-
Notifications
You must be signed in to change notification settings - Fork 4
/
constants.py
110 lines (104 loc) · 3.36 KB
/
constants.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
INFERENCE_OUTPUT = "inference_outputs"
EVALUATION_OUTPUT = "evaluation_outputs"
RESULT_OUTPUT = "results"
MODELS = [
"GPT-3.5-Turbo-16K",
"Llama-2-chat-7B",
"Llama-2-chat-13B",
"Llama-2-chat-70B",
"Vicuna-13B-16K-v1.5",
"ChatGLM2-6B-32K",
"ChatGLM3-6B-32K",
"MistralLite-7B-32K",
"Zephyr-7b-beta",
]
DATASET_MAP = {
"mnds-news": "MN-DS",
"mnds-news_autoregressive": "MN-DS (wo gold)",
# "mnds-news-retrieval_autoregressive": "MN-DS Retrieval (wo gold)",
"newsqa": "NewsQA",
"newsqa_autoregressive": "NewsQA (wo gold)",
"tedtalks": "TED-talks",
"news_commentary": "News-Commentary",
"cnnnews": "CNN",
"multi-task-single-inst_qa": "Multi-Task (QA)",
# "multi-task_qa": "QA",
"multi-task-single-inst_sum": "Multi-Task (SUM)",
# "multi-task_sum": "SUM",
"multi-task-single-inst_trans": "Multi-Task (TRANS)",
# "multi-task_trans": "TRANS",
"multi-task-single-inst_ner": "Multi-Task (NER)",
# "multi-task_ner": "NER",
"multi-task-single-inst_pos": "Multi-Task (POS)",
# "multi-task_pos": "POS",
"multi-task-single-inst_cls": "Multi-Task (CLS)",
# "multi-task_cls": "CLS",
"multi-task-single-inst_rel": "Multi-Task (REL)",
# "multi-task_rel": "REL",
"multi-task-multi-inst-multi-turn": "Multi-Task Multi-Inst Multi-Turn",
"multi-task-multi-inst-single-turn": "Multi-Task Multi-Inst Single-Turn",
"multi-task-single-inst-multi-turn": "Multi-Task Single-Inst Multi-Turn",
"multi-task-single-inst-single-turn": "Multi-Task Single-Inst Single-Turn",
"cyrus-qa": "QA",
"cyrus-qa_autoregressive": "QA (wo gold)",
"cyrus-cls": "CLS",
"cyrus-cls_autoregressive": "CLS (wo gold)",
}
MODEL_MAP = {
# "gpt-3.5-turbo-16k": "GPT-3.5-Turbo-16k",
"gpt-3.5-turbo": "GPT-3.5-Turbo",
# "chatglm2-6b-32k": "ChatGLM2-6B-32K",
"chatglm3-6b-32k": "ChatGLM3-6B-32K",
# "vicuna-13b-16k": "Vicuna-13B-16K-v1.5",
"vicuna-7b": "Vicuna-7B-v1.5",
"vicuna-13b": "Vicuna-13B-v1.5",
"llama2-chat-7b": "Llama-2-chat-7B",
"llama2-chat-13b": "Llama-2-chat-13B",
"llama2-chat-70b": "Llama-2-chat-70B",
# "mistrallite": "MistralLite-7B-32K",
# "zephyr-7b-beta": "Zephyr-7B",
# "openchat-3.5": "OpenChat-3.5-7B",
# "openhermes-2.5": "OpenHermes-2.5-7B",
# "starling-7b": "Starling-7B",
"qwen-chat-7b": "Qwen-chat-7B",
"qwen-chat-14b": "Qwen-chat-14B",
"qwen-chat-72b": "Qwen-chat-72B",
}
END_TOKENS = {
"chatglm3-6b-32k": ["<|user|>"],
"chatglm2-6b-32k": ["问："],
"llama2-chat-7b": [
"</s>",
],
"llama2-chat-13b": ["</s>"],
"llama2-chat-70b": ["</s>"],
"vicuna-13b-16k": ["</s>", "USER:"],
"vicuna-7b": ["</s>", "USER:"],
"vicuna-13b": ["</s>", "USER:"],
"gpt-3.5-turbo-16k": [],
"mistrallite": ["</s>", "<|prompter|>"],
"zephyr-7b-beta": ["</s>", "<|user|>"],
"openchat-3.5": ["<|end_of_turn|>", "GPT4 Correct User"],
"openhermes-2.5": ["<|im_end|>", "user\n"],
"starling-7b": ["<|end_of_turn|>", "GPT4 Correct User"],
"qwen-chat-7b": ["<|im_end|>"],
"qwen-chat-14b": ["<|im_end|>"],
"qwen-chat-72b": ["<|im_end|>"],
"gpt-3.5-turbo": [],
"gpt-4": [],
}
topics = [
"ethics",
"sports",
"music",
"art",
"science",
"literature",
"economics",
"medicine",
"food",
"technology",
"history",
"travel",
"education",
]