Implementation of GPT from scratch. Design to be lightweight and easy to modify.
[IJCV 2024] P3Former: Position-Guided Point Cloud Panoptic Segmentation Transformer
Complete code for the proposed CNN-Transformer model for natural language understanding.
Symbolic music generation taking inspiration from NLP and human composition process
This notebook shows a basic implementation of a transformer (decoder) architecture for image generation in TensorFlow 2.
A from-scratch implementation of GPT-2 built for learning Transformer architectures. Optimized with DDP, Flash Attention, and torch.compile.
Magic The GPT - GPT inspired model to generate Magic the Gathering cards
GPT (Decoder only Transformer - from scratch) generated fake/phoney taxonomies (based on NCBI taxonomy dataset)
Official Pytorch Implementation of: "Enhancing High-Vocabulary Image Annotation with a Novel Attention-Based Pooling"
⚡ Implement an exact GPT training and inference algorithm in pure JavaScript for browser and Node.js environments without dependencies
实现了两种序列建模（时序分析）解决方案：FMLSTMAttentionModel、FMLlamaModel。 FMLSTMAttentionModel：基于 DeepFM、LSTM 和多头自注意力（Multi-Head Self-Attention）模型的分类及回归任务解决方案。 FMLlamaModel：基于 DeepFM、Llama 自回归大模型架构模型的分类及回归任务解决方案。
In this we explore detailed architecture of Transformer
The goal of this project was to implement the encoder only transformer in order to recreate a mini version of GPT.
Minimal encoder for text classification, decoder for text generation, ViT for image classification
Fully vectorized Transformer decoder implemented from scratch in NumPy with causal masking, autoregressive training, and empirical O(n²) complexity analysis.
Add a description, image, and links to the transformer-decoder topic page so that developers can more easily learn about it.
To associate your repository with the transformer-decoder topic, visit your repo's landing page and select "manage topics."