From 19944818724c98d2f06a10f619acf622ba0876ad Mon Sep 17 00:00:00 2001 From: Heiner Lohaus Date: Sun, 19 May 2024 05:09:55 +0200 Subject: Update model list / providers --- g4f/Provider/Cohere.py | 2 +- g4f/Provider/DuckDuckGo.py | 7 +- g4f/Provider/Llama.py | 2 +- g4f/Provider/PerplexityLabs.py | 7 +- g4f/Provider/__init__.py | 2 +- g4f/Provider/needs_auth/OpenaiChat.py | 3 +- g4f/models.py | 154 ++++++++++------------------------ g4f/providers/retry_provider.py | 41 ++++----- 8 files changed, 75 insertions(+), 143 deletions(-) diff --git a/g4f/Provider/Cohere.py b/g4f/Provider/Cohere.py index 4f9fd30a..eac04ab4 100644 --- a/g4f/Provider/Cohere.py +++ b/g4f/Provider/Cohere.py @@ -9,7 +9,7 @@ from .helper import format_prompt class Cohere(AbstractProvider): url = "https://cohereforai-c4ai-command-r-plus.hf.space" - working = True + working = False supports_gpt_35_turbo = False supports_gpt_4 = False supports_stream = True diff --git a/g4f/Provider/DuckDuckGo.py b/g4f/Provider/DuckDuckGo.py index 2fa0612a..9379660b 100644 --- a/g4f/Provider/DuckDuckGo.py +++ b/g4f/Provider/DuckDuckGo.py @@ -16,8 +16,11 @@ class DuckDuckGo(AsyncGeneratorProvider, ProviderModelMixin): supports_message_history = True default_model = "gpt-3.5-turbo-0125" - models = ["gpt-3.5-turbo-0125", "claude-instant-1.2"] - model_aliases = {"gpt-3.5-turbo": "gpt-3.5-turbo-0125"} + models = ["gpt-3.5-turbo-0125", "claude-3-haiku-20240307"] + model_aliases = { + "gpt-3.5-turbo": "gpt-3.5-turbo-0125", + "claude-3-haiku": "claude-3-haiku-20240307" + } status_url = "https://duckduckgo.com/duckchat/v1/status" chat_url = "https://duckduckgo.com/duckchat/v1/chat" diff --git a/g4f/Provider/Llama.py b/g4f/Provider/Llama.py index f2c78b36..235c0994 100644 --- a/g4f/Provider/Llama.py +++ b/g4f/Provider/Llama.py @@ -9,7 +9,7 @@ from .base_provider import AsyncGeneratorProvider, ProviderModelMixin class Llama(AsyncGeneratorProvider, ProviderModelMixin): url = "https://www.llama2.ai" - working = True + working = False supports_message_history = True default_model = "meta/meta-llama-3-70b-instruct" models = [ diff --git a/g4f/Provider/PerplexityLabs.py b/g4f/Provider/PerplexityLabs.py index 91ba63f2..4a2cc9e5 100644 --- a/g4f/Provider/PerplexityLabs.py +++ b/g4f/Provider/PerplexityLabs.py @@ -21,11 +21,14 @@ class PerplexityLabs(AsyncGeneratorProvider, ProviderModelMixin): "related" ] model_aliases = { - "mistralai/Mistral-7B-Instruct-v0.1": "mistral-7b-instruct", + "mistralai/Mistral-7B-Instruct-v0.1": "mistral-7b-instruct", + "mistralai/Mistral-7B-Instruct-v0.2": "mistral-7b-instruct", "mistralai/Mixtral-8x7B-Instruct-v0.1": "mixtral-8x7b-instruct", "codellama/CodeLlama-70b-Instruct-hf": "codellama-70b-instruct", "llava-v1.5-7b": "llava-v1.5-7b-wrapper", - 'databricks/dbrx-instruct': "dbrx-instruct" + "databricks/dbrx-instruct": "dbrx-instruct", + "meta-llama/Meta-Llama-3-70B-Instruct": "llama-3-70b-instruct", + "meta-llama/Meta-Llama-3-8B-Instruct": "llama-3-8b-instruct" } @classmethod diff --git a/g4f/Provider/__init__.py b/g4f/Provider/__init__.py index d0c0d8b6..e60e1310 100644 --- a/g4f/Provider/__init__.py +++ b/g4f/Provider/__init__.py @@ -1,7 +1,7 @@ from __future__ import annotations from ..providers.types import BaseProvider, ProviderType -from ..providers.retry_provider import RetryProvider, IterProvider +from ..providers.retry_provider import RetryProvider, IterListProvider from ..providers.base_provider import AsyncProvider, AsyncGeneratorProvider from ..providers.create_images import CreateImagesProvider diff --git a/g4f/Provider/needs_auth/OpenaiChat.py b/g4f/Provider/needs_auth/OpenaiChat.py index d8ea4fad..f40ae961 100644 --- a/g4f/Provider/needs_auth/OpenaiChat.py +++ b/g4f/Provider/needs_auth/OpenaiChat.py @@ -373,6 +373,7 @@ class OpenaiChat(AsyncGeneratorProvider, ProviderModelMixin): except NoValidHarFileError as e: if cls._api_key is None and cls.needs_auth: raise e + cls._create_request_args() if cls.default_model is None: cls.default_model = cls.get_model(await cls.get_default_model(session, cls._headers)) @@ -420,7 +421,7 @@ class OpenaiChat(AsyncGeneratorProvider, ProviderModelMixin): **requirements["proofofwork"], user_agent=cls._headers["user-agent"], proofTokens=proofTokens - ) + ) if debug.logging: print( 'Arkose:', False if not need_arkose else arkose_token[:12]+"...", diff --git a/g4f/models.py b/g4f/models.py index 0c5eaa9f..40de22ba 100644 --- a/g4f/models.py +++ b/g4f/models.py @@ -2,27 +2,27 @@ from __future__ import annotations from dataclasses import dataclass -from .Provider import RetryProvider, ProviderType +from .Provider import IterListProvider, ProviderType from .Provider import ( Aichatos, Bing, Blackbox, - Chatgpt4Online, ChatgptAi, ChatgptNext, - Cohere, Cnote, DeepInfra, + DuckDuckGo, + Ecosia, Feedough, FreeGpt, Gemini, - GeminiProChat, + GeminiPro, GigaChat, HuggingChat, HuggingFace, Koala, Liaobots, - Llama, + MetaAI, OpenaiChat, PerplexityLabs, Replicate, @@ -32,7 +32,6 @@ from .Provider import ( Reka ) - @dataclass(unsafe_hash=True) class Model: """ @@ -55,12 +54,12 @@ class Model: default = Model( name = "", base_provider = "", - best_provider = RetryProvider([ + best_provider = IterListProvider([ Bing, ChatgptAi, You, - Chatgpt4Online, - OpenaiChat + OpenaiChat, + Ecosia, ]) ) @@ -68,11 +67,14 @@ default = Model( gpt_35_long = Model( name = 'gpt-3.5-turbo', base_provider = 'openai', - best_provider = RetryProvider([ + best_provider = IterListProvider([ FreeGpt, You, ChatgptNext, OpenaiChat, + Koala, + Ecosia, + DuckDuckGo, ]) ) @@ -80,7 +82,7 @@ gpt_35_long = Model( gpt_35_turbo = Model( name = 'gpt-3.5-turbo', base_provider = 'openai', - best_provider = RetryProvider([ + best_provider = IterListProvider([ FreeGpt, You, ChatgptNext, @@ -95,7 +97,7 @@ gpt_35_turbo = Model( gpt_4 = Model( name = 'gpt-4', base_provider = 'openai', - best_provider = RetryProvider([ + best_provider = IterListProvider([ Bing, Liaobots, ]) ) @@ -103,8 +105,8 @@ gpt_4 = Model( gpt_4o = Model( name = 'gpt-4o', base_provider = 'openai', - best_provider = RetryProvider([ - You + best_provider = IterListProvider([ + You, Liaobots ]) ) @@ -120,46 +122,22 @@ gigachat = Model( best_provider = GigaChat ) -gigachat_plus = Model( - name = 'GigaChat-Plus', - base_provider = 'gigachat', - best_provider = GigaChat -) - -gigachat_pro = Model( - name = 'GigaChat-Pro', - base_provider = 'gigachat', - best_provider = GigaChat -) - -llama2_7b = Model( - name = "meta-llama/Llama-2-7b-chat-hf", - base_provider = 'meta', - best_provider = RetryProvider([Llama, DeepInfra]) -) - -llama2_13b = Model( - name = "meta-llama/Llama-2-13b-chat-hf", - base_provider = 'meta', - best_provider = RetryProvider([Llama, DeepInfra]) -) - -llama2_70b = Model( - name = "meta-llama/Llama-2-70b-chat-hf", +meta = Model( + name = "meta", base_provider = "meta", - best_provider = RetryProvider([Llama, DeepInfra]) + best_provider = MetaAI ) llama3_8b_instruct = Model( name = "meta-llama/Meta-Llama-3-8B-Instruct", base_provider = "meta", - best_provider = RetryProvider([Llama, DeepInfra, Replicate]) + best_provider = IterListProvider([DeepInfra, PerplexityLabs, Replicate]) ) llama3_70b_instruct = Model( name = "meta-llama/Meta-Llama-3-70B-Instruct", base_provider = "meta", - best_provider = RetryProvider([Llama, DeepInfra]) + best_provider = IterListProvider([DeepInfra, PerplexityLabs, Replicate]) ) codellama_34b_instruct = Model( @@ -171,61 +149,30 @@ codellama_34b_instruct = Model( codellama_70b_instruct = Model( name = "codellama/CodeLlama-70b-Instruct-hf", base_provider = "meta", - best_provider = RetryProvider([DeepInfra, PerplexityLabs]) + best_provider = IterListProvider([DeepInfra, PerplexityLabs]) ) # Mistral mixtral_8x7b = Model( name = "mistralai/Mixtral-8x7B-Instruct-v0.1", base_provider = "huggingface", - best_provider = RetryProvider([DeepInfra, HuggingFace, PerplexityLabs]) + best_provider = IterListProvider([DeepInfra, HuggingFace, PerplexityLabs]) ) mistral_7b = Model( name = "mistralai/Mistral-7B-Instruct-v0.1", base_provider = "huggingface", - best_provider = RetryProvider([HuggingChat, HuggingFace, PerplexityLabs]) + best_provider = IterListProvider([HuggingChat, HuggingFace, PerplexityLabs]) ) mistral_7b_v02 = Model( name = "mistralai/Mistral-7B-Instruct-v0.2", base_provider = "huggingface", - best_provider = DeepInfra -) - -mixtral_8x22b = Model( - name = "HuggingFaceH4/zephyr-orpo-141b-A35b-v0.1", - base_provider = "huggingface", - best_provider = DeepInfra -) - -# Misc models -dolphin_mixtral_8x7b = Model( - name = "cognitivecomputations/dolphin-2.6-mixtral-8x7b", - base_provider = "huggingface", - best_provider = DeepInfra -) - -lzlv_70b = Model( - name = "lizpreciatior/lzlv_70b_fp16_hf", - base_provider = "huggingface", - best_provider = DeepInfra -) - -airoboros_70b = Model( - name = "deepinfra/airoboros-70b", - base_provider = "huggingface", - best_provider = DeepInfra -) - -openchat_35 = Model( - name = "openchat/openchat_3.5", - base_provider = "huggingface", - best_provider = DeepInfra + best_provider = IterListProvider([DeepInfra, HuggingFace, PerplexityLabs]) ) # Bard -gemini = bard = palm = Model( +gemini = Model( name = 'gemini', base_provider = 'google', best_provider = Gemini @@ -234,7 +181,7 @@ gemini = bard = palm = Model( claude_v2 = Model( name = 'claude-v2', base_provider = 'anthropic', - best_provider = RetryProvider([Vercel]) + best_provider = IterListProvider([Vercel]) ) claude_3_opus = Model( @@ -249,6 +196,12 @@ claude_3_sonnet = Model( best_provider = You ) +claude_3_haiku = Model( + name = 'claude-3-haiku', + base_provider = 'anthropic', + best_provider = DuckDuckGo +) + gpt_35_turbo_16k = Model( name = 'gpt-3.5-turbo-16k', base_provider = 'openai', @@ -288,7 +241,7 @@ gpt_4_32k_0613 = Model( gemini_pro = Model( name = 'gemini-pro', base_provider = 'google', - best_provider = RetryProvider([GeminiProChat, You]) + best_provider = IterListProvider([GeminiPro, You]) ) pi = Model( @@ -300,13 +253,13 @@ pi = Model( dbrx_instruct = Model( name = 'databricks/dbrx-instruct', base_provider = 'mistral', - best_provider = RetryProvider([DeepInfra, PerplexityLabs]) + best_provider = IterListProvider([DeepInfra, PerplexityLabs]) ) command_r_plus = Model( name = 'CohereForAI/c4ai-command-r-plus', base_provider = 'mistral', - best_provider = RetryProvider([HuggingChat, Cohere]) + best_provider = IterListProvider([HuggingChat]) ) blackbox = Model( @@ -334,9 +287,8 @@ class ModelUtils: 'gpt-3.5-turbo-0613' : gpt_35_turbo_0613, 'gpt-3.5-turbo-16k' : gpt_35_turbo_16k, 'gpt-3.5-turbo-16k-0613' : gpt_35_turbo_16k_0613, - 'gpt-3.5-long': gpt_35_long, - + # gpt-4 'gpt-4o' : gpt_4o, 'gpt-4' : gpt_4, @@ -345,52 +297,38 @@ class ModelUtils: 'gpt-4-32k-0613' : gpt_4_32k_0613, 'gpt-4-turbo' : gpt_4_turbo, - # Llama - 'llama2-7b' : llama2_7b, - 'llama2-13b': llama2_13b, - 'llama2-70b': llama2_70b, - - 'llama3-8b' : llama3_8b_instruct, # alias + "meta-ai": meta, + 'llama3-8b': llama3_8b_instruct, # alias 'llama3-70b': llama3_70b_instruct, # alias 'llama3-8b-instruct' : llama3_8b_instruct, 'llama3-70b-instruct': llama3_70b_instruct, - + 'codellama-34b-instruct': codellama_34b_instruct, 'codellama-70b-instruct': codellama_70b_instruct, - # GigaChat - 'gigachat' : gigachat, - 'gigachat_plus': gigachat_plus, - 'gigachat_pro' : gigachat_pro, - # Mistral Opensource 'mixtral-8x7b': mixtral_8x7b, 'mistral-7b': mistral_7b, 'mistral-7b-v02': mistral_7b_v02, - 'mixtral-8x22b': mixtral_8x22b, - 'dolphin-mixtral-8x7b': dolphin_mixtral_8x7b, - + # google gemini 'gemini': gemini, 'gemini-pro': gemini_pro, - + # anthropic 'claude-v2': claude_v2, 'claude-3-opus': claude_3_opus, 'claude-3-sonnet': claude_3_sonnet, - + 'claude-3-haiku': claude_3_haiku, + # reka core - 'reka-core': reka_core, 'reka': reka_core, - 'Reka Core': reka_core, - + # other 'blackbox': blackbox, 'command-r+': command_r_plus, 'dbrx-instruct': dbrx_instruct, - 'lzlv-70b': lzlv_70b, - 'airoboros-70b': airoboros_70b, - 'openchat_3.5': openchat_35, + 'gigachat': gigachat, 'pi': pi } diff --git a/g4f/providers/retry_provider.py b/g4f/providers/retry_provider.py index cde8c848..0061bcc1 100644 --- a/g4f/providers/retry_provider.py +++ b/g4f/providers/retry_provider.py @@ -4,11 +4,11 @@ import asyncio import random from ..typing import Type, List, CreateResult, Messages, Iterator, AsyncResult -from .types import BaseProvider, BaseRetryProvider +from .types import BaseProvider, BaseRetryProvider, ProviderType from .. import debug from ..errors import RetryProviderError, RetryNoProviderError -class NewBaseRetryProvider(BaseRetryProvider): +class IterListProvider(BaseRetryProvider): def __init__( self, providers: List[Type[BaseProvider]], @@ -45,21 +45,17 @@ class NewBaseRetryProvider(BaseRetryProvider): Raises: Exception: Any exception encountered during the completion process. """ - providers = [p for p in self.providers if stream and p.supports_stream] if stream else self.providers - if self.shuffle: - random.shuffle(providers) - exceptions = {} started: bool = False - for provider in providers: + for provider in self.get_providers(stream): self.last_provider = provider try: if debug.logging: print(f"Using {provider.__name__} provider") for token in provider.create_completion(model, messages, stream, **kwargs): yield token - started = True + started = True if started: return except Exception as e: @@ -87,13 +83,9 @@ class NewBaseRetryProvider(BaseRetryProvider): Raises: Exception: Any exception encountered during the asynchronous completion process. """ - providers = self.providers - if self.shuffle: - random.shuffle(providers) - exceptions = {} - for provider in providers: + for provider in self.get_providers(False): self.last_provider = provider try: if debug.logging: @@ -109,8 +101,8 @@ class NewBaseRetryProvider(BaseRetryProvider): raise_exceptions(exceptions) - def get_providers(self, stream: bool): - providers = [p for p in self.providers if stream and p.supports_stream] if stream else self.providers + def get_providers(self, stream: bool) -> list[ProviderType]: + providers = [p for p in self.providers if p.supports_stream] if stream else self.providers if self.shuffle: random.shuffle(providers) return providers @@ -138,7 +130,7 @@ class NewBaseRetryProvider(BaseRetryProvider): else: for token in provider.create_completion(model, messages, stream, **kwargs): yield token - started = True + started = True if started: return except Exception as e: @@ -150,7 +142,7 @@ class NewBaseRetryProvider(BaseRetryProvider): raise_exceptions(exceptions) -class RetryProvider(NewBaseRetryProvider): +class RetryProvider(IterListProvider): def __init__( self, providers: List[Type[BaseProvider]], @@ -188,11 +180,10 @@ class RetryProvider(NewBaseRetryProvider): Raises: Exception: Any exception encountered during the completion process. """ - providers = self.get_providers(stream) - if self.single_provider_retry and len(providers) == 1: + if self.single_provider_retry: exceptions = {} started: bool = False - provider = providers[0] + provider = self.providers[0] self.last_provider = provider for attempt in range(self.max_retries): try: @@ -200,7 +191,7 @@ class RetryProvider(NewBaseRetryProvider): print(f"Using {provider.__name__} provider (attempt {attempt + 1})") for token in provider.create_completion(model, messages, stream, **kwargs): yield token - started = True + started = True if started: return except Exception as e: @@ -229,14 +220,10 @@ class RetryProvider(NewBaseRetryProvider): Raises: Exception: Any exception encountered during the asynchronous completion process. """ - providers = self.providers - if self.shuffle: - random.shuffle(providers) - exceptions = {} - if self.single_provider_retry and len(providers) == 1: - provider = providers[0] + if self.single_provider_retry: + provider = self.providers[0] self.last_provider = provider for attempt in range(self.max_retries): try: -- cgit v1.2.3