Batch Size Parameters
The number of images generated simultaneously in a single operation. Larger batches are more efficient but require more VRAM.
Batch size determines how many images are generated in parallel.
How Batching Works
- Batch of 1: Generate one image at a time
- Batch of 4: Generate four images together
- Batch count: Repeat the batch multiple times
VRAM Requirements
Each additional image requires more memory:
| Batch | VRAM Increase |
|---|---|
| 1 | Baseline |
| 2 | ~1.8x |
| 4 | ~3.2x |
| 8 | ~6x |
Batch Size vs Batch Count
- Batch Size: Parallel generation, uses more VRAM
- Batch Count: Sequential batches, same VRAM
Example: 8 images
- Batch size 8, count 1: Fast, high VRAM
- Batch size 2, count 4: Slower, lower VRAM
- Batch size 1, count 8: Slowest, minimum VRAM
When to Use Larger Batches
- Exploring prompt variations
- Random seed exploration
- Faster total generation
- When you have VRAM headroom
Tips
- Start with batch size 1-2
- Increase if VRAM allows
- Use batch count for exploration
- Each image in batch gets different seed (by default)
- Same seed across batch = same image