Update README.md
Browse files
README.md
CHANGED
@@ -9,4 +9,30 @@ datasets:
|
|
9 |
library_name: transformers
|
10 |
---
|
11 |
|
12 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
9 |
library_name: transformers
|
10 |
---
|
11 |
|
12 |
+
## Examples
|
13 |
+
```python
|
14 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
15 |
+
|
16 |
+
model_path = 'chinoll/Yi-6b-200k-dpo'
|
17 |
+
|
18 |
+
tokenizer = AutoTokenizer.from_pretrained(model_path, use_fast=False)
|
19 |
+
|
20 |
+
# Since transformers 4.35.0, the GPT-Q/AWQ model can be loaded using AutoModelForCausalLM.
|
21 |
+
model = AutoModelForCausalLM.from_pretrained(
|
22 |
+
model_path,
|
23 |
+
device_map="auto",
|
24 |
+
torch_dtype='auto'
|
25 |
+
).eval()
|
26 |
+
|
27 |
+
# Prompt content: "hi"
|
28 |
+
messages = [
|
29 |
+
{"role": "user", "content": "hi"}
|
30 |
+
]
|
31 |
+
|
32 |
+
input_ids = tokenizer.apply_chat_template(conversation=messages, tokenize=True, add_generation_prompt=True, return_tensors='pt')
|
33 |
+
output_ids = model.generate(input_ids.to('cuda'))
|
34 |
+
response = tokenizer.decode(output_ids[0][input_ids.shape[1]:], skip_special_tokens=True)
|
35 |
+
|
36 |
+
# Model response: "Hello! How can I assist you today?"
|
37 |
+
print(response)
|
38 |
+
```
|