I find the "prompts to file" feature very useful but would be happy about these improvements:
add an index column. When using Combinatorial generation with several iterations, it would be helpful to have an index in the CSV, representing the nth iteration of the very image that was created. I myself use [generation_number] in the Images filename pattern. So when comparing good with bad results, it's easier to identify the row that matches the generated image.
for each variant add a separate column that represents its specific value that has been used within that specific prompt.
When using one random sampler and two combinatorial samplers within the positive prompt, the result might look like this:
index
positive_prompt
negative_prompt
variant_1
variant_2
variant_3
0
a {~hot|cold|mild} {summer|spring} is coming to {mexico|germany}
I find the "prompts to file" feature very useful but would be happy about these improvements:
index
column. When usingCombinatorial generation
with several iterations, it would be helpful to have an index in the CSV, representing the nth iteration of the very image that was created. I myself use[generation_number]
in theImages filename pattern
. So when comparing good with bad results, it's easier to identify the row that matches the generated image.variant
add a separate column that represents its specific value that has been used within that specific prompt.When using one random sampler and two combinatorial samplers within the positive prompt, the result might look like this: