Spaces:
Sleeping
Sleeping
johnsmith253325
commited on
Commit
·
4cae7cc
1
Parent(s):
cebe276
feat: 支持ChatGLM3 resolve #941
Browse files- modules/models/ChatGLM.py +24 -1
- modules/models/base_model.py +5 -0
- modules/presets.py +2 -0
modules/models/ChatGLM.py
CHANGED
@@ -4,6 +4,8 @@ import logging
|
|
4 |
import os
|
5 |
import platform
|
6 |
|
|
|
|
|
7 |
import colorama
|
8 |
|
9 |
from ..index_func import *
|
@@ -18,6 +20,7 @@ class ChatGLM_Client(BaseLLMModel):
|
|
18 |
import torch
|
19 |
from transformers import AutoModel, AutoTokenizer
|
20 |
global CHATGLM_TOKENIZER, CHATGLM_MODEL
|
|
|
21 |
if CHATGLM_TOKENIZER is None or CHATGLM_MODEL is None:
|
22 |
system_name = platform.system()
|
23 |
model_path = None
|
@@ -53,7 +56,12 @@ class ChatGLM_Client(BaseLLMModel):
|
|
53 |
model = model.eval()
|
54 |
CHATGLM_MODEL = model
|
55 |
|
56 |
-
def
|
|
|
|
|
|
|
|
|
|
|
57 |
history = [x["content"] for x in self.history]
|
58 |
query = history.pop()
|
59 |
logging.debug(colorama.Fore.YELLOW +
|
@@ -65,6 +73,12 @@ class ChatGLM_Client(BaseLLMModel):
|
|
65 |
for i in range(0, len(history), 2)]
|
66 |
return history, query
|
67 |
|
|
|
|
|
|
|
|
|
|
|
|
|
68 |
def get_answer_at_once(self):
|
69 |
history, query = self._get_glm_style_input()
|
70 |
response, _ = CHATGLM_MODEL.chat(
|
@@ -82,3 +96,12 @@ class ChatGLM_Client(BaseLLMModel):
|
|
82 |
temperature=self.temperature,
|
83 |
):
|
84 |
yield response
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4 |
import os
|
5 |
import platform
|
6 |
|
7 |
+
import gc
|
8 |
+
import torch
|
9 |
import colorama
|
10 |
|
11 |
from ..index_func import *
|
|
|
20 |
import torch
|
21 |
from transformers import AutoModel, AutoTokenizer
|
22 |
global CHATGLM_TOKENIZER, CHATGLM_MODEL
|
23 |
+
self.deinitialize()
|
24 |
if CHATGLM_TOKENIZER is None or CHATGLM_MODEL is None:
|
25 |
system_name = platform.system()
|
26 |
model_path = None
|
|
|
56 |
model = model.eval()
|
57 |
CHATGLM_MODEL = model
|
58 |
|
59 |
+
def _get_glm3_style_input(self):
|
60 |
+
history = self.history
|
61 |
+
query = history.pop()["content"]
|
62 |
+
return history, query
|
63 |
+
|
64 |
+
def _get_glm2_style_input(self):
|
65 |
history = [x["content"] for x in self.history]
|
66 |
query = history.pop()
|
67 |
logging.debug(colorama.Fore.YELLOW +
|
|
|
73 |
for i in range(0, len(history), 2)]
|
74 |
return history, query
|
75 |
|
76 |
+
def _get_glm_style_input(self):
|
77 |
+
if "glm2" in self.model_name:
|
78 |
+
return self._get_glm2_style_input()
|
79 |
+
else:
|
80 |
+
return self._get_glm3_style_input()
|
81 |
+
|
82 |
def get_answer_at_once(self):
|
83 |
history, query = self._get_glm_style_input()
|
84 |
response, _ = CHATGLM_MODEL.chat(
|
|
|
96 |
temperature=self.temperature,
|
97 |
):
|
98 |
yield response
|
99 |
+
|
100 |
+
def deinitialize(self):
|
101 |
+
# 释放显存
|
102 |
+
global CHATGLM_MODEL, CHATGLM_TOKENIZER
|
103 |
+
CHATGLM_MODEL = None
|
104 |
+
CHATGLM_TOKENIZER = None
|
105 |
+
gc.collect()
|
106 |
+
torch.cuda.empty_cache()
|
107 |
+
logging.info("ChatGLM model deinitialized")
|
modules/models/base_model.py
CHANGED
@@ -847,6 +847,11 @@ class BaseLLMModel:
|
|
847 |
"""
|
848 |
return gr.update()
|
849 |
|
|
|
|
|
|
|
|
|
|
|
850 |
|
851 |
class Base_Chat_Langchain_Client(BaseLLMModel):
|
852 |
def __init__(self, model_name, user_name=""):
|
|
|
847 |
"""
|
848 |
return gr.update()
|
849 |
|
850 |
+
def deinitialize(self):
|
851 |
+
"""deinitialize the model, implement if needed
|
852 |
+
"""
|
853 |
+
pass
|
854 |
+
|
855 |
|
856 |
class Base_Chat_Langchain_Client(BaseLLMModel):
|
857 |
def __init__(self, model_name, user_name=""):
|
modules/presets.py
CHANGED
@@ -87,6 +87,8 @@ LOCAL_MODELS = [
|
|
87 |
"chatglm-6b-int4-ge",
|
88 |
"chatglm2-6b",
|
89 |
"chatglm2-6b-int4",
|
|
|
|
|
90 |
"StableLM",
|
91 |
"MOSS",
|
92 |
"Llama-2-7B-Chat",
|
|
|
87 |
"chatglm-6b-int4-ge",
|
88 |
"chatglm2-6b",
|
89 |
"chatglm2-6b-int4",
|
90 |
+
"chatglm3-6b",
|
91 |
+
"chatglm3-6b-32k",
|
92 |
"StableLM",
|
93 |
"MOSS",
|
94 |
"Llama-2-7B-Chat",
|