Deci Launches DeciCoder to Augment Code Development with Generative AI

Generative AI startups are really focusing on the future of code.

Friends,

Here we go again! Yet another Startup unveiling a tool to ease code development. I’m not going to lie, I find these products fascinating. I’m of the belief that Generative A.I. has a lasting impact on the future of software development, no-code and how we use code and accessibility to coding at large.

So what is Deci? Deci’s deep learning development platform bridges the AI efficiency gap, empowering AI teams to efficiently build next generation deep learning applications. 

To get more articles from us:

If you want to learn more about Deci there’s a Webinar by someone I know from Linkedin soon.

So back in July, Deci, a startup company with 50 employees who are developing a platform to build and optimize AI-powered systems, announced that it closed a $25 million Series B financing round led by Insight Partners with participation from Square Peg, Emerge, Jibe Ventures, Fort Ross Ventures and ICON that brings the company’s total raised to $55.1 million.

Their mission is to empower AI developers with powerful tools for building innovative AI based solutions.

Arrival of DeciCoder

Deci’s DeciCoder is a novel generative AI foundation model designed to streamline code development.

Future of Code? (the pitch by Saboo)

🔥 Introducing DeciCoder - an extraordinary open-source LLM for code generation. It's not only fast and precise, but it's also budget-friendly.

🔍 What is DeciCoder?
It's an auto-regressive language model rooted in the transformer decoder architecture. Tailored for Python, Java, and Javascript enthusiasts, its performance is outstanding with superior memory utilization even on budget GPUs. We've struck the perfect balance between rapid code generation and minimal latency.

🧠 Deci's AutoNAC Revolution
Remember the days when designing the "optimal" neural network structure was a backbreaking task? Say goodbye to manual customizations and exhaustive trials! Deci’s AutoNAC has made strides in uniting accuracy with speedy inference.

DeciCoder was generated using Deci's proprietary Automated Neural Architecture Construction (AutoNAC) engine, the most advanced Neural Architecture Search (NAS)-based technology on the market. AutoNAC identifies the ideal architecture that strikes a perfect balance between accuracy and processing speed, tailored for distinct data features, tasks, performance goals, and inference environment.

From models like Yolo-NAS (for object detection) to DeciBERT (Q&A) and DeciSeg (semantic segmentation), AutoNAC has been the brain behind them. For DeciCoder, it has adeptly maneuvered transformers to craft an architecture fine-tuned for NVIDIA’s A10 cloud GPU. The precision? It's comparable (perhaps superior) to the likes of SantaCoder! 🎅

📊 Key Details:
- 1 Billion parameters
-  Trained on a whopping 6TB of source code across 358 languages, with a primary focus on Python, Java, and Javascript.
- A generous context window of 2048 tokens.

What are you waiting for? 🤔 Try it now for free:
🕹️ Demo: https://lnkd.in/gzxAapuJ
🌎 Access the Model: https://lnkd.in/g-EeMqWs
📜 Detailed Insights: https://lnkd.in/gnMVfh9Q

In Brief

• With a robust architecture featuring 1 billion parameters and a 2048-token context window, DeciCoder can generate diverse and high-quality code snippets across multiple programming languages.

• DeciCoder’s efficiency outshines SantaCoder, offering faster inference speeds on more affordable hardware without compromising accuracy.

• It outperforms SantaCoder in accuracy across Python, JavaScript, and Java languages.

https://huggingface.co/Deci/DeciCoder-1b

Some of the specs are fairly encouraging for Deci.

When DeciCoder was benchmarked on Hugging Face Inference Endpoints against well-established code LLMs such as SantaCoder, DeciCoder showcased a 22% increase in throughput, a significant reduction in memory usage, and a 1.5-2.4 percentage point improvement in accuracy on the HumanEval benchmark.

I have no affiliations with this company and this is not a sponsored post, I really do find Generative A.I. startups going after coding fairly interesting.

The product was first unveiled on August 15th, 2023.

Apparently, when running on NVIDIA’s A10G, a less expensive hardware, DeciCoder’s inference speed surpasses that of SantaCoder, the most popular model in the 1-billion parameter range, running on the pricier NVIDIA’s A100. Moreover, DeciCoder on the A10 is 3.5 times faster than SantaCoder on the A10 and 1.6 times faster than SantaCoder on the A100. I obviously could not verify these claims of the company.

About Deci

Deci enables deep learning to live up to its true potential by using AI to build better AI. With the company’s deep learning development platform, AI developers can build, optimize, and deploy faster and more accurate models for any environment including cloud, edge, and mobile, allowing them to revolutionize industries with innovative products.

A NAC/NAS Startup

The platform is powered by Deci’s proprietary automated Neural Architecture Construction technology (AutoNAC), which automatically generates and optimizes deep learning models’ architecture and allows teams to accelerate inference performance, enable new use cases on limited hardware, shorten development cycles and reduce computing costs. Founded in 2019, Deci’s team of deep learning engineers and scientists are dedicated to eliminating production-related bottlenecks across the AI lifecycle.

NAS, (Neural Architecture Search) a family of techniques on which Deci heavily relies, can help automatically discover low-cost, optimal models for a given problem. Deci isn’t unique in this — Google’s Vertex AI service leverages NAS to optimize the performance of models on specific, customer-specified tasks. But Geifman argues that Deci’s platform offers access to NAS capabilities at a lower cost.

The company is based in Israel. Deci enables deep learning to live up to its true potential by using AI to build better AI. With the company’s end-to-end deep learning acceleration platform, AI developers can build, optimize, and deploy faster and more accurate models for any environment, including cloud, edge, or mobile.