short language model (SLM) Phi3, and new models from Core42 and Nixtla. Soon to be released are models from NTT DATA, Bria AI, Gretel, Cohere Rerank, AI21, and Stability AI. The most popular models that have been packed and optimised for use on the Azure AI platform are those that Azure AI has curated. In addition, the Hugging Face collection offers a wide range of hundreds of models, enabling users to select the precise model that best suits their needs. And there are a tonne more options available!

With the model benchmark dashboard in Azure AI Studio, developers can assess how well different models perform on different industry-standard datasets and determine which ones work best. Using measures like accuracy, coherence, fluency, and GPT similarity, benchmarks evaluate models. Users are able to compare models side by side by seeing benchmark results in list and dashboard graph forms.

Models as a Platform (MaaP) and Models as a Service (MaaS) are the two model deployment options provided by the model catalogue. Whereas MaaP offers models deployed on dedicated virtual machines (VMs) and paid as VMs per-hour, MaaS offers pay-as-you-go per-token pricing.

Before integrating open models into the Azure AI collection, Azure AI Studio additionally checks them for security flaws and vulnerabilities. This ensures that model cards have validations, allowing developers to confidently deploy models.

Create a copilot to expedite the operations of call centers

With the help of AI Studio, Vodafone was able to update their customer care chatbot TOBi and create SuperAgent, a new copilot with a conversational AI search interface that would assist human agents in handling intricate customer queries.

In order to assist consumers, TOBi responds to frequently asked queries about account status and basic technical troubleshooting. Call centre transcripts are summarised by SuperAgent, which reduces long calls into succinct summaries that are kept in the customer relationship management system (CRM). This speeds up response times and raises customer satisfaction by enabling agents to rapidly identify new problems and determine the cause of a client’s previous call. All calls are automatically transcribed and summarised by Microsoft Azure OpenAI Service in Azure AI Studio, giving agents relevant and useful information.

When combined, Vodafone’s call centre is managing about 45 million customer calls monthly, fully resolving 70% of them. The results are outstanding. Customer call times have decreased by at least one minute on average, saving both customers’ and agents’ crucial time.

Create a copilot to enhance client interactions

With the help of AI Studio, H&R Block created AI Tax Assist, “a generative AI experience that streamlines online tax filing by enabling clients to ask questions during the workflow.”

In addition to assisting people with tax preparation and filing, AI Tax Assist may also provide tax theory clarification and guidance when necessary. To assist consumers in maximising their possible refunds and lowering their tax obligations, it might offer information on tax forms, deductions, and credits. Additionally, AI Tax Assist responds dynamically to consumer inquiries and provides answers to free-form tax-related queries.

Construct a copilot to increase worker output

Leading European architecture and engineering firm Sweco realised that employees needed a customised copilot solution to support them in their work flow. They used AI Studio to create SwecoGPT, their own copilot that offers advanced search, language translation, and automates document generation and analysis.

The “one-click deployment of the models in Azure AI Studio and that it makes Microsoft Azure AI offerings transparent and available to the user,” according to Shah Muhammad, Head of AI Innovation at Sweco, is greatly appreciated. Since SwecoGPT was implemented, almost 50% of the company’s staff members have reported greater productivity, which frees up more time for them to concentrate on their creative work and customer service.

Read more on Govindhtech.com