gptopus
🐙GPTopus is a decoder-only transformer that generates new text based on input text
Science Score: 44.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
✓CITATION.cff file
Found CITATION.cff file -
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
○DOI references
-
○Academic publication links
-
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (6.9%) to scientific vocabulary
Repository
🐙GPTopus is a decoder-only transformer that generates new text based on input text
Basic Info
- Host: GitHub
- Owner: dantethemartian
- License: mit
- Language: Python
- Default Branch: main
- Size: 350 KB
Statistics
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
- Releases: 0
Metadata Files
README.md
🐙GPTopus is a decoder-only transformer that generates new text based on input text.
GPTopus uses tiktoken, the OpenAI byte pair encoding (BPE) tokeniser, to encode and decode text.
📦 Install
To install required libraries for GPTopus from pip:
$ pip install torch, tiktoken
⛵ Quickstart
Figure 1. Training and Validation Loss for 4000 Iterations with frankestein.txt Dataset on 15GB TP4 GPU
Using OpenGPT:
* Get input text and place in same folder as gptopus.py
eg. frankenstein.txt
```
Letter 1
To Mrs. Saville, England.
St. Petersburgh, Dec. 11th, 17—.
You will rejoice to hear that no disaster has accompanied the commencement of an enterprise which you have regarded with such evil forebodings. I arrived here yesterday, and my first task is to assure my dear sister of my welfare and increasing confidence in the success of my undertaking.
I am already far north of London, and as I walk in the streets of Petersburgh, I feel a cold northern breeze play upon my cheeks, which braces my nerves and fills me with delight. Do you understand this feeling? This breeze, which has travelled from the regions towards which I am advancing, gives me a foretaste of those icy climes.
```
- Change hyperparameters
eg.
max_new_tkns = 1000 # generated text length
- set
filenameas input text
eg.
filename = 'frankenstein.txt' # set as input text
- Set gpt model
eg.
enc = tiktoken.encoding_for_model("gpt-4o") # change gpt model
- Generated text file created and placed in same folder as
gptopus.py, generated text file will use input filename inserting_out.txt
eg. frankenstein_out.txt
!uing all my family are for your destruction and
str yourself: I felt a strange companions of spirit as far different branches, my gaishown will be the winds, in
a gloom he said, so rapidly. But an
you answer
and I am so sister, Felix describe the lovely ship to this he would that I discovered as your fellow. My kind and warmth of my ryg had
tum, after the fut to others I am not describe; I darted by the
from the advocate of justice of my father whilst he
exception of age of possession to give our comfort on her so and almost trees him, that I distinctly
possible him, I reflect, Victor so dear to humanity of a
soension. It is very pause to fill the scene were extinguish percept me above
in the court days;, but one of the
shel from my father calmed himselfive the fresh weary.
🪪 License
Owner
- Name: Danté
- Login: dantethemartian
- Kind: user
- Location: The Moon
- Website: https://dantevangelista.com
- Repositories: 1
- Profile: https://github.com/dantethemartian
💭 ML 🧠 AI 🗣️ NLP/LLMs🔥Deep Learning 🧩 Lin Alg 🧬 Langs 👓 Comp Vis 👀 New Experiences & Traveling 🎲 Solving Complex Probs
Citation (citation.cff)
cff-version: 1.2.0 message: "If you use this software, please cite it as below." authors: - family-names: "Evangelista" given-names: "Danté" title: "gptopus" version: 1.0.0 doi: 10.5281/zenodo.1234 date-released: 2024-10-23 url: "https://github.com/dantevangelista/gptopus"