add model config : gpt_35_turbo_0613

This commit is contained in:
tony 2023-09-20 11:16:42 +08:00
parent 4772b830d1
commit fa80db85ca

View File

@ -1,226 +1,229 @@
from __future__ import annotations from __future__ import annotations
from dataclasses import dataclass from dataclasses import dataclass
from .typing import Union from .typing import Union
from .Provider import BaseProvider from .Provider import BaseProvider
from .Provider import ( from .Provider import (
ChatgptLogin, ChatgptLogin,
CodeLinkAva, CodeLinkAva,
ChatgptAi, ChatgptAi,
ChatBase, ChatBase,
Vercel, Vercel,
DeepAi, DeepAi,
Aivvm, Aivvm,
Bard, Bard,
H2o H2o,
Yqcloud
) )
@dataclass @dataclass
class Model: class Model:
name: str name: str
base_provider: str base_provider: str
best_provider: Union[type[BaseProvider], tuple[type[BaseProvider]]] = None best_provider: Union[type[BaseProvider], tuple[type[BaseProvider]]] = None
# Config for HuggingChat, OpenAssistant # Config for HuggingChat, OpenAssistant
# Works for Liaobots, H2o, OpenaiChat, Yqcloud, You # Works for Liaobots, H2o, OpenaiChat, Yqcloud, You
default = Model( default = Model(
name = "", name="",
base_provider = "huggingface") base_provider="huggingface")
# GPT-3.5 / GPT-4 # GPT-3.5 / GPT-4
gpt_35_turbo = Model( gpt_35_turbo = Model(
name = 'gpt-3.5-turbo', name='gpt-3.5-turbo',
base_provider = 'openai', base_provider='openai',
best_provider = [ best_provider=[
DeepAi, CodeLinkAva, ChatgptLogin, ChatgptAi, ChatBase, Aivvm DeepAi, CodeLinkAva, ChatgptLogin, ChatgptAi, ChatBase, Aivvm
] ]
) )
gpt_4 = Model( gpt_4 = Model(
name = 'gpt-4', name='gpt-4',
base_provider = 'openai') base_provider='openai')
# Bard # Bard
palm = Model( palm = Model(
name = 'palm', name='palm',
base_provider = 'google', base_provider='google',
best_provider = Bard) best_provider=Bard)
# H2o # H2o
falcon_7b = Model( falcon_7b = Model(
name = 'h2oai/h2ogpt-gm-oasst1-en-2048-falcon-7b-v3', name='h2oai/h2ogpt-gm-oasst1-en-2048-falcon-7b-v3',
base_provider = 'huggingface', base_provider='huggingface',
best_provider = H2o) best_provider=H2o)
falcon_40b = Model( falcon_40b = Model(
name = 'h2oai/h2ogpt-gm-oasst1-en-2048-falcon-40b-v1', name='h2oai/h2ogpt-gm-oasst1-en-2048-falcon-40b-v1',
base_provider = 'huggingface', base_provider='huggingface',
best_provider = H2o) best_provider=H2o)
llama_13b = Model( llama_13b = Model(
name = 'h2oai/h2ogpt-gm-oasst1-en-2048-open-llama-13b', name='h2oai/h2ogpt-gm-oasst1-en-2048-open-llama-13b',
base_provider = 'huggingface', base_provider='huggingface',
best_provider = H2o) best_provider=H2o)
# Vercel # Vercel
claude_instant_v1 = Model( claude_instant_v1 = Model(
name = 'anthropic:claude-instant-v1', name='anthropic:claude-instant-v1',
base_provider = 'anthropic', base_provider='anthropic',
best_provider = Vercel) best_provider=Vercel)
claude_v1 = Model( claude_v1 = Model(
name = 'anthropic:claude-v1', name='anthropic:claude-v1',
base_provider = 'anthropic', base_provider='anthropic',
best_provider = Vercel) best_provider=Vercel)
claude_v2 = Model( claude_v2 = Model(
name = 'anthropic:claude-v2', name='anthropic:claude-v2',
base_provider = 'anthropic', base_provider='anthropic',
best_provider = Vercel) best_provider=Vercel)
command_light_nightly = Model( command_light_nightly = Model(
name = 'cohere:command-light-nightly', name='cohere:command-light-nightly',
base_provider = 'cohere', base_provider='cohere',
best_provider = Vercel) best_provider=Vercel)
command_nightly = Model( command_nightly = Model(
name = 'cohere:command-nightly', name='cohere:command-nightly',
base_provider = 'cohere', base_provider='cohere',
best_provider = Vercel) best_provider=Vercel)
gpt_neox_20b = Model( gpt_neox_20b = Model(
name = 'huggingface:EleutherAI/gpt-neox-20b', name='huggingface:EleutherAI/gpt-neox-20b',
base_provider = 'huggingface', base_provider='huggingface',
best_provider = Vercel) best_provider=Vercel)
oasst_sft_1_pythia_12b = Model( oasst_sft_1_pythia_12b = Model(
name = 'huggingface:OpenAssistant/oasst-sft-1-pythia-12b', name='huggingface:OpenAssistant/oasst-sft-1-pythia-12b',
base_provider = 'huggingface', base_provider='huggingface',
best_provider = Vercel) best_provider=Vercel)
oasst_sft_4_pythia_12b_epoch_35 = Model( oasst_sft_4_pythia_12b_epoch_35 = Model(
name = 'huggingface:OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5', name='huggingface:OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5',
base_provider = 'huggingface', base_provider='huggingface',
best_provider = Vercel) best_provider=Vercel)
santacoder = Model( santacoder = Model(
name = 'huggingface:bigcode/santacoder', name='huggingface:bigcode/santacoder',
base_provider = 'huggingface', base_provider='huggingface',
best_provider = Vercel) best_provider=Vercel)
bloom = Model( bloom = Model(
name = 'huggingface:bigscience/bloom', name='huggingface:bigscience/bloom',
base_provider = 'huggingface', base_provider='huggingface',
best_provider = Vercel) best_provider=Vercel)
flan_t5_xxl = Model( flan_t5_xxl = Model(
name = 'huggingface:google/flan-t5-xxl', name='huggingface:google/flan-t5-xxl',
base_provider = 'huggingface', base_provider='huggingface',
best_provider = Vercel) best_provider=Vercel)
code_davinci_002 = Model( code_davinci_002 = Model(
name = 'openai:code-davinci-002', name='openai:code-davinci-002',
base_provider = 'openai', base_provider='openai',
best_provider = Vercel) best_provider=Vercel)
gpt_35_turbo_16k = Model( gpt_35_turbo_16k = Model(
name = 'openai:gpt-3.5-turbo-16k', name='openai:gpt-3.5-turbo-16k',
base_provider = 'openai', base_provider='openai',
best_provider = Vercel) best_provider=Vercel)
gpt_35_turbo_0613 = Model( gpt_35_turbo_0613 = Model(
name = 'openai:gpt-3.5-turbo-0613', name='openai:gpt-3.5-turbo-0613',
base_provider = 'openai', base_provider='openai',
best_provider = [ best_provider=[
DeepAi, CodeLinkAva, ChatgptLogin, ChatgptAi, ChatBase, Aivvm Yqcloud, Aivvm,ChatgptLogin
]) ])
gpt_35_turbo_16k_0613 = Model( gpt_35_turbo_16k_0613 = Model(
name = 'openai:gpt-3.5-turbo-16k-0613', name='openai:gpt-3.5-turbo-16k-0613',
base_provider = 'openai') base_provider='openai')
gpt_4_0613 = Model( gpt_4_0613 = Model(
name = 'openai:gpt-4-0613', name='openai:gpt-4-0613',
base_provider = 'openai', base_provider='openai',
best_provider = Vercel) best_provider=Vercel)
text_ada_001 = Model( text_ada_001 = Model(
name = 'openai:text-ada-001', name='openai:text-ada-001',
base_provider = 'openai', base_provider='openai',
best_provider = Vercel) best_provider=Vercel)
text_babbage_001 = Model( text_babbage_001 = Model(
name = 'openai:text-babbage-001', name='openai:text-babbage-001',
base_provider = 'openai', base_provider='openai',
best_provider = Vercel) best_provider=Vercel)
text_curie_001 = Model( text_curie_001 = Model(
name = 'openai:text-curie-001', name='openai:text-curie-001',
base_provider = 'openai', base_provider='openai',
best_provider = Vercel) best_provider=Vercel)
text_davinci_002 = Model( text_davinci_002 = Model(
name = 'openai:text-davinci-002', name='openai:text-davinci-002',
base_provider = 'openai', base_provider='openai',
best_provider = Vercel) best_provider=Vercel)
text_davinci_003 = Model( text_davinci_003 = Model(
name = 'openai:text-davinci-003', name='openai:text-davinci-003',
base_provider = 'openai', base_provider='openai',
best_provider = Vercel) best_provider=Vercel)
llama13b_v2_chat = Model( llama13b_v2_chat = Model(
name = 'replicate:a16z-infra/llama13b-v2-chat', name='replicate:a16z-infra/llama13b-v2-chat',
base_provider = 'replicate', base_provider='replicate',
best_provider = Vercel) best_provider=Vercel)
llama7b_v2_chat = Model( llama7b_v2_chat = Model(
name = 'replicate:a16z-infra/llama7b-v2-chat', name='replicate:a16z-infra/llama7b-v2-chat',
base_provider = 'replicate', base_provider='replicate',
best_provider = Vercel) best_provider=Vercel)
class ModelUtils: class ModelUtils:
convert: dict[str, Model] = { convert: dict[str, Model] = {
# GPT-3.5 / GPT-4 # GPT-3.5 / GPT-4
'gpt-3.5-turbo' : gpt_35_turbo, 'gpt-3.5-turbo': gpt_35_turbo,
'gpt-4' : gpt_4, 'gpt-4': gpt_4,
# Bard # Bard
'palm2' : palm, 'palm2': palm,
'palm' : palm, 'palm': palm,
'google' : palm, 'google': palm,
'google-bard' : palm, 'google-bard': palm,
'google-palm' : palm, 'google-palm': palm,
'bard' : palm, 'bard': palm,
# H2o # H2o
'falcon-40b' : falcon_40b, 'falcon-40b': falcon_40b,
'falcon-7b' : falcon_7b, 'falcon-7b': falcon_7b,
'llama-13b' : llama_13b, 'llama-13b': llama_13b,
# Vercel # Vercel
'claude-instant-v1' : claude_instant_v1, 'claude-instant-v1': claude_instant_v1,
'claude-v1' : claude_v1, 'claude-v1': claude_v1,
'claude-v2' : claude_v2, 'claude-v2': claude_v2,
'command-nightly' : command_nightly, 'command-nightly': command_nightly,
'gpt-neox-20b' : gpt_neox_20b, 'gpt-neox-20b': gpt_neox_20b,
'santacoder' : santacoder, 'santacoder': santacoder,
'bloom' : bloom, 'bloom': bloom,ChatgptLogin
'flan-t5-xxl' : flan_t5_xxl, 'flan-t5-xxl': flan_t5_xxl,
'code-davinci-002' : code_davinci_002, 'code-davinci-002': code_davinci_002,
'gpt-3.5-turbo-16k' : gpt_35_turbo_16k, 'gpt-3.5-turbo-16k': gpt_35_turbo_16k,
'gpt-4-0613' : gpt_4_0613, 'gpt-4-0613': gpt_4_0613,
'text-ada-001' : text_ada_001, 'text-ada-001': text_ada_001,
'text-babbage-001' : text_babbage_001, 'text-babbage-001': text_babbage_001,
'text-curie-001' : text_curie_001, 'text-curie-001': text_curie_001,
'text-davinci-002' : text_davinci_002, 'text-davinci-002': text_davinci_002,
'text-davinci-003' : text_davinci_003, 'text-davinci-003': text_davinci_003,
'llama13b-v2-chat' : llama13b_v2_chat, 'llama13b-v2-chat': llama13b_v2_chat,
'llama7b-v2-chat' : llama7b_v2_chat, 'llama7b-v2-chat': llama7b_v2_chat,
'oasst-sft-1-pythia-12b' : oasst_sft_1_pythia_12b, '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, 'oasst-sft-4-pythia-12b-epoch-3.5': oasst_sft_4_pythia_12b_epoch_35,
'command-light-nightly' : command_light_nightly, 'command-light-nightly': command_light_nightly,
'gpt-3.5-turbo-16k-0613' : gpt_35_turbo_16k_0613, 'gpt-3.5-turbo-16k-0613': gpt_35_turbo_16k_0613,
'gpt-3.5-turbo-0613' : gpt_35_turbo_0613, 'gpt-3.5-turbo-0613': gpt_35_turbo_0613,
} }