|
|
--- |
|
|
library_name: transformers |
|
|
datasets: |
|
|
- HuggingFaceTB/smollm-corpus |
|
|
--- |
|
|
|
|
|
# Doge-tokenizer |
|
|
Tokenizer for the training model on [smollm-corpus](https://huggingface.co/datasets/HuggingFaceTB/smollm-corpus), and support reasoning fine-tuning like R1. |
|
|
This tokenizer was trained on 2M samples from: |
|
|
- FineWeb-Edu 70% |
|
|
- Cosmopedia v2 20% |
|
|
- Python-Edu 5% |
|
|
- FineMath 5% |
|
|
|
|
|
## How to use |
|
|
|
|
|
<details> |
|
|
<summary>Only conversation</summary> |
|
|
|
|
|
```python |
|
|
from transformers import AutoTokenizer |
|
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained("SmallDoge/Doge-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 |
|
|
<|begin_of_text|><|start_header_id|>system<|end_header_id|> |
|
|
Cutting Knowledge Date: December 2024 |
|
|
Today Date: December 2025 |
|
|
You are a helpful assistant.<|end_of_text|> |
|
|
|
|
|
<|start_header_id|>user<|end_header_id|> |
|
|
What's wrong with Cheems? What treatment does he need?<|end_of_text|> |
|
|
|
|
|
<|start_header_id|>assistant<|end_header_id|> |
|
|
Based on Cheems' symptoms, I recommend the following treatment: For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters.<|end_of_text|> |
|
|
``` |
|
|
</details> |
|
|
|
|
|
<details> |
|
|
<summary>With documents</summary> |
|
|
|
|
|
```python |
|
|
from transformers import AutoTokenizer |
|
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained("SmallDoge/Doge-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 |
|
|
<|begin_of_text|><|start_header_id|>system<|end_header_id|> |
|
|
Cutting Knowledge Date: December 2024 |
|
|
Today Date: December 2025 |
|
|
You are a helpful assistant.<|end_of_text|> |
|
|
|
|
|
<|start_header_id|>user<|end_header_id|> |
|
|
Given the following documents, please use them to answer the user's question. |
|
|
|
|
|
Title: Cheems' 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. |
|
|
|
|
|
What's wrong with Cheems? What treatment does he need?<|end_of_text|> |
|
|
|
|
|
<|start_header_id|>assistant<|end_header_id|> |
|
|
Based on Cheems' symptoms, I recommend the following treatment: For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters.<|end_of_text|> |
|
|
``` |
|
|
</details> |
|
|
|
|
|
<details> |
|
|
<summary>With tools</summary> |
|
|
|
|
|
```python |
|
|
from transformers import AutoTokenizer |
|
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained("SmallDoge/Doge-tokenizer") |
|
|
|
|
|
tools = [ |
|
|
{ |
|
|
"name": "recommend_treatment", |
|
|
"description": "Treatment is recommended based on patient symptoms and diagnosis.", |
|
|
"parameters": { |
|
|
"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", "tool_calls": [ |
|
|
{ |
|
|
"id": "call_1", |
|
|
"function": { |
|
|
"name": "recommend_treatment", |
|
|
"arguments": "{\"diagnosis\": \"Deficiency of kidney\", \"severity\": \"severe\"}" |
|
|
} |
|
|
} |
|
|
]}, |
|
|
{"role": "tool", "tool_call_id": "call_1", "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 |
|
|
<|begin_of_text|><|start_header_id|>system<|end_header_id|> |
|
|
Environment: ipython |
|
|
Cutting Knowledge Date: December 2024 |
|
|
Today Date: December 2025 |
|
|
You are a helpful assistant.<|end_of_text|> |
|
|
|
|
|
<|start_header_id|>user<|end_header_id|> |
|
|
Given the following functions, please respond with a JSON for a function call with its proper arguments that best answers the given prompt. |
|
|
|
|
|
{ |
|
|
"name": "recommend_treatment", |
|
|
"description": "Treatment is recommended based on patient symptoms and diagnosis.", |
|
|
"parameters": { |
|
|
"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" |
|
|
] |
|
|
} |
|
|
} |
|
|
|
|
|
Respond in the format {"name": function name, "parameters": dictionary of argument name and its value}. Do not use variables. |
|
|
|
|
|
What's wrong with Cheems? What treatment does he need?<|end_of_text|> |
|
|
|
|
|
<|start_header_id|>assistant<|end_header_id|> |
|
|
{"name": "recommend_treatment", "parameters": "{\"diagnosis\": \"Deficiency of kidney\", \"severity\": \"severe\"}"}<|end_of_text|> |
|
|
|
|
|
<|start_header_id|>ipython<|end_header_id|> |
|
|
"For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters."<|end_of_text|> |
|
|
|
|
|
<|start_header_id|>assistant<|end_header_id|> |
|
|
Based on Cheems' symptoms, I recommend the following treatment: For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters.<|end_of_text|> |
|
|
``` |
|
|
</details> |
|
|
|
|
|
<details> |
|
|
<summary>With documents and tools</summary> |
|
|
|
|
|
```python |
|
|
from transformers import AutoTokenizer |
|
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained("SmallDoge/Doge-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.", |
|
|
"parameters": { |
|
|
"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", "tool_calls": [ |
|
|
{ |
|
|
"id": "call_1", |
|
|
"function": { |
|
|
"name": "recommend_treatment", |
|
|
"arguments": "{\"diagnosis\": \"Deficiency of kidney\", \"severity\": \"severe\"}" |
|
|
} |
|
|
} |
|
|
]}, |
|
|
{"role": "tool", "tool_call_id": "call_1", "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( |
|
|
documents=documents, |
|
|
tools=tools, |
|
|
conversation=conversation, |
|
|
tokenize=False, |
|
|
) |
|
|
print(inputs) |
|
|
``` |
|
|
```shell |
|
|
<|begin_of_text|><|start_header_id|>system<|end_header_id|> |
|
|
Environment: ipython |
|
|
Cutting Knowledge Date: December 2024 |
|
|
Today Date: December 2025 |
|
|
You are a helpful assistant.<|end_of_text|> |
|
|
|
|
|
<|start_header_id|>user<|end_header_id|> |
|
|
Given the following documents, please use them to answer the user's question. |
|
|
|
|
|
Title: Cheems' 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. |
|
|
|
|
|
Given the following functions, please respond with a JSON for a function call with its proper arguments that best answers the given prompt. |
|
|
|
|
|
{ |
|
|
"name": "recommend_treatment", |
|
|
"description": "Treatment is recommended based on patient symptoms and diagnosis.", |
|
|
"parameters": { |
|
|
"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" |
|
|
] |
|
|
} |
|
|
} |
|
|
|
|
|
Respond in the format {"name": function name, "parameters": dictionary of argument name and its value}. Do not use variables. |
|
|
|
|
|
What's wrong with Cheems? What treatment does he need?<|end_of_text|> |
|
|
|
|
|
<|start_header_id|>assistant<|end_header_id|> |
|
|
{"name": "recommend_treatment", "parameters": "{\"diagnosis\": \"Deficiency of kidney\", \"severity\": \"severe\"}"}<|end_of_text|> |
|
|
|
|
|
<|start_header_id|>ipython<|end_header_id|> |
|
|
"For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters."<|end_of_text|> |
|
|
|
|
|
<|start_header_id|>assistant<|end_header_id|> |
|
|
Based on Cheems' symptoms, I recommend the following treatment: For Cheems with severe kidney deficiency, it is recommended to eat more leeks and oysters.<|end_of_text|> |
|
|
``` |
|
|
</details> |
