Ingredient Photo AI Training
Budget: ₹600 – ₹1,500 INR
I run several cooking blogs and want to automate the creation of clean, consistent ingredient photos. Your task is to train (or fine-tune) an image-generation model so that it reliably produces minimalistic ingredient shots laid out on a wooden surface.
Here’s what I need from you:
• Help me assemble, clean, and label a high-quality training set of ingredient images that match the desired look.
• Select or build the best model architecture (Stable Diffusion, DreamBooth, or a comparable approach)—I’m open, as long as it learns quickly and can run locally after training.
• Train and validate the model, then deliver the final weights, a straightforward inference script, and a short guide so I can generate new photos on demand (Python preferred).
• Provide at least 20 sample outputs that prove the model can create various pantry staples, herbs, and produce while maintaining the same minimalistic style and wood background.
Acceptance criteria
1. Generated images show individual ingredients clearly, with no extra props, shadows kept natural, and the wood grain visible but not distracting.
2. Resolution of 1024×1024 px or higher.
3. Reproducible environment: requirements.txt or environment.yml plus a quick-start README.
If this sounds like your wheelhouse, let’s work together to streamline my visual content workflow.
Here’s what I need from you:
• Help me assemble, clean, and label a high-quality training set of ingredient images that match the desired look.
• Select or build the best model architecture (Stable Diffusion, DreamBooth, or a comparable approach)—I’m open, as long as it learns quickly and can run locally after training.
• Train and validate the model, then deliver the final weights, a straightforward inference script, and a short guide so I can generate new photos on demand (Python preferred).
• Provide at least 20 sample outputs that prove the model can create various pantry staples, herbs, and produce while maintaining the same minimalistic style and wood background.
Acceptance criteria
1. Generated images show individual ingredients clearly, with no extra props, shadows kept natural, and the wood grain visible but not distracting.
2. Resolution of 1024×1024 px or higher.
3. Reproducible environment: requirements.txt or environment.yml plus a quick-start README.
If this sounds like your wheelhouse, let’s work together to streamline my visual content workflow.