Ever tried solving optimization problems with PyGAD while fueled by a banana-macha-chocolate smoothie?
Watch this to see how last semester vibes and genetic algorithms actually mix.
Good luck, Josué with your upcoming exam.
https://t.co/ytUQITyIam
LLM Fine-Tuning with Evolutionary Algorithms
Fine-tuning large language models using evolutionary algorithms is especially exciting because it doesn’t require gradient descent—only inference. You modify the network, evaluate how it performs, and iterate.
This approach avoids getting stuck in local minima and sidesteps many of the constraints associated with gradient-based optimization. Instead of relying on backpropagation, you explore the solution space more freely.
Several techniques can make inference-time models far more efficient than their original versions: distillation, quantization, 1-bit networks, specialized hardware, and more. When you take these advantages into account, evolutionary fine-tuning can potentially become significantly more efficient than traditional gradient descent–based methods.
https://t.co/lLBWEXXQDb
Alex has been a loyal contributor to the PyGAD library. Huge thanks to Alex for being a loyal supporter and helping us keep the mutation rates stable and the crossovers clean!
Open Collective: https://t.co/CNr8fbAKJY
PayPal: Use either this link: https://t.co/kFAuTjGOS4
PyGAD 3.6.0 Released
Quick summary of the release changes:
1. Pass a class as fitness function.
2. Optimizing and refactoring the code.
3. More tests.
4. Other bug fixes.
Documentation: https://t.co/TO8a7zEesX
GitHub repository: https://t.co/02yBdh49p1
A new AI review! ahmedfgad/GeneticAlgorithmPython ⭐3.7/5.0
PyGAD (published as the `pygad` package) is a mature, feature-rich genetic algorithm library for Python focused on practical optimization workflows, including integ...
https://t.co/tbulBzwgSR
@PyGADLib 3.5.0 Released!
The 3.5.0 release introduces the new gene_constraint parameter enabling users to define custom rules for gene values.
Source code: https://t.co/L0ryFlpUxq
Documentation: https://t.co/haG9uSBdfH
@PyGADLib 3.5.0 Released!
The 3.5.0 release introduces the new gene_constraint parameter enabling users to define custom rules for gene values using callables.
Source code at GitHub: https://t.co/L0ryFlpUxq
Documentation: https://t.co/haG9uSBdfH
Thank you Hasan Can Beydili for the comparison between @PyGADLib and PyerualJetwork.
Unfortunately, I have no Turkish background to benefit from the comparison.
I just used Google Gemini to create a quick summary of the comparison.
https://t.co/IflU2MZwfr
@PyGADLib 3.4.0 Released!
PyGAD is a Python library for solving general-purpose optimization problems using the genetic algorithm.
GitHub repository: https://t.co/L0ryFlpmHS
Documentation: https://t.co/pCyHteNEmi
🚀🎉 We Did It! 2 Million Downloads! 🎉🚀
A huge THANK YOU to our incredible community for making this milestone possible! PyGAD has officially reached 2 MILLION downloads! 🥳✨
https://t.co/CseRlSJ1uT
https://t.co/02yBdh49p1
Here’s to the next million! 🚀🔥
🚀 The Optimization Gadget is Almost Here!
Stay ahead of the curve—subscribe to our newsletter for exclusive updates and early access! 🌟
https://t.co/MWYiiGGYIZ
combining compression around extreme values (with jax,numpy) with genetic algorithms (pygad) to optimize distribution of compression thresholds in #python
NSGA-ii is so good (Non-Dominated Sorting Genetic Algorithm II), like implemented in pygad #pythonprogramming. It easily handles more than 10 optimization objectives at the same time. Here used to find a nice function to plot and to optimize the color gradients.