Made by AI with Compute
- CLIP – Contrastive Language–Image Pretraining (2021)
- ViT – An Image is Worth 16x16 Words (2020)
- AlexNet – ImageNet Classification with Deep Convolutional Neural Networks (2012)
- LLaMA 2 – LLaMA 2: Open Foundation and Fine-Tuned Chat Models (2023)
- Mistral MoE – Mixtral of Experts (2023)
- GPT-2 – Language Models are Unsupervised Multitask Learners (2019)
- Word2Vec – Word2Vec (2014)
- Gemma – Gemma: A Generalized Multi-Modal Foundation Model (2024)
- VAE (MLP) – Variational Autoencoder using fully-connected networks
- VAE (U-Net CNN) – Convolutional VAE with U-Net-style encoder/decoder
- PPO – Proximal Policy Optimization (2017)
- DQN – Deep Q Learning (2013)
- DPO - Direct Preference Optimization for LLM-Post Training
- (Work in progress)
- LoRA – LoRA: Low-Rank Adaptation for Fast Fine-Tuning (2023)
- DPO - Direct Preference Optimization for Post-training
- Hinton Distillation - Learning from Parent Model