Commit
·
b0ba900
1
Parent(s):
292a662
update to inference engine
Browse files- README.md +14 -28
- RunTinyStories.cs +32 -59
- merges.txt → data/merges.txt +0 -0
- vocab.json → data/vocab.json +0 -0
- info.json +4 -4
- tinystories.onnx → models/tinystories.onnx +0 -0
- tinystories.sentis +0 -3
README.md
CHANGED
@@ -2,42 +2,28 @@
|
|
2 |
license: mit
|
3 |
library_name: unity-sentis
|
4 |
pipeline_tag: text-generation
|
|
|
|
|
5 |
---
|
6 |
|
7 |
-
# Tiny Stories
|
8 |
-
*Version 1.3.0 Sentis files are not compatible with Sentis 1.4.0 and would need to be recreated/downloaded
|
9 |
-
|
10 |
-
This is the [Tiny Stories model](https://huggingface.co/roneneldan/TinyStories-33M) checked to run on Unity 2023. Tiny Stories is a Large Language Model that was trained on children's stories and can create stories based on the first couple of sentences.
|
11 |
|
|
|
12 |
|
13 |
## How to Use
|
14 |
-
* Create a new scene in Unity 2023
|
15 |
-
* Install `com.unity.sentis` and `com.unity.nuget.newtonsoft-json` packages
|
16 |
-
* Add the RunTinyStories.cs file to the Main Camera
|
17 |
-
* Put `tinystories.sentis`, `vocab.json` and `merges.txt` in the Assets/StreamingAssets folder
|
18 |
-
* Adjust some of the variables such as the `outputText` string to set the prompt
|
19 |
-
* Press run
|
20 |
-
* The output will appear in the console window
|
21 |
-
|
22 |
-
## Example Input
|
23 |
-
```
|
24 |
-
One day an alien came down from Mars. It saw a chicken
|
25 |
-
```
|
26 |
-
## Example Output
|
27 |
-
```
|
28 |
-
One day an alien came down from Mars. It saw a chicken and said, "Hello, little chicken. What are you doing here?"
|
29 |
-
|
30 |
-
The chicken replied, "I'm looking for a place to stay. I'm very tired."
|
31 |
-
|
32 |
-
The alien said, "You can stay here. I have a nice place for you. It's very comfortable."
|
33 |
|
34 |
-
|
|
|
|
|
|
|
|
|
|
|
35 |
|
36 |
-
|
37 |
-
|
38 |
|
39 |
-
##
|
40 |
-
|
41 |
|
42 |
## Disclaimer
|
43 |
The model was trained on children's stories so very unlikely to produce undesirable text. As an extra precaution, we removed a few tokens from vocab.json that might not be suitable for younger audiences. The original json can be found on the Tiny Stories original page.
|
|
|
2 |
license: mit
|
3 |
library_name: unity-sentis
|
4 |
pipeline_tag: text-generation
|
5 |
+
tags:
|
6 |
+
- unity-inference-engine
|
7 |
---
|
8 |
|
9 |
+
# Tiny Stories in Unity 6 with Inference Engine
|
|
|
|
|
|
|
10 |
|
11 |
+
This is the [Tiny Stories model](https://huggingface.co/roneneldan/TinyStories-33M) running in Unity 6 with Inference Engine. Tiny Stories is a Large Language Model that was trained on children's stories and can create stories based on the first couple of sentences.
|
12 |
|
13 |
## How to Use
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
14 |
|
15 |
+
* Create a new scene in Unity 6;
|
16 |
+
* Install `com.unity.ai.inference` from the package manager;
|
17 |
+
* Install `com.unity.nuget.newtonsoft-json` from the package manager;
|
18 |
+
* Drag the `tinystories.onnx` asset from the `models` folder into the `Model Asset` field;
|
19 |
+
* Drag the `vocab.json` asset from the `data` folder into the `Vocab Asset` field;
|
20 |
+
* Drag the `merges.txt` asset from the `data` folder into the `Merges Asset` field;
|
21 |
|
22 |
+
## Preview
|
23 |
+
Enter play mode. If working correctly the predicted text will be logged to the console.
|
24 |
|
25 |
+
## Inference Engine
|
26 |
+
Inference Engine is a neural network inference library for Unity. Find out more [here](https://docs.unity3d.com/Packages/com.unity.ai.inference@latest).
|
27 |
|
28 |
## Disclaimer
|
29 |
The model was trained on children's stories so very unlikely to produce undesirable text. As an extra precaution, we removed a few tokens from vocab.json that might not be suitable for younger audiences. The original json can be found on the Tiny Stories original page.
|
RunTinyStories.cs
CHANGED
@@ -1,33 +1,14 @@
|
|
1 |
-
using System.Collections;
|
2 |
using System.Collections.Generic;
|
3 |
using UnityEngine;
|
4 |
-
using Unity.
|
5 |
-
using System.IO;
|
6 |
using System.Text;
|
7 |
-
using FF = Unity.
|
8 |
-
|
9 |
-
/*
|
10 |
-
* Tiny Stories Inference Code
|
11 |
-
* ===========================
|
12 |
-
*
|
13 |
-
* Put this script on the Main Camera
|
14 |
-
*
|
15 |
-
* In Assets/StreamingAssets put:
|
16 |
-
*
|
17 |
-
* tinystories.sentis (or put in asset folder and drag onto field)
|
18 |
-
* vocab.json
|
19 |
-
* merges.txt
|
20 |
-
*
|
21 |
-
* Install package com.unity.nuget.newtonsoft-json from packagemanger
|
22 |
-
* Install package com.unity.sentis
|
23 |
-
*
|
24 |
-
*/
|
25 |
-
|
26 |
|
27 |
public class RunTinyStories : MonoBehaviour
|
28 |
{
|
29 |
-
|
30 |
-
|
|
|
31 |
const BackendType backend = BackendType.GPUCompute;
|
32 |
|
33 |
//string outputString = "Once upon a time, there were three bears";
|
@@ -45,22 +26,21 @@ public class RunTinyStories : MonoBehaviour
|
|
45 |
//Store the vocabulary
|
46 |
string[] tokens;
|
47 |
|
48 |
-
|
49 |
|
50 |
-
int currentToken
|
51 |
int[] outputTokens = new int[maxTokens];
|
52 |
|
53 |
// Used for special character decoding
|
54 |
int[] whiteSpaceCharacters = new int[256];
|
55 |
int[] encodedCharacters = new int[256];
|
56 |
|
57 |
-
bool runInference
|
58 |
-
|
59 |
|
60 |
//stop after this many tokens
|
61 |
const int stopAfter = 100;
|
62 |
|
63 |
-
int totalTokens
|
64 |
|
65 |
string[] merges;
|
66 |
Dictionary<string, int> vocab;
|
@@ -71,19 +51,17 @@ public class RunTinyStories : MonoBehaviour
|
|
71 |
|
72 |
LoadVocabulary();
|
73 |
|
74 |
-
var model1 = ModelLoader.Load(
|
75 |
-
//var model1 = ModelLoader.Load(asset);
|
76 |
//Create a new model to select the random token:
|
77 |
-
var model2 = FF.Compile(
|
78 |
-
(input, currentToken) =>
|
79 |
-
{
|
80 |
-
var row = FF.Select(model1.Forward(input)[8], 1, currentToken);
|
81 |
-
return FF.Multinomial(predictability * row, 1);
|
82 |
-
},
|
83 |
-
(model1.inputs[0], InputDef.Int(new TensorShape()))
|
84 |
-
);
|
85 |
|
86 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
87 |
|
88 |
DecodePrompt(outputString);
|
89 |
|
@@ -101,16 +79,14 @@ public class RunTinyStories : MonoBehaviour
|
|
101 |
|
102 |
void RunInference()
|
103 |
{
|
104 |
-
using var tokensSoFar = new
|
105 |
-
using var index = new
|
106 |
|
107 |
-
engine.
|
108 |
|
109 |
-
var probs = engine.PeekOutput() as
|
110 |
Debug.Log(probs.shape);
|
111 |
|
112 |
-
probs.CompleteOperationsAndDownload();
|
113 |
-
|
114 |
int ID = probs[0];
|
115 |
|
116 |
//shift window down if got to the end
|
@@ -130,23 +106,22 @@ public class RunTinyStories : MonoBehaviour
|
|
130 |
else outputString += GetUnicodeText(tokens[ID]);
|
131 |
|
132 |
Debug.Log(outputString);
|
133 |
-
|
134 |
}
|
135 |
|
136 |
void DecodePrompt(string text)
|
137 |
{
|
138 |
var inputTokens = GetTokens(text);
|
139 |
|
140 |
-
for(int i = 0; i < inputTokens.Count; i++)
|
141 |
{
|
142 |
outputTokens[i] = inputTokens[i];
|
143 |
}
|
144 |
currentToken = inputTokens.Count - 1;
|
145 |
}
|
146 |
-
|
147 |
void LoadVocabulary()
|
148 |
{
|
149 |
-
var jsonText =
|
150 |
vocab = Newtonsoft.Json.JsonConvert.DeserializeObject<Dictionary<string, int>>(jsonText);
|
151 |
tokens = new string[vocab.Count];
|
152 |
foreach (var item in vocab)
|
@@ -154,7 +129,7 @@ public class RunTinyStories : MonoBehaviour
|
|
154 |
tokens[item.Value] = item.Key;
|
155 |
}
|
156 |
|
157 |
-
merges =
|
158 |
}
|
159 |
|
160 |
// Translates encoded special characters to Unicode
|
@@ -174,8 +149,7 @@ public class RunTinyStories : MonoBehaviour
|
|
174 |
string outText = "";
|
175 |
foreach (char letter in text)
|
176 |
{
|
177 |
-
outText += (
|
178 |
-
(char)whiteSpaceCharacters[(int)(letter - 256)];
|
179 |
}
|
180 |
return outText;
|
181 |
}
|
@@ -185,7 +159,7 @@ public class RunTinyStories : MonoBehaviour
|
|
185 |
string outText = "";
|
186 |
foreach (char letter in text)
|
187 |
{
|
188 |
-
outText += (char)encodedCharacters[
|
189 |
}
|
190 |
return outText;
|
191 |
}
|
@@ -215,7 +189,7 @@ public class RunTinyStories : MonoBehaviour
|
|
215 |
|
216 |
// Start with a list of single characters
|
217 |
var inputTokens = new List<string>();
|
218 |
-
foreach(var letter in text)
|
219 |
{
|
220 |
inputTokens.Add(letter.ToString());
|
221 |
}
|
@@ -224,7 +198,7 @@ public class RunTinyStories : MonoBehaviour
|
|
224 |
|
225 |
//Find the ids of the words in the vocab
|
226 |
var ids = new List<int>();
|
227 |
-
foreach(var token in inputTokens)
|
228 |
{
|
229 |
if (vocab.TryGetValue(token, out int id))
|
230 |
{
|
@@ -237,7 +211,7 @@ public class RunTinyStories : MonoBehaviour
|
|
237 |
|
238 |
void ApplyMerges(List<string> inputTokens)
|
239 |
{
|
240 |
-
foreach(var merge in merges)
|
241 |
{
|
242 |
string[] pair = merge.Split(' ');
|
243 |
int n = 0;
|
@@ -254,9 +228,8 @@ public class RunTinyStories : MonoBehaviour
|
|
254 |
}
|
255 |
}
|
256 |
|
257 |
-
|
258 |
{
|
259 |
engine?.Dispose();
|
260 |
}
|
261 |
-
|
262 |
}
|
|
|
|
|
1 |
using System.Collections.Generic;
|
2 |
using UnityEngine;
|
3 |
+
using Unity.InferenceEngine;
|
|
|
4 |
using System.Text;
|
5 |
+
using FF = Unity.InferenceEngine.Functional;
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
6 |
|
7 |
public class RunTinyStories : MonoBehaviour
|
8 |
{
|
9 |
+
public ModelAsset modelAsset;
|
10 |
+
public TextAsset vocabAsset;
|
11 |
+
public TextAsset mergesAsset;
|
12 |
const BackendType backend = BackendType.GPUCompute;
|
13 |
|
14 |
//string outputString = "Once upon a time, there were three bears";
|
|
|
26 |
//Store the vocabulary
|
27 |
string[] tokens;
|
28 |
|
29 |
+
Worker engine;
|
30 |
|
31 |
+
int currentToken;
|
32 |
int[] outputTokens = new int[maxTokens];
|
33 |
|
34 |
// Used for special character decoding
|
35 |
int[] whiteSpaceCharacters = new int[256];
|
36 |
int[] encodedCharacters = new int[256];
|
37 |
|
38 |
+
bool runInference;
|
|
|
39 |
|
40 |
//stop after this many tokens
|
41 |
const int stopAfter = 100;
|
42 |
|
43 |
+
int totalTokens;
|
44 |
|
45 |
string[] merges;
|
46 |
Dictionary<string, int> vocab;
|
|
|
51 |
|
52 |
LoadVocabulary();
|
53 |
|
54 |
+
var model1 = ModelLoader.Load(modelAsset);
|
|
|
55 |
//Create a new model to select the random token:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
56 |
|
57 |
+
var graph = new FunctionalGraph();
|
58 |
+
var input = graph.AddInput(model1, 0);
|
59 |
+
var currentTokenInput = graph.AddInput<int>(new TensorShape(), "currentToken");
|
60 |
+
var row = FF.Select(Functional.Forward(model1, input)[0], 1, currentTokenInput);
|
61 |
+
var output = FF.Multinomial(predictability * row, 1);
|
62 |
+
var model2 = graph.Compile(output);
|
63 |
+
|
64 |
+
engine = new Worker(model2, backend);
|
65 |
|
66 |
DecodePrompt(outputString);
|
67 |
|
|
|
79 |
|
80 |
void RunInference()
|
81 |
{
|
82 |
+
using var tokensSoFar = new Tensor<int>(new TensorShape(1, maxTokens), outputTokens);
|
83 |
+
using var index = new Tensor<int>(new TensorShape(), new[] { currentToken });
|
84 |
|
85 |
+
engine.Schedule(tokensSoFar, index);
|
86 |
|
87 |
+
using var probs = (engine.PeekOutput() as Tensor<int>).ReadbackAndClone();
|
88 |
Debug.Log(probs.shape);
|
89 |
|
|
|
|
|
90 |
int ID = probs[0];
|
91 |
|
92 |
//shift window down if got to the end
|
|
|
106 |
else outputString += GetUnicodeText(tokens[ID]);
|
107 |
|
108 |
Debug.Log(outputString);
|
|
|
109 |
}
|
110 |
|
111 |
void DecodePrompt(string text)
|
112 |
{
|
113 |
var inputTokens = GetTokens(text);
|
114 |
|
115 |
+
for (int i = 0; i < inputTokens.Count; i++)
|
116 |
{
|
117 |
outputTokens[i] = inputTokens[i];
|
118 |
}
|
119 |
currentToken = inputTokens.Count - 1;
|
120 |
}
|
121 |
+
|
122 |
void LoadVocabulary()
|
123 |
{
|
124 |
+
var jsonText = vocabAsset.text;
|
125 |
vocab = Newtonsoft.Json.JsonConvert.DeserializeObject<Dictionary<string, int>>(jsonText);
|
126 |
tokens = new string[vocab.Count];
|
127 |
foreach (var item in vocab)
|
|
|
129 |
tokens[item.Value] = item.Key;
|
130 |
}
|
131 |
|
132 |
+
merges = mergesAsset.text.Split("\r\n");
|
133 |
}
|
134 |
|
135 |
// Translates encoded special characters to Unicode
|
|
|
149 |
string outText = "";
|
150 |
foreach (char letter in text)
|
151 |
{
|
152 |
+
outText += (letter <= 256) ? letter : (char)whiteSpaceCharacters[letter - 256];
|
|
|
153 |
}
|
154 |
return outText;
|
155 |
}
|
|
|
159 |
string outText = "";
|
160 |
foreach (char letter in text)
|
161 |
{
|
162 |
+
outText += (char)encodedCharacters[letter];
|
163 |
}
|
164 |
return outText;
|
165 |
}
|
|
|
189 |
|
190 |
// Start with a list of single characters
|
191 |
var inputTokens = new List<string>();
|
192 |
+
foreach (var letter in text)
|
193 |
{
|
194 |
inputTokens.Add(letter.ToString());
|
195 |
}
|
|
|
198 |
|
199 |
//Find the ids of the words in the vocab
|
200 |
var ids = new List<int>();
|
201 |
+
foreach (var token in inputTokens)
|
202 |
{
|
203 |
if (vocab.TryGetValue(token, out int id))
|
204 |
{
|
|
|
211 |
|
212 |
void ApplyMerges(List<string> inputTokens)
|
213 |
{
|
214 |
+
foreach (var merge in merges)
|
215 |
{
|
216 |
string[] pair = merge.Split(' ');
|
217 |
int n = 0;
|
|
|
228 |
}
|
229 |
}
|
230 |
|
231 |
+
void OnDestroy()
|
232 |
{
|
233 |
engine?.Dispose();
|
234 |
}
|
|
|
235 |
}
|
merges.txt → data/merges.txt
RENAMED
File without changes
|
vocab.json → data/vocab.json
RENAMED
File without changes
|
info.json
CHANGED
@@ -3,13 +3,13 @@
|
|
3 |
"RunTinyStories.cs"
|
4 |
],
|
5 |
"models": [
|
6 |
-
"tinystories.
|
7 |
],
|
8 |
"data": [
|
9 |
-
"vocab.json",
|
10 |
-
"merges.txt"
|
11 |
],
|
12 |
"version": [
|
13 |
-
"
|
14 |
]
|
15 |
}
|
|
|
3 |
"RunTinyStories.cs"
|
4 |
],
|
5 |
"models": [
|
6 |
+
"models/tinystories.onnx"
|
7 |
],
|
8 |
"data": [
|
9 |
+
"data/vocab.json",
|
10 |
+
"data/merges.txt"
|
11 |
],
|
12 |
"version": [
|
13 |
+
"2.2.0"
|
14 |
]
|
15 |
}
|
tinystories.onnx → models/tinystories.onnx
RENAMED
File without changes
|
tinystories.sentis
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:c7962eb7db56b241cc19cd3f0cffcf5d76d3c35639917f07effa6b3c242c91e9
|
3 |
-
size 478818076
|
|
|
|
|
|
|
|