Featured projects
Griffin
Mixing Gated Linear Recurrences with
Local Attention for Efficient Language Models.
Paper implementation
RNN
Attention
Hybrid
Confidence Transformer
Transformer model that reports confidence in its predictions.
Transformer
Confidence Estimation
F5 TTS
Fast and efficient Text-to-Speech model.
TTS
Speech Synthesis
LLM Memory Modules
Exploring memory mechanisms for Large Language Models at scale.
LLM
Memory
Scaling
Simple ViT
A straightforward implementation of the Vision Transformer.
Vision Transformer
Computer Vision
Simple CLIP
A straightforward implementation of CLIP.
CLIP
Multimodal
Tiny MoE
A small Mixture of Experts language model.
MoE
LLM
Conversational Transformer
Transformer model focused on conversational AI.
Transformer
Chatbot
Action Transformer
Transformer model for predicting sequences of actions.
Transformer
Reinforcement Learning
YOLOv5 DeepQ Learning Model
Combining YOLOv5 object detection with Deep Q-Learning.
Object Detection
Reinforcement Learning
Single Architecture Ensemble
Multi Input-Multi Output Transformer models with early exit
Paper implementation
Ensemble Methods
Deep Learning
T5 SpeechtoSpeech
Speech-to-speech translation model based on T5.
T5
Speech Translation
