Spaces:
Paused
Paused
Your Name
commited on
Commit
•
40d91e9
1
Parent(s):
bf44dd1
improve dockerfile
Browse files- docs/Dockerfile+ChatGLM +3 -0
- request_llm/bridge_all.py +8 -5
docs/Dockerfile+ChatGLM
CHANGED
@@ -36,6 +36,9 @@ from transformers import AutoModel, AutoTokenizer \n\
|
|
36 |
chatglm_tokenizer = AutoTokenizer.from_pretrained("THUDM/chatglm-6b", trust_remote_code=True) \n\
|
37 |
chatglm_model = AutoModel.from_pretrained("THUDM/chatglm-6b", trust_remote_code=True).float() ' >> warm_up_chatglm.py
|
38 |
RUN python3 -u warm_up_chatglm.py
|
|
|
|
|
|
|
39 |
RUN $useProxyNetwork git pull
|
40 |
|
41 |
# 为chatgpt-academic配置代理和API-KEY (非必要 可选步骤)
|
|
|
36 |
chatglm_tokenizer = AutoTokenizer.from_pretrained("THUDM/chatglm-6b", trust_remote_code=True) \n\
|
37 |
chatglm_model = AutoModel.from_pretrained("THUDM/chatglm-6b", trust_remote_code=True).float() ' >> warm_up_chatglm.py
|
38 |
RUN python3 -u warm_up_chatglm.py
|
39 |
+
|
40 |
+
# 禁用缓存,确保更新代码
|
41 |
+
ADD "https://www.random.org/cgi-bin/randbyte?nbytes=10&format=h" skipcache
|
42 |
RUN $useProxyNetwork git pull
|
43 |
|
44 |
# 为chatgpt-academic配置代理和API-KEY (非必要 可选步骤)
|
request_llm/bridge_all.py
CHANGED
@@ -23,6 +23,9 @@ from .bridge_tgui import predict as tgui_ui
|
|
23 |
|
24 |
colors = ['#FF00FF', '#00FFFF', '#FF0000', '#990099', '#009999', '#990044']
|
25 |
|
|
|
|
|
|
|
26 |
model_info = {
|
27 |
# openai
|
28 |
"gpt-3.5-turbo": {
|
@@ -31,7 +34,7 @@ model_info = {
|
|
31 |
"endpoint": "https://api.openai.com/v1/chat/completions",
|
32 |
"max_token": 4096,
|
33 |
"tokenizer": tiktoken.encoding_for_model("gpt-3.5-turbo"),
|
34 |
-
"token_cnt":
|
35 |
},
|
36 |
|
37 |
"gpt-4": {
|
@@ -40,7 +43,7 @@ model_info = {
|
|
40 |
"endpoint": "https://api.openai.com/v1/chat/completions",
|
41 |
"max_token": 8192,
|
42 |
"tokenizer": tiktoken.encoding_for_model("gpt-4"),
|
43 |
-
"token_cnt":
|
44 |
},
|
45 |
|
46 |
# api_2d
|
@@ -50,7 +53,7 @@ model_info = {
|
|
50 |
"endpoint": "https://openai.api2d.net/v1/chat/completions",
|
51 |
"max_token": 4096,
|
52 |
"tokenizer": tiktoken.encoding_for_model("gpt-3.5-turbo"),
|
53 |
-
"token_cnt":
|
54 |
},
|
55 |
|
56 |
"api2d-gpt-4": {
|
@@ -59,7 +62,7 @@ model_info = {
|
|
59 |
"endpoint": "https://openai.api2d.net/v1/chat/completions",
|
60 |
"max_token": 8192,
|
61 |
"tokenizer": tiktoken.encoding_for_model("gpt-4"),
|
62 |
-
"token_cnt":
|
63 |
},
|
64 |
|
65 |
# chatglm
|
@@ -69,7 +72,7 @@ model_info = {
|
|
69 |
"endpoint": None,
|
70 |
"max_token": 1024,
|
71 |
"tokenizer": tiktoken.encoding_for_model("gpt-3.5-turbo"),
|
72 |
-
"token_cnt":
|
73 |
},
|
74 |
|
75 |
}
|
|
|
23 |
|
24 |
colors = ['#FF00FF', '#00FFFF', '#FF0000', '#990099', '#009999', '#990044']
|
25 |
|
26 |
+
get_token_num_gpt35 = lambda txt: len(tiktoken.encoding_for_model("gpt-3.5-turbo").encode(txt, disallowed_special=()))
|
27 |
+
get_token_num_gpt4 = lambda txt: len(tiktoken.encoding_for_model("gpt-4").encode(txt, disallowed_special=()))
|
28 |
+
|
29 |
model_info = {
|
30 |
# openai
|
31 |
"gpt-3.5-turbo": {
|
|
|
34 |
"endpoint": "https://api.openai.com/v1/chat/completions",
|
35 |
"max_token": 4096,
|
36 |
"tokenizer": tiktoken.encoding_for_model("gpt-3.5-turbo"),
|
37 |
+
"token_cnt": get_token_num_gpt35,
|
38 |
},
|
39 |
|
40 |
"gpt-4": {
|
|
|
43 |
"endpoint": "https://api.openai.com/v1/chat/completions",
|
44 |
"max_token": 8192,
|
45 |
"tokenizer": tiktoken.encoding_for_model("gpt-4"),
|
46 |
+
"token_cnt": get_token_num_gpt4,
|
47 |
},
|
48 |
|
49 |
# api_2d
|
|
|
53 |
"endpoint": "https://openai.api2d.net/v1/chat/completions",
|
54 |
"max_token": 4096,
|
55 |
"tokenizer": tiktoken.encoding_for_model("gpt-3.5-turbo"),
|
56 |
+
"token_cnt": get_token_num_gpt35,
|
57 |
},
|
58 |
|
59 |
"api2d-gpt-4": {
|
|
|
62 |
"endpoint": "https://openai.api2d.net/v1/chat/completions",
|
63 |
"max_token": 8192,
|
64 |
"tokenizer": tiktoken.encoding_for_model("gpt-4"),
|
65 |
+
"token_cnt": get_token_num_gpt4,
|
66 |
},
|
67 |
|
68 |
# chatglm
|
|
|
72 |
"endpoint": None,
|
73 |
"max_token": 1024,
|
74 |
"tokenizer": tiktoken.encoding_for_model("gpt-3.5-turbo"),
|
75 |
+
"token_cnt": get_token_num_gpt35,
|
76 |
},
|
77 |
|
78 |
}
|