Update README.md
Browse files
README.md
CHANGED
|
@@ -3,15 +3,56 @@ license: unlicense
|
|
| 3 |
tags:
|
| 4 |
- text2text-generation
|
| 5 |
---
|
| 6 |
-
# Alphabet
|
| 7 |
|
| 8 |
-
|
| 9 |
|
| 10 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
|
| 12 |
-
|
|
|
|
| 13 |
|
| 14 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
|
| 16 |
```txt
|
| 17 |
['A'] -> B
|
|
@@ -39,4 +80,4 @@ You can judge for yourself
|
|
| 39 |
['W'] -> Z
|
| 40 |
['X'] -> Z
|
| 41 |
['Y'] -> Z
|
| 42 |
-
```
|
|
|
|
| 3 |
tags:
|
| 4 |
- text2text-generation
|
| 5 |
---
|
|
|
|
| 6 |
|
| 7 |
+
# The Alphabetizer™️
|
| 8 |
|
| 9 |
+
## Overview
|
| 10 |
+
**Model Name**: The Alphabetizer™️
|
| 11 |
+
**Version**: 1.
|
| 12 |
+
**Purpose**: To predict the next letter in the alphabet, because reciting ABCs is hard.
|
| 13 |
+
**Date**: September 6, 2023
|
| 14 |
|
| 15 |
+
## Intended Use
|
| 16 |
+
For those moments when you're too overwhelmed to remember what comes after "A". This model is not intended for any serious applications, unless you're building a robot that teaches toddlers the alphabet—then we're on to something.
|
| 17 |
|
| 18 |
+
## Performance Metrics
|
| 19 |
+
- Accuracy: Probably around 100% on a good day.
|
| 20 |
+
- Latency: Faster than you can say "Alphabetti Spaghetti."
|
| 21 |
+
|
| 22 |
+
## Limitations
|
| 23 |
+
- Cannot predict the next letter in any sequence other than the English alphabet.
|
| 24 |
+
- Will not improve your Scrabble game.
|
| 25 |
+
- Does not know the difference between 'a' and 'A'; case-sensitive like a sensitive poet.
|
| 26 |
+
|
| 27 |
+
## Ethical Considerations
|
| 28 |
+
No alphabets were harmed during the training of this model.
|
| 29 |
+
|
| 30 |
+
## Data
|
| 31 |
+
**Source**: The 26 letters of the English alphabet.
|
| 32 |
+
**Quality**: Top-notch, handpicked, and farm-to-table alphabets.
|
| 33 |
+
**Size**: A whopping 26 letters!
|
| 34 |
+
|
| 35 |
+
## Architecture
|
| 36 |
+
Built on a single-layer LSTM network because let's not get carried away. It's just the alphabet, folks.
|
| 37 |
+
|
| 38 |
+
## Training
|
| 39 |
+
**Algorithm**: TensorFlow + Keras
|
| 40 |
+
**Epochs**: 500, because overfitting is just a number, right?
|
| 41 |
+
**Batch Size**: 1, we give individual attention to each letter.
|
| 42 |
+
|
| 43 |
+
## Output Interpretation
|
| 44 |
+
The model will output a letter, which will invariably be the next letter in the alphabet. Brace yourselves.
|
| 45 |
+
|
| 46 |
+
## Responsible AI Practices
|
| 47 |
+
We're still searching for the part of this that could be considered "AI".
|
| 48 |
+
|
| 49 |
+
## Update Policy
|
| 50 |
+
We might consider adding numbers if the model gets bored.
|
| 51 |
+
|
| 52 |
+
## Contact
|
| 53 |
+
For feedback, compliments, or your best alphabet jokes, please contact: `[email protected]`
|
| 54 |
+
|
| 55 |
+
## Output
|
| 56 |
|
| 57 |
```txt
|
| 58 |
['A'] -> B
|
|
|
|
| 80 |
['W'] -> Z
|
| 81 |
['X'] -> Z
|
| 82 |
['Y'] -> Z
|
| 83 |
+
```
|