Found 39 repositories(showing 30)
ghosthamlet
Pytorch model for https://github.com/imcaspar/gpt2-ml
iyaja
A companion repository for the GPT-2 article on the FloydHub blog.
CyberZHG
Load GPT-2 checkpoint and generate texts in PyTorch
gzroy
Pytorch implementation for gpt2
BenjaminWegener
text generation using GPT2
ngocthinh09
A from-scratch implementation of GPT-2 built for learning Transformer architectures. Optimized with DDP, Flash Attention, and torch.compile.
CaptainJa
No description available
cyyeh
tiny torch-like engine implemented in Rust with Python bindings via PyO3 using Cursor agent mode(Opus 4.5, GPT 5.2)
ParikshitGehlaut
PyTorch implementation of GPT-2 124M model along with training script
denma98
Reproduced the GPT-2 124M parameter LLM from scratch using PyTorch, referencing ”At tention is All You Need” and GPT papers. Optimized performance with Flash Attention, Torch Compile, Gradient Accumulation, and DDP for multi-GPU training. Evaluated on the Hellaswag dataset.
warrenzha
GPT-2 Torch.
samnet
Implementation of GPT 2.0 using Torch
S4vyss
No description available
JulianSprung
Pytorch implementation of GPT-2
shamashel
Basic language model using torch, based on gpt-2
gkswjdzz
No description available
AnshDhalla1
No description available
AnshDhalla1
No description available
VantaTomat
Minimal GPT-2 inference in pure PyTorch (no transformers, no safetensors)
No description available
Parsagh05
No description available
nafisadipra
No description available
SehbazSingh
This repository demonstrates how to use the GPT-2 tokenizer from OpenAI's `tiktoken` library to tokenize text data, and then apply a simple PyTorch embedding layer on the tokenized input.
gkswjdzz
No description available
No description available
thaddavis
No description available
A minimalist implementation of the GPT-2 architecture built entirely from scratch using PyTorch.
thaddavis
No description available
thaddavis
No description available
KshitijK288
This project implements the GPT-2 (124M) transformer model entirely from scratch using PyTorch, including custom multi-head causal self-attention, LayerNorm, training loop, and text generation. It also supports loading and running OpenAI’s pretrained GPT-2 weights, enabling both training from scratch and pretrained inference.