Transformers
File size: 11,655 Bytes
8626dcc
 
0f5ba3f
 
8626dcc
 
0f5ba3f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
---
library_name: transformers
datasets:
- SmallDoge/SmallCorpus
---

# Doge2-tokenizer

Tokenizer for the training model on [SmallCorpus](https://huggingface.co/datasets/SmallDoge/SmallCorpus), supporting document retrieval, tool invocation, and reasoning.
This tokenizer was trained on 2M samples from:
- Web-EN 50%
- Web-ZH 20%
- TextBook-EN 15%
- TextBook-ZH 5%
- Code 5%
- Math 5%

## How to use

<details>
<summary>Only conversation</summary>

```python
from transformers import AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("SmallDoge/Doge2-tokenizer")

conversation = [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "What's wrong with Cheems? What treatment does he need?"},
    {"role": "assistant", "content": "Based on Cheems' symptoms, I recommend the following treatment: For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters."},
]

inputs = tokenizer.apply_chat_template(
    conversation=conversation,
    tokenize=False,
)
print(inputs)
```
```shell
<|im_start|>system
You are a helpful assistant.<|im_end|>
<|im_start|>user
What's wrong with Cheems? What treatment does he need?<|im_end|>
<|im_start|>assistant
<think>

</think>

Based on Cheems' symptoms, I recommend the following treatment: For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters.<|im_end|>

```
</details>

<details>
<summary>With documents</summary>

```python
from transformers import AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("SmallDoge/Doge2-tokenizer")

documents = [
    {
        "title": "Cheems's case",
        "text": "Cheems is a 233-year-old alchemist, but he is very kidney deficient."
    },
]

conversation = [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "What's wrong with Cheems? What treatment does he need?"},
    {"role": "assistant", "content": "Based on Cheems' symptoms, I recommend the following treatment: For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters."},
]

inputs = tokenizer.apply_chat_template(
    documents=documents,
    conversation=conversation,
    tokenize=False,
)
print(inputs)
```
```shell
<|im_start|>system
You are a helpful assistant.

You have access to the following documents. Please use them to answer the user's question.

Title: Cheems's case
Content: Cheems is a 233-year-old alchemist, but he is very kidney deficient.

If the documents don't contain relevant information, rely on your general knowledge but acknowledge when you're doing so.<|im_end|>
<|im_start|>user
What's wrong with Cheems? What treatment does he need?<|im_end|>
<|im_start|>assistant
<think>

</think>

Based on Cheems' symptoms, I recommend the following treatment: For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters.<|im_end|>
```
</details>

<details>
<summary>With tools</summary>

```python
from transformers import AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("SmallDoge/Doge2-tokenizer")

tools = [
    {
        "name": "recommend_treatment",
        "description": "Treatment is recommended based on patient symptoms and diagnosis.",
        "arguments": {
            "type": "object",
            "properties": {
                "diagnosis": {
                    "type": "string",
                    "description": "Patient diagnostic results"
                },
                "severity": {
                    "type": "string",
                    "enum": ["mild", "moderate", "severe"],
                    "description": "Severity of the condition, can be mild, moderate, or severe."
                }
            },
            "required": ["diagnosis", "severity"]
        }
    }
]

conversation = [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "What's wrong with Cheems? What treatment does he need?"},
    {"role": "assistant", "content": "<tool_call>\n{\"name\": \"recommend_treatment\", \"arguments\": {\"diagnosis\": \"Deficiency of kidney\", \"severity\": \"severe\"}}\n</tool_call>"},
    {"role": "tool", "content": "For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters."},
    {"role": "assistant", "content": "Based on Cheems' symptoms, I recommend the following treatment: For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters."},
]

inputs = tokenizer.apply_chat_template(
    tools=tools,
    conversation=conversation,
    tokenize=False,
)
print(inputs)
```
```shell
<|im_start|>system
You are a helpful assistant.

You may call one or more functions to assist with the user query. You are provided with function signatures within <tools></tools> XML tags:

<tools>
{"name": "recommend_treatment", "description": "Treatment is recommended based on patient symptoms and diagnosis.", "arguments": {"type": "object", "properties": {"diagnosis": {"type": "string", "description": "Patient diagnostic results"}, "severity": {"type": "string", "enum": ["mild", "moderate", "severe"], "description": "Severity of the condition, can be mild, moderate, or severe."}}, "required": ["diagnosis", "severity"]}}
</tools>

For each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:

<tool_call>
{"name": <function-name>, "arguments": <args-json-object>}
</tool_call><|im_end|>
<|im_start|>user
What's wrong with Cheems? What treatment does he need?<|im_end|>
<|im_start|>assistant
<tool_call>
{"name": "recommend_treatment", "arguments": {"diagnosis": "Deficiency of kidney", "severity": "severe"}}
</tool_call><|im_end|>
<|im_start|>user
<tool_response>
For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters.
</tool_response><|im_end|>
<|im_start|>assistant
<think>

</think>

Based on Cheems' symptoms, I recommend the following treatment: For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters.<|im_end|>
```
</details>

<details>
<summary>With reasoning</summary>

```python
from transformers import AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("SmallDoge/Doge2-tokenizer")

conversation = [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "What's wrong with Cheems? What treatment does he need?"},
    {"role": "assistant", "reasoning_content": "Cheems is a dog, a old dog. He is been around for a long time, so his innards are not as good as before. If he is not careful, he will get kidney deficiency. He needs to eat more leeks and oysters.", "content": "Based on Cheems' symptoms, I recommend the following treatment: For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters."},
]

inputs = tokenizer.apply_chat_template(
    conversation=conversation,
    tokenize=False,
)
print(inputs)
```
```shell
<|im_start|>system
You are a helpful assistant.<|im_end|>
<|im_start|>user
What's wrong with Cheems? What treatment does he need?<|im_end|>
<|im_start|>assistant
<think>
Cheems is a dog, a old dog. He is been around for a long time, so his innards are not as good as before. If he is not careful, he will get kidney deficiency. He needs to eat more leeks and oysters.
</think>

Based on Cheems' symptoms, I recommend the following treatment: For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters.<|im_end|>
```
</details>

<details>
<summary>With documents tools reasoning</summary>

```python
from transformers import AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("SmallDoge/Doge2-tokenizer")

documents = [
    {
        "title": "Cheems's case",
        "text": "Cheems is a 233-year-old alchemist, but he is very kidney deficient."
    },
]

tools = [
    {
        "name": "recommend_treatment",
        "description": "Treatment is recommended based on patient symptoms and diagnosis.",
        "arguments": {
            "type": "object",
            "properties": {
                "diagnosis": {
                    "type": "string",
                    "description": "Patient diagnostic results"
                },
                "severity": {
                    "type": "string",
                    "enum": ["mild", "moderate", "severe"],
                    "description": "Severity of the condition, can be mild, moderate, or severe."
                }
            },
            "required": ["diagnosis", "severity"]
        }
    }
]

conversation = [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "What's wrong with Cheems? What treatment does he need?"},
    {"role": "assistant", "content": "<tool_call>\n{\"name\": \"recommend_treatment\", \"arguments\": {\"diagnosis\": \"Deficiency of kidney\", \"severity\": \"severe\"}}\n</tool_call>"},
    {"role": "tool", "content": "For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters."},
    {"role": "assistant", "reasoning_content": "Cheems is a dog, a old dog. He is been around for a long time, so his innards are not as good as before. If he is not careful, he will get kidney deficiency. He needs to eat more leeks and oysters.", "content": "Based on Cheems' symptoms, I recommend the following treatment: For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters."},
]

inputs = tokenizer.apply_chat_template(
    documents=documents,
    tools=tools,
    conversation=conversation,
    tokenize=False,
)
print(inputs)
```
```shell
<|im_start|>system
You are a helpful assistant.

You have access to the following documents. Please use them to answer the user's question.

Title: Cheems's case
Content: Cheems is a 233-year-old alchemist, but he is very kidney deficient.

If the documents don't contain relevant information, rely on your general knowledge but acknowledge when you're doing so.

You may call one or more functions to assist with the user query. You are provided with function signatures within <tools></tools> XML tags:

<tools>
{"name": "recommend_treatment", "description": "Treatment is recommended based on patient symptoms and diagnosis.", "arguments": {"type": "object", "properties": {"diagnosis": {"type": "string", "description": "Patient diagnostic results"}, "severity": {"type": "string", "enum": ["mild", "moderate", "severe"], "description": "Severity of the condition, can be mild, moderate, or severe."}}, "required": ["diagnosis", "severity"]}}
</tools>

For each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:

<tool_call>
{"name": <function-name>, "arguments": <args-json-object>}
</tool_call><|im_end|>
<|im_start|>user
What's wrong with Cheems? What treatment does he need?<|im_end|>
<|im_start|>assistant
<tool_call>
{"name": "recommend_treatment", "arguments": {"diagnosis": "Deficiency of kidney", "severity": "severe"}}
</tool_call><|im_end|>
<|im_start|>user
<tool_response>
For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters.
</tool_response><|im_end|>
<|im_start|>assistant
<think>
Cheems is a dog, a old dog. He is been around for a long time, so his innards are not as good as before. If he is not careful, he will get kidney deficiency. He needs to eat more leeks and oysters.
</think>

Based on Cheems' symptoms, I recommend the following treatment: For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters.<|im_end|>
```
</details>