How Much Energy Does Your Prompt Use? Measuring AI Impact with Ecologits скачать в хорошем качестве

How Much Energy Does Your Prompt Use? Measuring AI Impact with Ecologits

In this episode of Is it Observable?, we dive into the technical side of measuring the environmental impact of AI workloads. As large language models (LLMs) like ChatGPT, Claude, and Gemini become integral to our workflows, it’s time to ask: how much energy do they really consume? 🌍⚡ Join us as we explore: 🔍 The key differences between traditional applications and LLMs. 📊 How to monitor and observe AI workloads with tools like OpenLLMetry and OpenLit. 🌱 Estimating energy usage and emissions using the open-source project Ecologits. 🖥️ Leveraging CodeCarbon for energy tracking in self-hosted LLM environments. 📈 Exporting energy data as logs or OpenTelemetry metrics for deeper analysis. Whether you’re an AI developer, data scientist, or just curious about the environmental impact of AI, this episode provides a step-by-step guide to understanding and reducing energy consumption. 💻 GitHub Repo As always, I’ve prepared a GitHub repository with all the examples and code used in this episode: 👉 https://github.com/isItObservable/eco... 👍 Like the episode? 🔔 Subscribe for more deep dives into observability, Kubernetes, and sustainable AI. 💬 Drop a comment if you’re using any of these tools or have others to recommend! ▬▬▬▬▬▬ 🔗 Additional Info 🔗 ▬▬▬▬▬▬ 🔗Dynatrace Trial: https://bit.ly/3KxWDvY 🔗Github repo for the tutorial: https://github.com/isItObservable/eco... 🔗ecolotigits : https://ecologits.ai/latest 🔗ecologits algorithm: https://ecologits.ai/latest/methodolo... 🔗Pr on the Otel support: https://github.com/genai-impact/ecolo... 🔗OpenLLmetry: https://www.traceloop.com/docs/openll... 🔗OpenLit : https://docs.openlit.io/latest/introd... 🔗CodeCarbon: https://codecarbon.io 🔗Blog : https://isitobservable.io ▬▬▬▬▬▬ 👋 Contact me 👋 ▬▬▬▬▬▬ ➡ Twitter:   / isitobservable   ➡ LinkedIn:   / isitobservable   ▬▬▬▬▬▬ ⏱ Timecodes ⏱ ▬▬▬▬▬▬ 00:00 Intro 02:20 Differences between traditional apps and LLMs 07:45 LLM Observability 13:11 Ecologits 19:48 Conclusion