Update README.md
Browse files
README.md
CHANGED
@@ -50,7 +50,7 @@ Moreover, we scale up our base model to LLaMA-1-13B to see if **our method is si
|
|
50 |
## Domain-Specific LLaMA-2-Chat
|
51 |
Our method is also effective for aligned models! LLaMA-2-Chat requires a [specific data format](https://huggingface.co/blog/llama2#how-to-prompt-llama-2), and our **reading comprehension can perfectly fit the data format** by transforming the reading comprehension into a multi-turn conversation. We have also open-sourced chat models in different domains: [Biomedicine-Chat](https://huggingface.co/AdaptLLM/medicine-chat), [Finance-Chat](https://huggingface.co/AdaptLLM/finance-chat) and [Law-Chat](https://huggingface.co/AdaptLLM/law-chat)
|
52 |
|
53 |
-
For example, to chat with the law model:
|
54 |
```python
|
55 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
56 |
|
@@ -67,7 +67,7 @@ Options:
|
|
67 |
|
68 |
Please provide your choice first and then provide explanations if possible.'''
|
69 |
|
70 |
-
# We use the prompt template of LLaMA-2-Chat demo
|
71 |
prompt = f"<s>[INST] <<SYS>>\nYou are a helpful, respectful and honest assistant. Always answer as helpfully as possible, while being safe. Your answers should not include any harmful, unethical, racist, sexist, toxic, dangerous, or illegal content. Please ensure that your responses are socially unbiased and positive in nature.\n\nIf a question does not make any sense, or is not factually coherent, explain why instead of answering something not correct. If you don't know the answer to a question, please don't share false information.\n<</SYS>>\n\n{user_input} [/INST]"
|
72 |
|
73 |
inputs = tokenizer(prompt, return_tensors="pt", add_special_tokens=False).input_ids.to(model.device)
|
|
|
50 |
## Domain-Specific LLaMA-2-Chat
|
51 |
Our method is also effective for aligned models! LLaMA-2-Chat requires a [specific data format](https://huggingface.co/blog/llama2#how-to-prompt-llama-2), and our **reading comprehension can perfectly fit the data format** by transforming the reading comprehension into a multi-turn conversation. We have also open-sourced chat models in different domains: [Biomedicine-Chat](https://huggingface.co/AdaptLLM/medicine-chat), [Finance-Chat](https://huggingface.co/AdaptLLM/finance-chat) and [Law-Chat](https://huggingface.co/AdaptLLM/law-chat)
|
52 |
|
53 |
+
For example, to chat with the law-chat model:
|
54 |
```python
|
55 |
from transformers import AutoModelForCausalLM, AutoTokenizer
|
56 |
|
|
|
67 |
|
68 |
Please provide your choice first and then provide explanations if possible.'''
|
69 |
|
70 |
+
# We use the prompt template of LLaMA-2-Chat demo for chat models (NOTE: NO prompt template is required for base models!)
|
71 |
prompt = f"<s>[INST] <<SYS>>\nYou are a helpful, respectful and honest assistant. Always answer as helpfully as possible, while being safe. Your answers should not include any harmful, unethical, racist, sexist, toxic, dangerous, or illegal content. Please ensure that your responses are socially unbiased and positive in nature.\n\nIf a question does not make any sense, or is not factually coherent, explain why instead of answering something not correct. If you don't know the answer to a question, please don't share false information.\n<</SYS>>\n\n{user_input} [/INST]"
|
72 |
|
73 |
inputs = tokenizer(prompt, return_tensors="pt", add_special_tokens=False).input_ids.to(model.device)
|