WangZeJun commited on
Commit
6e19446
1 Parent(s): 40fdf78

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +25 -1
README.md CHANGED
@@ -1,4 +1,28 @@
1
  ---
2
  license: bigscience-bloom-rail-1.0
3
  ---
4
- https://github.com/zejunwang1/bloom_tuning
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: bigscience-bloom-rail-1.0
3
  ---
4
+ https://github.com/zejunwang1/bloom_tuning
5
+
6
+ 可以通过如下代码调用 bloom-396m-chat 模型来生成对话:
7
+ ```python
8
+ from transformers import BloomTokenizerFast, BloomForCausalLM
9
+
10
+ model_name_or_path = "WangZeJun/bloom-396m-chat"
11
+
12
+ tokenizer = BloomTokenizerFast.from_pretrained(model_name_or_path)
13
+ model = BloomForCausalLM.from_pretrained(model_name_or_path).cuda()
14
+ model = model.eval()
15
+
16
+ input_pattern = "{}</s>"
17
+ text = "你好"
18
+ input_ids = tokenizer(input_pattern.format(text), return_tensors="pt").input_ids
19
+ input_ids = input_ids.cuda()
20
+
21
+ outputs = model.generate(input_ids, do_sample=True, max_new_tokens=1024, top_p=0.85,
22
+ temperature=0.3, repetition_penalty=1.2, eos_token_id=tokenizer.eos_token_id)
23
+
24
+ output = tokenizer.decode(outputs[0])
25
+ response = output.replace(text, "").replace('</s>', "")
26
+ print(response)
27
+ ```
28
+