sentis-tiny-stories / RunTinyStories.cs
Paul Bird
Upload RunTinyStories.cs
1d9995e verified
raw
history blame
6.81 kB
using System.Collections;
using System.Collections.Generic;
using UnityEngine;
using Unity.Sentis;
using System.IO;
using Newtonsoft.Json;
using System.Text;
/*
* Tiny Stories Inference Code
* ===========================
*
* Put this script on the Main Camera
*
* In Assets/StreamingAssets put:
*
* tinystories.sentis
* vocab.json
* merges.txt
*
* Install package com.unity.nuget.newtonsoft-json from packagemanger
* Install package com.unity.sentis
*
*/
public class RunTinyStories : MonoBehaviour
{
const BackendType backend = BackendType.GPUCompute;
//string outputString = "Once upon a time, there were three bears";
string outputString = "One day an alien came down from Mars. It saw a chicken";
// This is how many tokens you want. It can be adjusted.
const int maxTokens = 100;
//Make this smaller for more randomness
const float predictability = 5f;
//Special tokens
const int END_OF_TEXT = 50256;
Ops ops;
ITensorAllocator allocator;
//Store the vocabulary
string[] tokens;
IWorker engine;
int currentToken = 0;
int[] outputTokens = new int[maxTokens];
// Used for special character decoding
int[] whiteSpaceCharacters = new int[256];
int[] encodedCharacters = new int[256];
bool runInference = false;
//stop after this many tokens
const int stopAfter = 100;
int totalTokens = 0;
string[] merges;
Dictionary<string, int> vocab;
void Start()
{
allocator = new TensorCachingAllocator();
ops = WorkerFactory.CreateOps(backend, allocator);
SetupWhiteSpaceShifts();
LoadVocabulary();
Model model = ModelLoader.Load(Application.streamingAssetsPath + "/tinystories.sentis");
engine = WorkerFactory.CreateWorker(backend, model);
DecodePrompt(outputString);
runInference = true;
}
// Update is called once per frame
void Update()
{
if (runInference)
{
RunInference();
}
}
void RunInference()
{
using var tokensSoFar = new TensorInt(new TensorShape(1, maxTokens), outputTokens);
engine.Execute(tokensSoFar);
var tokensOut = engine.PeekOutput() as TensorFloat;
using var row = ops.Slice(tokensOut, new[] { currentToken }, new[] { currentToken + 1 }, new[] { 1 }, new[] { 1 });
using var rowB = ops.Mul(predictability, row);
using var probs = ops.Softmax(rowB, 2);
probs.MakeReadable();
int ID = SelectRandomToken(probs.ToReadOnlyArray());
if (currentToken >= maxTokens - 1)
{
for (int i = 0; i < maxTokens - 1; i++) outputTokens[i] = outputTokens[i + 1];
currentToken--;
}
outputTokens[++currentToken] = ID;
totalTokens++;
if (ID == END_OF_TEXT || totalTokens >= stopAfter)
{
runInference = false;
}
else outputString += GetUnicodeText(tokens[ID]);
Debug.Log(outputString);
}
void DecodePrompt(string text)
{
var inputTokens = GetTokens(text);
for(int i = 0; i < inputTokens.Count; i++)
{
outputTokens[i] = inputTokens[i];
}
currentToken = inputTokens.Count - 1;
}
void LoadVocabulary()
{
var jsonText = File.ReadAllText(Application.streamingAssetsPath + "/vocab.json");
vocab = Newtonsoft.Json.JsonConvert.DeserializeObject<Dictionary<string, int>>(jsonText);
tokens = new string[vocab.Count];
foreach (var item in vocab)
{
tokens[item.Value] = item.Key;
}
merges = File.ReadAllLines(Application.streamingAssetsPath + "/merges.txt");
}
int SelectRandomToken(float[] probs)
{
float p = UnityEngine.Random.Range(0, 1f);
float t = 0;
for (int i = 0; i < probs.Length; i++)
{
t += probs[i];
if (p < t)
{
return i;
}
}
return probs.Length - 1;
}
// Translates encoded special characters to Unicode
string GetUnicodeText(string text)
{
var bytes = Encoding.GetEncoding("ISO-8859-1").GetBytes(ShiftCharacterDown(text));
return Encoding.UTF8.GetString(bytes);
}
string GetASCIIText(string newText)
{
var bytes = Encoding.UTF8.GetBytes(newText);
return ShiftCharacterUp(Encoding.GetEncoding("ISO-8859-1").GetString(bytes));
}
string ShiftCharacterDown(string text)
{
string outText = "";
foreach (char letter in text)
{
outText += ((int)letter <= 256) ? letter :
(char)whiteSpaceCharacters[(int)(letter - 256)];
}
return outText;
}
string ShiftCharacterUp(string text)
{
string outText = "";
foreach (char letter in text)
{
outText += (char)encodedCharacters[(int)letter];
}
return outText;
}
void SetupWhiteSpaceShifts()
{
for (int i = 0, n = 0; i < 256; i++)
{
encodedCharacters[i] = i;
if (IsWhiteSpace((char)i))
{
encodedCharacters[i] = n + 256;
whiteSpaceCharacters[n++] = i;
}
}
}
bool IsWhiteSpace(char c)
{
return !(('!' <= c && c <= '~') || ('�' <= c && c <= '�') || ('�' <= c && c <= '�'));
}
List<int> GetTokens(string text)
{
text = GetASCIIText(text);
// Start with a list of single characters
var inputTokens = new List<string>();
foreach(var letter in text)
{
inputTokens.Add(letter.ToString());
}
ApplyMerges(inputTokens);
//Find the ids of the words in the vocab
var ids = new List<int>();
foreach(var token in inputTokens)
{
if (vocab.TryGetValue(token, out int id))
{
ids.Add(id);
}
}
return ids;
}
void ApplyMerges(List<string> inputTokens)
{
foreach(var merge in merges)
{
string[] pair = merge.Split(' ');
int n = 0;
while (n >= 0)
{
n = inputTokens.IndexOf(pair[0], n);
if (n != -1 && n < inputTokens.Count - 1 && inputTokens[n + 1] == pair[1])
{
inputTokens[n] += inputTokens[n + 1];
inputTokens.RemoveAt(n + 1);
}
if (n != -1) n++;
}
}
}
private void OnDestroy()
{
engine?.Dispose();
ops?.Dispose();
allocator?.Dispose();
}
}