Update README.md
Browse files
README.md
CHANGED
@@ -4,17 +4,20 @@ language:
|
|
4 |
- zh
|
5 |
---
|
6 |
A instruction-tuned LoRA model of https://huggingface.co/baichuan-inc/baichuan-7B
|
|
|
7 |
Training framework: https://github.com/hiyouga/LLaMA-Factory
|
|
|
8 |
Please follow the baichuan-7B License to use this model.
|
9 |
|
10 |
Usage:
|
|
|
11 |
from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer
|
12 |
|
13 |
tokenizer = AutoTokenizer.from_pretrained("hiyouga/baichuan-7b-sft", trust_remote_code=True)
|
14 |
model = AutoModelForCausalLM.from_pretrained("hiyouga/baichuan-7b-sft", trust_remote_code=True).cuda()
|
15 |
streamer = TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
|
16 |
|
17 |
-
query = "
|
18 |
template = (
|
19 |
"你是一名经验丰富的心理咨询师,专长于认知行为疗法, 以心理咨询师的身份回答以下问题。\n"
|
20 |
"Human: {}\nAssistant: "
|
@@ -23,3 +26,4 @@ template = (
|
|
23 |
inputs = tokenizer([template.format(query)], return_tensors="pt")
|
24 |
inputs = inputs.to("cuda")
|
25 |
generate_ids = model.generate(**inputs, max_new_tokens=256, streamer=streamer)
|
|
|
|
4 |
- zh
|
5 |
---
|
6 |
A instruction-tuned LoRA model of https://huggingface.co/baichuan-inc/baichuan-7B
|
7 |
+
|
8 |
Training framework: https://github.com/hiyouga/LLaMA-Factory
|
9 |
+
|
10 |
Please follow the baichuan-7B License to use this model.
|
11 |
|
12 |
Usage:
|
13 |
+
```python
|
14 |
from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer
|
15 |
|
16 |
tokenizer = AutoTokenizer.from_pretrained("hiyouga/baichuan-7b-sft", trust_remote_code=True)
|
17 |
model = AutoModelForCausalLM.from_pretrained("hiyouga/baichuan-7b-sft", trust_remote_code=True).cuda()
|
18 |
streamer = TextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
|
19 |
|
20 |
+
query = "为什么生怕一点点事情做不好被人批评?做事情,别人告诉了我方法,但身体不会按方法来,非要折腾几遍,才发现别人告诉的方法和口诀,是最高效的;而且最近一个月,睡眠不好,整个白天都是无精打采的,每天活的很丧,知道自己的问题出在哪里,不晓得怎么去做出改变"
|
21 |
template = (
|
22 |
"你是一名经验丰富的心理咨询师,专长于认知行为疗法, 以心理咨询师的身份回答以下问题。\n"
|
23 |
"Human: {}\nAssistant: "
|
|
|
26 |
inputs = tokenizer([template.format(query)], return_tensors="pt")
|
27 |
inputs = inputs.to("cuda")
|
28 |
generate_ids = model.generate(**inputs, max_new_tokens=256, streamer=streamer)
|
29 |
+
```
|