feat:llm support for chatglm2 (#271)

Add support for ChatGLM2-6B，which is the second-generation version of the open-source bilingual (Chinese-English) chat model [ChatGLM-6B](https://github.com/THUDM/ChatGLM-6B).
eosphoros-ai · Jun 25, 2023 · 8a54ae7 · 8a54ae7
2 parents 0558a8b + aa4115e
commit 8a54ae7
Show file tree

Hide file tree

Showing 8 changed files with 15 additions and 4 deletions.
diff --git a/README.md b/README.md
@@ -21,10 +21,11 @@ As large models are released and iterated upon, they are becoming increasingly i
 DB-GPT is an experimental open-source project that uses localized GPT large models to interact with your data and environment. With this solution, you can be assured that there is no risk of data leakage, and your data is 100% private and secure.
 
 ## News
+- [2023/06/25]🔥support chatglm2-6b model. [documents](https://db-gpt.readthedocs.io/en/latest/modules/llms.html) 
 - [2023/06/14] support gpt4all model, which can run at M1/M2, or cpu machine. [documents](https://db-gpt.readthedocs.io/en/latest/modules/llms.html) 
 - [2023/06/01]🔥 On the basis of the Vicuna-13B basic model, task chain calls are implemented through plugins. For example, the implementation of creating a database with a single sentence.[demo](./assets/auto_plugin.gif)
 - [2023/06/01]🔥 QLoRA guanaco(7b, 13b, 33b) support.
-- [2023/05/28]🔥 Learning from crawling data from the Internet [demo](./assets/chaturl_en.gif)
+- [2023/05/28] Learning from crawling data from the Internet [demo](./assets/chaturl_en.gif)
 - [2023/05/21] Generate SQL and execute it automatically. [demo](./assets/auto_sql_en.gif)
 - [2023/05/15] Chat with documents. [demo](./assets/new_knownledge_en.gif)
 - [2023/05/06] SQL generation and diagnosis. [demo](./assets/demo_en.gif)

diff --git a/README.zh.md b/README.zh.md
@@ -23,6 +23,7 @@ DB-GPT 是一个开源的以数据库为基础的GPT实验项目，使用本地
 
 
 ## 最新发布
+- [2023/06/25]🔥 支持ChatGLM2-6B模型。 [使用文档](https://db-gpt.readthedocs.io/projects/db-gpt-docs-zh-cn/zh_CN/latest/modules/llms.html)
 - [2023/06/14]🔥 支持gpt4all模型，可以在M1/M2 或者CPU机器上运行。 [使用文档](https://db-gpt.readthedocs.io/projects/db-gpt-docs-zh-cn/zh_CN/latest/modules/llms.html)
 - [2023/06/01]🔥 在Vicuna-13B基础模型的基础上，通过插件实现任务链调用。例如单句创建数据库的实现.
 - [2023/06/01]🔥 QLoRA guanaco(原驼)支持, 支持4090运行33B

diff --git a/docs/conf.py b/docs/conf.py
@@ -14,7 +14,7 @@
 copyright = "2023, csunny"
 author = "csunny"
 
-version = "👏👏 0.2.2"
+version = "👏👏 0.2.3"
 html_title = project + " " + version
 
 # -- General configuration ---------------------------------------------------

diff --git a/docs/getting_started/getting_started.md b/docs/getting_started/getting_started.md
@@ -38,6 +38,7 @@ Once the environment is installed, we have to create a new folder "models" in th
 git clone https://huggingface.co/Tribbiani/vicuna-13b 
 git clone https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2
 git clone https://huggingface.co/GanymedeNil/text2vec-large-chinese
+git clone https://huggingface.co/THUDM/chatglm2-6b
 ```
 
 The model files are large and will take a long time to download. During the download, let's configure the .env file, which needs to be copied and created from the .env.template

diff --git a/docs/locales/zh_CN/LC_MESSAGES/modules/llms.po b/docs/locales/zh_CN/LC_MESSAGES/modules/llms.po
@@ -6,7 +6,7 @@
 #, fuzzy
 msgid ""
 msgstr ""
-"Project-Id-Version: DB-GPT 0.1.0\n"
+"Project-Id-Version: DB-GPT 0.2.3\n"
 "Report-Msgid-Bugs-To: \n"
 "POT-Creation-Date: 2023-06-14 22:33+0800\n"
 "PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"

diff --git a/docs/modules/llms.md b/docs/modules/llms.md
@@ -17,6 +17,12 @@ if you want use other model, such as chatglm-6b, you just need update .env confi
 ```
 LLM_MODEL=chatglm-6b
 ```
+or chatglm2-6b, which  is the second-generation version of the open-source bilingual (Chinese-English) chat model ChatGLM-6B. 
+```
+LLM_MODEL=chatglm2-6b
+```
+
+
 
 ## Run Model with cpu.
 we alse support smaller models, like gpt4all.  you can use it with cpu/mps(M1/M2), Download from [gpt4all model](https://gpt4all.io/models/ggml-gpt4all-j-v1.3-groovy.bin)

diff --git a/pilot/configs/model_config.py b/pilot/configs/model_config.py
@@ -39,6 +39,8 @@
     "codet5p-2b": os.path.join(MODEL_PATH, "codet5p-2b"),
     "chatglm-6b-int4": os.path.join(MODEL_PATH, "chatglm-6b-int4"),
     "chatglm-6b": os.path.join(MODEL_PATH, "chatglm-6b"),
+    "chatglm2-6b": os.path.join(MODEL_PATH, "chatglm2-6b"),
+    "chatglm2-6b-int4": os.path.join(MODEL_PATH, "chatglm2-6b-int4"),
     "text2vec-base": os.path.join(MODEL_PATH, "text2vec-base-chinese"),
     "guanaco-33b-merged": os.path.join(MODEL_PATH, "guanaco-33b-merged"),
     "falcon-40b": os.path.join(MODEL_PATH, "falcon-40b"),

diff --git a/setup.py b/setup.py
@@ -19,7 +19,7 @@ def parse_requirements(file_name: str) -> List[str]:
 setuptools.setup(
     name="DB-GPT",
     packages=find_packages(),
-    version="0.2.2",
+    version="0.2.3",
     author="csunny",
     author_email="[email protected]",
     description="DB-GPT is an experimental open-source project that uses localized GPT large models to interact with your data and environment."