Recently, I’ve been generating many images to serve as illustrations in my Spanish vocabulary courses published on Flashcard Space. Last week I switched from Stable Diffusion XL to Flux.1 dev. It is a significant upgrade in image quality, but I also noticed the illustrations have less variety than SDXL, at least when I use the … Read more
Generative AI endpoints might return Base64-encoded images Recently, I explored the possibility of generating images using locally hosted Stable Diffusion models. Thanks to the AUTOMATIC1111 stable-diffusion-WebUI project, I got the local API working pretty quickly. However, one aspect of working with such APIs is that images come base64-encoded in a JSON property, and one can’t … Read more
What is JSON Mode in Google Gemini? I recently blogged about JSON Mode and Structured Outputs in OpenAI APIs. Today, I’ve examined the OpenAI competitor Google Gemini and tested whether it can adhere to the requested schema. The JSON Mode is a feature that helps enforce the output of Chatbots to be in JSON format … Read more
What problem does the Structured Outputs feature solve When using GPT models from the code, we developers don’t normally want the output to be in a natural language like in ChatGPT conversations. Instead, we want it to have some structure, and the lingua franca today for data interchange is JSON format. Until recently, reliably achieving … Read more
TL/DR: JSON mode remains available as an API feature, but the newest models include the Structured Outputs feature, which solves this problem 100% without workarounds needed. JSON Mode vs Structured Output When we integrate our applications with chatbot-like API, we often prefer to work with data in a structured format like JSON rather than natural … Read more
Discovering Llama: where does it stand compared with gpt-4o? Today, I was looking for an AI model that would perform best with a non-mainstream task: translating text between two Slavic languages. This was a good opportunity to use the LMSYS Chatbot Arena. The service allows you to perform a blind test of two random modern … Read more
What is OpenAI Batch API? The pricing page for OpenAI services has an annotation that drew my attention recently: *Batch API pricing requires requests to be submitted as a batch. Responses will be returned within 24 hours for a 50% discount. I pretended not to see this option for a while because using batch jobs … Read more
Here’s a short hint that helped me resolve an annoying issue of ChatGPT signing me out very often. It’s a service I have bookmarked and I use frequently, yet it seemed to be more stubborn than most apps and I was signed out every day I opened it 😉 And especially if you use two-factor … Read more
Observing default behavior: JSON is indented, and we pay for it While working with ChatGPT API, I noticed that the JSON response I normally receive contains many whitespace characters used for indentation. Here’s a fragment of the response as an illustration: Output tokens still cost quite a lot in 2024 (assuming we process lots of … Read more
Does OpenAI Text to Speech support languages other than English? OpenAI’s Text-to-Speech service transforms text into audio files with impressive quality when we consider English language. I haven’t performed a blind test, but based on a few samples I heard, I don’t think I could easily recognize if the audio were recorded by a real … Read more