Connect with us

Latest News

Theta EdgeCloud to Launch with Meta Llama 2, Google Gemma, Stable Diffusion, and Other Popular AI Models | IDOs News

Avatar

Published

on

Theta EdgeCloud to Launch with Meta Llama 2, Google Gemma, Stable Diffusion, and Other Popular AI Models | IDOs News
Theta EdgeCloud to Launch with Meta Llama 2, Google Gemma, Stable Diffusion, and Other Popular AI Models | IDOs News


Theta Labs is launching EdgeCloud, a platform offering open-source AI models, aiming to revolutionize the AI industry by providing developers with access to cutting-edge technologies.

Theta Labs is excited to announce the upcoming launch of Theta EdgeCloud, a groundbreaking platform that will support a variety of open-source AI models. The team is currently in the final stages of development and is working on finalizing the set of AI models that will be available on launch.

With Theta EdgeCloud, AI developers will have the option to select and deploy popular generative AI and large language models. Some of the notable models that will be supported include Meta Llama 2, Google Gemma, Stable Diffusion, and more. These models have gained significant popularity in the industry, with millions of downloads.

One of the key advantages of Theta EdgeCloud is its provision of immediate access to crucial GPU resources. The platform will offer GPU resources such as NVIDIA A100s, V100s, T4s, and other GPUs, along with accompanying dashboards and metrics. This will enable developers to optimize their AI models for various business use cases.

The list of supported AI models on Theta EdgeCloud is subject to change, but currently includes the following:

  1. LLMs: Mistral-8x7B, Mistral-7B, Google Gemma-7B, Meta Llama-2-7B
  2. Image generation: Stable Diffusion XL (SDXL) Turbo, ControlNet
  3. Text-to-video: Stable Video Diffusion
  4. Speech recognition: OpenAI Whisper
  5. Code generation: CodeLlama

Mistral AI, a French AI startup, has developed the Mistral-8x7B model, which is set to rival the latest GPT-4. It is fluent in multiple languages, including English, French, Italian, German, and Spanish.

Google Gemma-7B, launched by Google DeepMind in February 2024, is a family of lightweight open language models optimized for Nvidia GPUs and Google Cloud TPUs. The model is designed for efficient deployment on consumer-size GPUs, as well as cloud-based GPUs.

Meta Llama-2, developed in partnership with Microsoft, is an open-source large language model that is free for both research and commercial use. It offers improved accuracy and safety of chat content through human feedback and reinforcement learning. Theta EdgeCloud will launch with Llama-2 support.

Stable Diffusion, developed by Stability AI, is a groundbreaking text-to-image generative AI technology. It has gained immense popularity since its launch in 2022 and will be supported by Theta EdgeCloud. The platform will also support Stable Video Diffusion, which enables text-to-video generation.

Additionally, Theta EdgeCloud will offer support for ControlNet, an innovative genAI model in the text-to-image category. This model allows users to have greater control over image generation by taking an additional input image.

OpenAI’s Whisper, a state-of-the-art speech recognition model, will also be supported on Theta EdgeCloud. It boasts “semi-supervised learning” and has been trained on a vast amount of multilingual and multitask data, making it more robust and accurate than other models.

Lastly, Theta EdgeCloud will launch with support for Meta’s CodeLlama, a large language model specifically designed for coding. It has the potential to enhance workflows and efficiency for developers, as well as lower the barrier to entry for beginners learning to code.

Theta EdgeCloud aims to revolutionize the AI industry by providing developers with easy access to powerful AI models and GPU resources. With its upcoming launch, it promises to empower AI developers and enable them to create innovative solutions using cutting-edge technologies.

Image source: Shutterstock


Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Latest News

NVIDIA and Mistral Launch NeMo 12B: A High-Performance Language Model on a Single GPU | IDOs News

Avatar

Published

on

NVIDIA and Mistral Launch NeMo 12B: A High-Performance Language Model on a Single GPU | IDOs News
NVIDIA and Mistral Launch NeMo 12B: A High-Performance Language Model on a Single GPU | IDOs News




Iris Coleman
Jul 27, 2024 05:35

NVIDIA and Mistral have developed NeMo 12B, a high-performance language model optimized to run on a single GPU, enhancing text-generation applications.





NVIDIA, in collaboration with Mistral, has unveiled the Mistral NeMo 12B, a groundbreaking language model that promises leading performance across various benchmarks. This advanced model is optimized to run on a single GPU, making it a cost-effective and efficient solution for text-generation applications, according to the NVIDIA Technical Blog.

Mistral NeMo 12B

The Mistral NeMo 12B model is a dense transformer model with 12 billion parameters, trained on a vast multilingual vocabulary of 131,000 words. It excels in a wide range of tasks, including common sense reasoning, coding, math, and multilingual chat. The model’s performance on benchmarks such as HellaSwag, Winograd, and TriviaQA highlights its superior capabilities compared to other models like Gemma 2 9B and Llama 3 8B.







Model Context Window HellaSwag (0-shot) Winograd (0-shot) NaturalQ (5-shot) TriviaQA (5-shot) MMLU (5-shot) OpenBookQA (0-shot) CommonSenseQA (0-shot) TruthfulQA (0-shot) MBPP (pass@1 3-shots)
Mistral NeMo 12B 128k 83.5% 76.8% 31.2% 73.8% 68.0% 60.6% 70.4% 50.3% 61.8%
Gemma 2 9B 8k 80.1% 74.0% 29.8% 71.3% 71.5% 50.8% 60.8% 46.6% 56.0%
Llama 3 8B 8k 80.6% 73.5% 28.2% 61.0% 62.3% 56.4% 66.7% 43.0% 57.2%

Table 1. Mistral NeMo model performance across popular benchmarks

With a 128K context length, Mistral NeMo can process extensive and complex information, resulting in coherent and contextually relevant outputs. The model is trained on Mistral’s proprietary dataset, which includes a significant amount of multilingual and code data, enhancing feature learning and reducing bias.

Optimized Training and Inference

The training of Mistral NeMo is powered by NVIDIA Megatron-LM, a PyTorch-based library that provides GPU-optimized techniques and system-level innovations. This library includes core components such as attention mechanisms, transformer blocks, and distributed checkpointing, facilitating large-scale model training.

For inference, Mistral NeMo leverages TensorRT-LLM engines, which compile the model layers into optimized CUDA kernels. These engines maximize inference performance through techniques like pattern matching and fusion. The model also supports inference in FP8 precision using NVIDIA TensorRT-Model-Optimizer, making it possible to create smaller models with lower memory footprints without sacrificing accuracy.

The ability to run the Mistral NeMo model on a single GPU improves compute efficiency, reduces costs, and enhances security and privacy. This makes it suitable for various commercial applications, including document summarization, classification, multi-turn conversations, language translation, and code generation.

Deployment with NVIDIA NIM

The Mistral NeMo model is available as an NVIDIA NIM inference microservice, designed to streamline the deployment of generative AI models across NVIDIA’s accelerated infrastructure. NIM supports a wide range of generative AI models, offering high-throughput AI inference that scales with demand. Enterprises can benefit from increased token throughput, which directly translates to higher revenue.

Use Cases and Customization

The Mistral NeMo model is particularly effective as a coding copilot, providing AI-powered code suggestions, documentation, unit tests, and error fixes. The model can be fine-tuned with domain-specific data for higher accuracy, and NVIDIA offers tools for aligning the model to specific use cases.

The instruction-tuned variant of Mistral NeMo demonstrates strong performance across several benchmarks and can be customized using NVIDIA NeMo, an end-to-end platform for developing custom generative AI. NeMo supports various fine-tuning techniques such as parameter-efficient fine-tuning (PEFT), supervised fine-tuning (SFT), and reinforcement learning from human feedback (RLHF).

Getting Started

To explore the capabilities of the Mistral NeMo model, visit the Artificial Intelligence solution page. NVIDIA also offers free cloud credits to test the model at scale and build a proof of concept by connecting to the NVIDIA-hosted API endpoint.

Image source: Shutterstock



Continue Reading

Latest News

Binance Introduces FDUSD Flexible Products with Up to 11.5% APR Rewards | IDOs News

Avatar

Published

on

Binance Introduces FDUSD Flexible Products with Up to 11.5% APR Rewards | IDOs News
Binance Introduces FDUSD Flexible Products with Up to 11.5% APR Rewards | IDOs News




James Ding
Jul 26, 2024 08:57

Binance launches a promotion for FDUSD Simple Earn Flexible Products, offering up to 11.5% APR rewards for eligible users during the promotion period.





Binance, a leading cryptocurrency exchange, has launched a new promotional offer for its FDUSD Simple Earn Flexible Products. According to binance.com, eligible users can enjoy an exclusive Bonus Tiered Annual Percentage Rate (APR) of 10.5% on top of Real-Time APR rewards, bringing the total potential APR to 11.5%.

Promotion Details

The promotion period will run from July 26, 2024 (00:00 UTC) to August 8, 2024 (23:59 UTC). During this period, users who subscribe to the FDUSD Simple Earn Flexible Products on a first-come, first-served basis can benefit from the enhanced APR rewards.

Subscription and Rewards

Subscribers will start accruing Bonus Tiered APR rewards the day after their subscription, starting from 00:00 UTC. These rewards will be distributed daily between 00:00 UTC and 08:00 UTC to the user’s Spot Wallet. All users holding open positions in FDUSD Flexible Products will receive both Real-Time APR and Bonus Tiered APR rewards during the promotion period.

Once the promotion ends, only Real-Time APR rewards will continue. The Bonus Tiered APR rewards are calculated based on the subscribed amounts and are subject to specific tier limits. Users can view their Flexible Products assets in their Binance wallets under the ‘Earn’ section.

Terms and Conditions

Binance has outlined several terms and conditions for the promotion:

  • Bonus Tiered APR rewards start accruing the day after subscription.
  • Rewards are distributed daily to the user’s Spot Wallet.
  • Real-Time APR and Bonus Tiered APR rewards are funded by Binance’s own resources and are not affiliated with the issuer of FDUSD.
  • Redemption for Flexible Products subscriptions is instant, though daily redemption limits apply and may be subject to change.
  • Binance reserves the right to disqualify participants who tamper with the system or act fraudulently.
  • Binance may cancel or amend the promotion at its discretion.

Additional Information

Users are advised to refer to the Binance Simple Earn Terms & Conditions before participating. Discrepancies may exist between translated and original versions of the announcement, and users should refer to the original English version for the most accurate information.

Binance’s announcement also includes a general disclaimer about the volatility of digital asset prices and the risks involved in cryptocurrency investments. Users are encouraged to make informed decisions and understand that APR is an estimate and may vary.

For more details, visit the official announcement on binance.com.

Image source: Shutterstock



Continue Reading

Latest News

BitMEX Announces Updates to Basket Indices | IDOs News

Avatar

Published

on

BitMEX Announces Updates to Basket Indices | IDOs News
BitMEX Announces Updates to Basket Indices | IDOs News




Jessie A Ellis
Jul 26, 2024 09:12

BitMEX has updated the weight and index multiplier of its basket indices, impacting several key indices as of July 26, 2024.





BitMEX, a leading cryptocurrency exchange, has announced significant updates to the weight and index multiplier of its basket indices, effective as of July 26, 2024. This update impacts several key indices, including .BMEMEMEXT, according to the BitMEX Blog.

Details of the Updates

The changes took effect at 04:00:05 UTC on the specified date. These adjustments are part of BitMEX’s ongoing efforts to maintain the accuracy and reliability of its trading platform. The specific indices affected by these updates have seen alterations in their weight and index multipliers, which are critical for traders to note as these factors directly influence their trading strategies and outcomes.

Impact on Traders

For traders using these indices, it is essential to review and adjust their positions accordingly. The changes could affect the performance and risk profile of their trading strategies. BitMEX has emphasized that these updates are designed to reflect the current market conditions and improve the overall trading experience on their platform.

Compliance and Restrictions

BitMEX also reiterated its commitment to compliance with regulatory standards. The platform prohibits access to trading or holding positions for individuals or entities located in restricted jurisdictions, including the United States. BitMEX has warned that any user found to have provided false information regarding their location or residency will have their accounts closed and positions liquidated immediately.

The exchange reserves the right to update its Restricted Jurisdiction Policy and implement new controls to prevent access from these areas. Users are advised to stay informed about these policies to avoid any disruptions to their trading activities.

Advice for Traders

BitMEX has cautioned that the material posted on its blog should not be considered as investment advice or recommendations. Traders are encouraged to conduct their due diligence and consult with financial advisors before making investment decisions. The updates to the basket indices are part of BitMEX’s broader strategy to ensure a fair and transparent trading environment.

This announcement underscores the importance of staying updated with platform changes and regulatory requirements to navigate the dynamic landscape of cryptocurrency trading effectively.

Image source: Shutterstock



Continue Reading

Trending