hunar4321/reweight-gpt
Reweight GPT - a simple neural network using transformer architecture for next character prediction
This project offers an alternative to the standard self-attention mechanism found in Transformer neural networks. It takes an input sequence of data, processes it using a direct re-weighting method, and produces predictions for the next element in the sequence. This is primarily useful for machine learning researchers and practitioners exploring new neural network architectures for sequence prediction tasks.
No commits in the last 6 months.
Use this if you are a machine learning researcher or engineer interested in experimenting with novel Transformer architectures for text or sequence generation.
Not ideal if you are looking for an out-of-the-box solution to train a language model or if you prefer standard, widely adopted Transformer implementations.
Stars
56
Forks
8
Language
Jupyter Notebook
License
MIT
Category
Last pushed
Aug 28, 2023
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/llm-tools/hunar4321/reweight-gpt"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
Nixtla/nixtla
TimeGPT-1: production ready pre-trained Time Series Foundation Model for forecasting and...
andrewdalpino/NoPE-GPT
A GPT-style small language model (SLM) with no positional embeddings (NoPE).
sigdelsanjog/gptmed
pip install gptmed
akanyaani/gpt-2-tensorflow2.0
OpenAI GPT2 pre-training and sequence prediction implementation in Tensorflow 2.0
samkamau81/FinGPT_
FinGPT is an AI language model designed to understand and generate financial content. Built upon...