summaryrefslogtreecommitdiffstats
path: root/g4f/models.py
blob: e61ebbcd06944b01e2fab835c78293c2b934c43c (plain) (blame)
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
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
from dataclasses import dataclass
from .Provider import Bard, BaseProvider, GetGpt, H2o, Liaobots, Vercel, Equing

@dataclass
class Model:
    name: str
    base_provider: str
    best_provider: type[BaseProvider]


# GPT-3.5 / GPT-4
gpt_35_turbo = Model(
    name          = 'gpt-3.5-turbo',
    base_provider = 'openai',
    best_provider = GetGpt)

gpt_4 = Model(
    name          = 'gpt-4',
    base_provider = 'openai',
    best_provider = Liaobots)

# Bard
palm = Model(
    name          = 'palm',
    base_provider = 'google',
    best_provider = Bard)

# H2o
falcon_7b = Model(
    name          = 'h2oai/h2ogpt-gm-oasst1-en-2048-falcon-7b-v3',
    base_provider = 'huggingface',
    best_provider = H2o)

falcon_40b = Model(
    name          = 'h2oai/h2ogpt-gm-oasst1-en-2048-falcon-40b-v1',
    base_provider = 'huggingface',
    best_provider = H2o)

llama_13b = Model(
    name          = 'h2oai/h2ogpt-gm-oasst1-en-2048-open-llama-13b',
    base_provider = 'huggingface',
    best_provider = H2o)

# Vercel
claude_instant_v1 = Model(
    name          = 'anthropic:claude-instant-v1',
    base_provider = 'anthropic',
    best_provider = Vercel)

claude_v1 = Model(
    name          = 'anthropic:claude-v1',
    base_provider = 'anthropic',
    best_provider = Vercel)

claude_v2 = Model(
    name          = 'anthropic:claude-v2',
    base_provider = 'anthropic',
    best_provider = Vercel)

command_light_nightly = Model(
    name          = 'cohere:command-light-nightly',
    base_provider = 'cohere',
    best_provider = Vercel)

command_nightly = Model(
    name          = 'cohere:command-nightly',
    base_provider = 'cohere',
    best_provider = Vercel)

gpt_neox_20b = Model(
    name          = 'huggingface:EleutherAI/gpt-neox-20b',
    base_provider = 'huggingface',
    best_provider = Vercel)

oasst_sft_1_pythia_12b = Model(
    name          = 'huggingface:OpenAssistant/oasst-sft-1-pythia-12b',
    base_provider = 'huggingface',
    best_provider = Vercel)

oasst_sft_4_pythia_12b_epoch_35 = Model(
    name          = 'huggingface:OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5',
    base_provider = 'huggingface',
    best_provider = Vercel)

santacoder = Model(
    name          = 'huggingface:bigcode/santacoder',
    base_provider = 'huggingface',
    best_provider = Vercel)

bloom = Model(
    name          = 'huggingface:bigscience/bloom',
    base_provider = 'huggingface',
    best_provider = Vercel)

flan_t5_xxl = Model(
    name          = 'huggingface:google/flan-t5-xxl',
    base_provider = 'huggingface',
    best_provider = Vercel)

code_davinci_002 = Model(
    name          = 'openai:code-davinci-002',
    base_provider = 'openai',
    best_provider = Vercel)

gpt_35_turbo_16k = Model(
    name          = 'openai:gpt-3.5-turbo-16k',
    base_provider = 'openai',
    best_provider = Vercel)

gpt_35_turbo_16k_0613 = Model(
    name          = 'openai:gpt-3.5-turbo-16k-0613',
    base_provider = 'openai',
    best_provider = Equing)

gpt_4_0613 = Model(
    name          = 'openai:gpt-4-0613',
    base_provider = 'openai',
    best_provider = Vercel)

text_ada_001 = Model(
    name          = 'openai:text-ada-001',
    base_provider = 'openai',
    best_provider = Vercel)

text_babbage_001 = Model(
    name          = 'openai:text-babbage-001',
    base_provider = 'openai',
    best_provider = Vercel)

text_curie_001 = Model(
    name          = 'openai:text-curie-001',
    base_provider = 'openai',
    best_provider = Vercel)

text_davinci_002 = Model(
    name          = 'openai:text-davinci-002',
    base_provider = 'openai',
    best_provider = Vercel)

text_davinci_003 = Model(
    name          = 'openai:text-davinci-003',
    base_provider = 'openai',
    best_provider = Vercel)

llama13b_v2_chat = Model(
    name          = 'replicate:a16z-infra/llama13b-v2-chat',
    base_provider = 'replicate',
    best_provider = Vercel)

llama7b_v2_chat = Model(
    name          = 'replicate:a16z-infra/llama7b-v2-chat',
    base_provider = 'replicate',
    best_provider = Vercel)


class ModelUtils:
    convert: dict[str, Model] = {
        # GPT-3.5 / GPT-4
        'gpt-3.5-turbo' : gpt_35_turbo,
        'gpt-4'         : gpt_4,
        
        # Bard
        'palm2'       : palm,
        'palm'        : palm,
        'google'      : palm,
        'google-bard' : palm,
        'google-palm' : palm,
        'bard'        : palm,
        
        # H2o
        'falcon-40b' : falcon_40b,
        'falcon-7b'  : falcon_7b,
        'llama-13b'  : llama_13b,
        
        # Vercel
        'claude-instant-v1' : claude_instant_v1,
        'claude-v1'         : claude_v1,
        'claude-v2'         : claude_v2,
        'command-nightly'   : command_nightly,
        'gpt-neox-20b'      : gpt_neox_20b,
        'santacoder'        : santacoder,
        'bloom'             : bloom,
        'flan-t5-xxl'       : flan_t5_xxl,
        'code-davinci-002'  : code_davinci_002,
        'gpt-3.5-turbo-16k' : gpt_35_turbo_16k,
        'gpt-4-0613'        : gpt_4_0613,
        'text-ada-001'      : text_ada_001,
        'text-babbage-001'  : text_babbage_001,
        'text-curie-001'    : text_curie_001,
        'text-davinci-002'  : text_davinci_002,
        'text-davinci-003'  : text_davinci_003,
        'llama13b-v2-chat'  : llama13b_v2_chat,
        'llama7b-v2-chat'   : llama7b_v2_chat,
        
        'oasst-sft-1-pythia-12b'           : oasst_sft_1_pythia_12b,
        'oasst-sft-4-pythia-12b-epoch-3.5' : oasst_sft_4_pythia_12b_epoch_35,
        'command-light-nightly'            : command_light_nightly,
        'gpt-3.5-turbo-16k-0613'           : gpt_35_turbo_16k_0613,
    }