|
--- |
|
tags: |
|
- autotrain |
|
- text-generation |
|
widget: |
|
- text: 'I love AutoTrain because ' |
|
license: other |
|
datasets: |
|
- SeanJIE250/llama2_law |
|
language: |
|
- en |
|
- zh |
|
--- |
|
# Contact Information |
|
Email:zhengjie1sun@gmail.com |
|
|
|
# English INTRO |
|
Give me some red stars ♥️ if u like this model! It's the model focused on Law field, honestly,doing bad as a daily chatbot however,start to know Mandarin and can handle the case study in details. |
|
|
|
# Mandarin INTRO |
|
老玩家点点红星♥️！中文法律对话机器人，具体案件审理较为不错。 |
|
|
|
# Usage |
|
First at first , implementing this command needs transformer library ,you can do the download directly.Hope u well! |
|
```python |
|
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer |
|
|
|
model_path = "SeanJIE250/chatbot_LAW" |
|
|
|
tokenizer = AutoTokenizer.from_pretrained(model_path) |
|
model = AutoModelForCausalLM.from_pretrained( |
|
model_path, |
|
device_map="auto", |
|
torch_dtype='auto' |
|
).eval() |
|
|
|
# Prompt content: "hi" |
|
messages = [ |
|
{"role": "user", "content": "杀了人在中国判多少年？"} |
|
] |
|
|
|
input_ids = tokenizer.apply_chat_template(conversation=messages, tokenize=True, add_generation_prompt=True, return_tensors='pt') |
|
outputs = model.generate(input_ids.to('cuda'),max_new_tokens=200)//you can adjust the max_new_tokens as you want. |
|
response = tokenizer.decode(outputs[0][input_ids.shape[1]:], skip_special_tokens=False) |
|
|
|
print(response) |
|
|
|
messages = [ |
|
{"role": "user", "content": "How to split the property if I divorced with my handsband?"} |
|
] |
|
|
|
input_ids = tokenizer.apply_chat_template(conversation=messages, tokenize=True, add_generation_prompt=True, return_tensors='pt') |
|
outputs = model.generate(input_ids.to('cuda'),max_new_tokens=200)//you can adjust the max_new_tokens as you want. |
|
response = tokenizer.decode(outputs[0][input_ids.shape[1]:], skip_special_tokens=False) |
|
|
|
print(response) |
|
``` |
