derrianx / photo-journey

documentation of film photography resources
0 stars 0 forks source link

photo-journey

documentation of film photography resources

Credits to chatgpt for generating this idea A tool for analyzing the exposure settings of your film photographs. This project could involve building a program that extracts metadata from your film photographs (e.g. ISO, aperture, shutter speed) and generates graphs or statistics to help you understand your shooting habits and identify any patterns or trends.

Initial problem

read.py is not able to read all the files

Resolution

I went on to download to exiftool and managed to output the entire directory onto a csv file

That is how i get my data set to begin analysis documentation: https://exiftool.org/faq.html#Q12 exiftool -csv -r image > out.csv 1 directories scanned 95 image files read

resources

pandas

https://datatofish.com/convert-string-to-float-dataframe/ https://www.w3schools.com/python/pandas/pandas_plotting.asp

exif

https://photo.stackexchange.com/ https://buildmedia.readthedocs.org/media/pdf/exif/latest/exif.pdf

csv file

https://exiftool.org/faq.html#Q12

new idea with dalle

The official DALL-E GitHub repository (https://github.com/lucidrains/DALLE-pytorch) provides a detailed tutorial on how to train DALL-E on a custom dataset. It includes instructions for preparing the dataset, training the model, and generating new images.

The OpenAI website (https://openai.com/dall-e/) also provides a tutorial on how to train DALL-E on a custom dataset, including instructions on how to use the OpenAI API to generate images from text prompts.

There is a tutorial in the Hugging Face website (https://huggingface.co/blog/dall-e-2-image-generation-tutorial) which provides a step-by-step guide on how to train DALL-E on a custom dataset and use it to generate new images.

It is important to note that the training of DALL-E is a computationally intensive task, it may require a powerful GPU and a lot of memory.