abdelhalim
/

Rec_Business_Names

Text2Text Generation

Text2Text Generation

Recommendation system

text-generation-inference

Inference Endpoints

Model card Files Files and versions Community

abdelhalim commited on Mar 27, 2022

Commit

de82735

·

1 Parent(s): 2cf4437

Update README.md

Files changed (1) hide show

README.md +2 -2

README.md CHANGED Viewed

@@ -16,12 +16,12 @@ Most of the business name generator systems based on Rule based approach and onl
 **Introduction**
-The goal is to create an AI service which is helpful to people and yet could turn into a small business. After fiddling around with T5, I have realized it has an immense creative potential that could prove useful in creative text generation. So, after scraping around 350.000 websites from different Domain list, I have fine-tuned T5 small parameter model. Results are weirdly accurate and also creative at the same time.
 T5 small is already pre-trained language model which is capable of creating text with a near human quality. It's able to understand the context of a given prefix to generate text. When fine tuned based on the domain names and their meta context, it was able to understand the relation between domain name and the content of the website.
 **Dataset**
-t5 small needs lots of data to be trained properly. Quality of the data that we will use for fine tuning will have a direct effect on the model quality therefore we need to make sure the data we are scraping from the websites are as clean as possible. The dateset will not be available but under request.
 # Usage
 In order to use the model in your Python script just copy the following code:

 **Introduction**
+The goal is to create an AI service which is helpful to people and yet could turn into a small business. After fiddling around with T5, I have realized it has an immense creative potential that could prove useful in creative text generation. So, after scraping around 350.000 websites from different Domain list, I have fine-tuned T5 small parameter on this dataset. Results are much depends to the context and creative at the same time.
 T5 small is already pre-trained language model which is capable of creating text with a near human quality. It's able to understand the context of a given prefix to generate text. When fine tuned based on the domain names and their meta context, it was able to understand the relation between domain name and the content of the website.
 **Dataset**
+t5 small needs lots of data to be trained properly. Quality of the data that we will use for fine tuning will have a direct effect on the model quality therefore we need to make sure the data we are scraping from the websites are as clean as possible. The dateset will be under request.
 # Usage
 In order to use the model in your Python script just copy the following code: