I am calling gemini flash 2.5 image generator using API key from an application. I have a paid account with tier 1. However, the generation keeps on failing. I am trying to generate 10-11 images in a row using a prompt but it always errors out after 2-9 images. Can someone help me with this?
1 Like
Hello,
Welcome to the Forum!
Could you please share the error you are getting along with the full payload and prompt, so that we can reproduce your issue and come up with more accurate response?
1 Like
Prompt: “Generate 11 images in total:
One main page with the full scene containing all 10 hidden objects. also provide the coordinates and dimensions of those images as a separate appropriate data structure for programming.
Ten separate images, transparent background, do not include any test or border or background scenery, each one isolating and clearly showing one of the hidden objects (Item 1 through Item 10).
Keep objects fun, recognizable, and age-appropriate.”
Error: “The model response was blocked, please clear your chat or start a new prompt to continue.”
I dont undestand which payload do you want? I get this error when I use the aistudio.google.com. after 3 to 9 images, no set number. From my react app, using google API key, the results are worse.
AI Studio is primarily designed for experimentation and may experience performance issues during peak usage times. For a more seamless and reliable experience, we recommend opting for the paid tier.
I have a paid tier. Still the same issue.
Hello,
Apologies for the confusion. We tried to reproduce your issue and test with different prompts. We have noticed that we get stuck after generating a total of 10 images per prompt. Could you please check if you observe a similar behavior? Additionally, if you have any extra information or observations please share them it will help us analyze this issue more effectively.
I see the errors after 3,4,5..10 images randomly. There is no predictive pattern of failure. Also, there are no images. I tried to break down the generation by doing 2 images at a time and even that fails randomly. The consistency is also a problem
When can we expect this issue to be resolved?