fix path
Browse files
README.md
CHANGED
@@ -247,12 +247,14 @@ pip install protobuf==3.20.0 transformers==4.27.1 icetk cpm_kernels
|
|
247 |
|
248 |
## 代码调用
|
249 |
|
250 |
-
|
|
|
|
|
251 |
|
252 |
```ipython
|
253 |
>>> from transformers import AutoTokenizer, AutoModel
|
254 |
-
>>> tokenizer = AutoTokenizer.from_pretrained("
|
255 |
-
>>> model = AutoModel.from_pretrained("
|
256 |
>>> response, history = model.chat(tokenizer, "你好", history=[])
|
257 |
>>> print(response)
|
258 |
你好,有什么我可以帮您解答的问题吗?
|
|
|
247 |
|
248 |
## 代码调用
|
249 |
|
250 |
+
夫子•明察基于 ChatGLM 进行司法垂直领域增量训练,所以其使用方法与 ChatGLM 相同。
|
251 |
+
|
252 |
+
将权重下载完毕后,可以通过如下代码调用夫子•明察模型来生成对话:(将 "/path/to/fuzi.mingcha-v1.0" 改为权重的放置路径)
|
253 |
|
254 |
```ipython
|
255 |
>>> from transformers import AutoTokenizer, AutoModel
|
256 |
+
>>> tokenizer = AutoTokenizer.from_pretrained("/path/to/fuzi.mingcha-v1.0", trust_remote_code=True)
|
257 |
+
>>> model = AutoModel.from_pretrained("/path/to/fuzi.mingcha-v1.0", trust_remote_code=True).half().cuda()
|
258 |
>>> response, history = model.chat(tokenizer, "你好", history=[])
|
259 |
>>> print(response)
|
260 |
你好,有什么我可以帮您解答的问题吗?
|