diff options
Diffstat (limited to 'g4f/models.py')
| -rw-r--r-- | g4f/models.py | 381 |
1 files changed, 274 insertions, 107 deletions
diff --git a/g4f/models.py b/g4f/models.py index 78e1d74e..e9016561 100644 --- a/g4f/models.py +++ b/g4f/models.py @@ -4,30 +4,34 @@ from dataclasses import dataclass from .Provider import IterListProvider, ProviderType from .Provider import ( - Aichatos, - Bing, - Blackbox, - ChatgptAi, - ChatgptNext, - Cnote, - DeepInfra, - Feedough, - FreeGpt, - Gemini, - GeminiPro, - GigaChat, - HuggingChat, - HuggingFace, - Koala, - Liaobots, - MetaAI, - OpenaiChat, - PerplexityLabs, - Replicate, - Pi, - Vercel, - You, - Reka + AI365VIP, + Bing, + Blackbox, + Chatgpt4o, + ChatgptFree, + DDG, + DeepInfra, + DeepInfraImage, + FreeChatgpt, + FreeGpt, + Gemini, + GeminiPro, + GeminiProChat, + GigaChat, + HuggingChat, + HuggingFace, + Koala, + Liaobots, + MetaAI, + OpenaiChat, + PerplexityLabs, + Pi, + Pizzagpt, + Reka, + Replicate, + ReplicateHome, + Vercel, + You, ) @dataclass(unsafe_hash=True) @@ -54,9 +58,15 @@ default = Model( base_provider = "", best_provider = IterListProvider([ Bing, - ChatgptAi, You, OpenaiChat, + FreeChatgpt, + AI365VIP, + Chatgpt4o, + DDG, + ChatgptFree, + Koala, + Pizzagpt, ]) ) @@ -67,28 +77,58 @@ gpt_35_long = Model( best_provider = IterListProvider([ FreeGpt, You, - ChatgptNext, OpenaiChat, Koala, + ChatgptFree, + FreeChatgpt, + DDG, + AI365VIP, + Pizzagpt, ]) ) -# GPT-3.5 / GPT-4 +############ +### Text ### +############ + +### OpenAI ### +### GPT-3.5 / GPT-4 ### +# gpt-3.5 gpt_35_turbo = Model( name = 'gpt-3.5-turbo', base_provider = 'openai', best_provider = IterListProvider([ FreeGpt, You, - ChatgptNext, Koala, OpenaiChat, - Aichatos, - Cnote, - Feedough, + ChatgptFree, + FreeChatgpt, + DDG, + AI365VIP, + Pizzagpt, ]) ) +gpt_35_turbo_16k = Model( + name = 'gpt-3.5-turbo-16k', + base_provider = 'openai', + best_provider = gpt_35_long.best_provider +) + +gpt_35_turbo_16k_0613 = Model( + name = 'gpt-3.5-turbo-16k-0613', + base_provider = 'openai', + best_provider = gpt_35_long.best_provider +) + +gpt_35_turbo_0613 = Model( + name = 'gpt-3.5-turbo-0613', + base_provider = 'openai', + best_provider = gpt_35_turbo.best_provider +) + +# gpt-4 gpt_4 = Model( name = 'gpt-4', base_provider = 'openai', @@ -97,12 +137,22 @@ gpt_4 = Model( ]) ) -gpt_4o = Model( - name = 'gpt-4o', +gpt_4_0613 = Model( + name = 'gpt-4-0613', base_provider = 'openai', - best_provider = IterListProvider([ - You, Liaobots - ]) + best_provider = gpt_4.best_provider +) + +gpt_4_32k = Model( + name = 'gpt-4-32k', + base_provider = 'openai', + best_provider = gpt_4.best_provider +) + +gpt_4_32k_0613 = Model( + name = 'gpt-4-32k-0613', + base_provider = 'openai', + best_provider = gpt_4.best_provider ) gpt_4_turbo = Model( @@ -111,18 +161,36 @@ gpt_4_turbo = Model( best_provider = Bing ) +gpt_4o = Model( + name = 'gpt-4o', + base_provider = 'openai', + best_provider = IterListProvider([ + You, Liaobots, Chatgpt4o, AI365VIP + ]) +) + + +### GigaChat ### gigachat = Model( name = 'GigaChat:latest', base_provider = 'gigachat', best_provider = GigaChat ) + +### Meta ### meta = Model( name = "meta", base_provider = "meta", best_provider = MetaAI ) +llama_2_70b_chat = Model( + name = "meta/llama-2-70b-chat", + base_provider = "meta", + best_provider = IterListProvider([ReplicateHome]) +) + llama3_8b_instruct = Model( name = "meta-llama/Meta-Llama-3-8B-Instruct", base_provider = "meta", @@ -132,7 +200,7 @@ llama3_8b_instruct = Model( llama3_70b_instruct = Model( name = "meta-llama/Meta-Llama-3-70B-Instruct", base_provider = "meta", - best_provider = IterListProvider([DeepInfra, PerplexityLabs, Replicate]) + best_provider = IterListProvider([DeepInfra, PerplexityLabs, Replicate, HuggingChat, DDG]) ) codellama_34b_instruct = Model( @@ -144,35 +212,77 @@ codellama_34b_instruct = Model( codellama_70b_instruct = Model( name = "codellama/CodeLlama-70b-Instruct-hf", base_provider = "meta", - best_provider = IterListProvider([DeepInfra, PerplexityLabs]) + best_provider = IterListProvider([DeepInfra]) ) -# Mistral + +### Mistral ### mixtral_8x7b = Model( name = "mistralai/Mixtral-8x7B-Instruct-v0.1", base_provider = "huggingface", - best_provider = IterListProvider([DeepInfra, HuggingFace, PerplexityLabs]) -) - -mistral_7b = Model( - name = "mistralai/Mistral-7B-Instruct-v0.1", - base_provider = "huggingface", - best_provider = IterListProvider([HuggingChat, HuggingFace, PerplexityLabs]) + best_provider = IterListProvider([DeepInfra, HuggingFace, PerplexityLabs, HuggingChat, DDG]) ) mistral_7b_v02 = Model( name = "mistralai/Mistral-7B-Instruct-v0.2", base_provider = "huggingface", - best_provider = IterListProvider([DeepInfra, HuggingFace, PerplexityLabs]) + best_provider = IterListProvider([DeepInfra, HuggingFace, HuggingChat, ReplicateHome]) +) + + +### NousResearch ### +Nous_Hermes_2_Mixtral_8x7B_DPO = Model( + name = "NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO", + base_provider = "NousResearch", + best_provider = IterListProvider([HuggingFace, HuggingChat]) +) + + +### 01-ai ### +Yi_1_5_34B_Chat = Model( + name = "01-ai/Yi-1.5-34B-Chat", + base_provider = "01-ai", + best_provider = IterListProvider([HuggingFace, HuggingChat]) +) + + +### Microsoft ### +Phi_3_mini_4k_instruct = Model( + name = "microsoft/Phi-3-mini-4k-instruct", + base_provider = "Microsoft", + best_provider = IterListProvider([HuggingFace, HuggingChat]) ) -# Bard + +### Google ### +# gemini gemini = Model( name = 'gemini', - base_provider = 'google', + base_provider = 'Google', best_provider = Gemini ) +gemini_pro = Model( + name = 'gemini-pro', + base_provider = 'Google', + best_provider = IterListProvider([GeminiPro, You, GeminiProChat]) +) + +# gemma +gemma_2_9b_it = Model( + name = 'gemma-2-9b-it', + base_provider = 'Google', + best_provider = IterListProvider([PerplexityLabs]) +) + +gemma_2_27b_it = Model( + name = 'gemma-2-27b-it', + base_provider = 'Google', + best_provider = IterListProvider([PerplexityLabs]) +) + + +### Anthropic ### claude_v2 = Model( name = 'claude-v2', base_provider = 'anthropic', @@ -194,79 +304,76 @@ claude_3_sonnet = Model( claude_3_haiku = Model( name = 'claude-3-haiku', base_provider = 'anthropic', - best_provider = None + best_provider = IterListProvider([DDG, AI365VIP]) ) -gpt_35_turbo_16k = Model( - name = 'gpt-3.5-turbo-16k', - base_provider = 'openai', - best_provider = gpt_35_long.best_provider -) -gpt_35_turbo_16k_0613 = Model( - name = 'gpt-3.5-turbo-16k-0613', - base_provider = 'openai', - best_provider = gpt_35_long.best_provider +### Reka AI ### +reka_core = Model( + name = 'reka-core', + base_provider = 'Reka AI', + best_provider = Reka ) -gpt_35_turbo_0613 = Model( - name = 'gpt-3.5-turbo-0613', - base_provider = 'openai', - best_provider = gpt_35_turbo.best_provider -) -gpt_4_0613 = Model( - name = 'gpt-4-0613', - base_provider = 'openai', - best_provider = gpt_4.best_provider +### NVIDIA ### +nemotron_4_340b_instruct = Model( + name = 'nemotron-4-340b-instruct', + base_provider = 'NVIDIA', + best_provider = IterListProvider([PerplexityLabs]) ) -gpt_4_32k = Model( - name = 'gpt-4-32k', - base_provider = 'openai', - best_provider = gpt_4.best_provider + +### Blackbox ### +blackbox = Model( + name = 'blackbox', + base_provider = 'Blackbox', + best_provider = Blackbox ) -gpt_4_32k_0613 = Model( - name = 'gpt-4-32k-0613', - base_provider = 'openai', - best_provider = gpt_4.best_provider + +### Databricks ### +dbrx_instruct = Model( + name = 'databricks/dbrx-instruct', + base_provider = 'Databricks', + best_provider = IterListProvider([DeepInfra]) ) -gemini_pro = Model( - name = 'gemini-pro', - base_provider = 'google', - best_provider = IterListProvider([GeminiPro, You]) + +### CohereForAI ### +command_r_plus = Model( + name = 'CohereForAI/c4ai-command-r-plus', + base_provider = 'CohereForAI', + best_provider = IterListProvider([HuggingChat]) ) + +### Other ### pi = Model( name = 'pi', base_provider = 'inflection', best_provider = Pi ) -dbrx_instruct = Model( - name = 'databricks/dbrx-instruct', - base_provider = 'mistral', - best_provider = IterListProvider([DeepInfra, PerplexityLabs]) -) -command_r_plus = Model( - name = 'CohereForAI/c4ai-command-r-plus', - base_provider = 'mistral', - best_provider = IterListProvider([HuggingChat]) -) +############# +### Image ### +############# -blackbox = Model( - name = 'blackbox', - base_provider = 'blackbox', - best_provider = Blackbox +### Stability AI ### +sdxl = Model( + name = 'stability-ai/sdxl', + base_provider = 'Stability AI', + best_provider = IterListProvider([ReplicateHome, DeepInfraImage]) + ) -reka_core = Model( - name = 'reka-core', - base_provider = 'Reka AI', - best_provider = Reka +### AI Forever ### +kandinsky_2_2 = Model( + name = 'ai-forever/kandinsky-2.2', + base_provider = 'AI Forever', + best_provider = IterListProvider([ReplicateHome]) + ) class ModelUtils: @@ -277,6 +384,13 @@ class ModelUtils: convert (dict[str, Model]): Dictionary mapping model string identifiers to Model instances. """ convert: dict[str, Model] = { + + ############ + ### Text ### + ############ + + ### OpenAI ### + ### GPT-3.5 / GPT-4 ### # gpt-3.5 'gpt-3.5-turbo' : gpt_35_turbo, 'gpt-3.5-turbo-0613' : gpt_35_turbo_0613, @@ -291,8 +405,12 @@ class ModelUtils: 'gpt-4-32k' : gpt_4_32k, 'gpt-4-32k-0613' : gpt_4_32k_0613, 'gpt-4-turbo' : gpt_4_turbo, - + + + ### Meta ### "meta-ai": meta, + + 'llama-2-70b-chat': llama_2_70b_chat, 'llama3-8b': llama3_8b_instruct, # alias 'llama3-70b': llama3_70b_instruct, # alias 'llama3-8b-instruct' : llama3_8b_instruct, @@ -301,30 +419,79 @@ class ModelUtils: 'codellama-34b-instruct': codellama_34b_instruct, 'codellama-70b-instruct': codellama_70b_instruct, - # Mistral Opensource + + ### Mistral (Opensource) ### 'mixtral-8x7b': mixtral_8x7b, - 'mistral-7b': mistral_7b, 'mistral-7b-v02': mistral_7b_v02, + + + ### NousResearch ### + 'Nous-Hermes-2-Mixtral-8x7B-DPO': Nous_Hermes_2_Mixtral_8x7B_DPO, - # google gemini + + ### 01-ai ### + 'Yi-1.5-34B-Chat': Yi_1_5_34B_Chat, + + + ### Microsoft ### + 'Phi-3-mini-4k-instruct': Phi_3_mini_4k_instruct, + + + ### Google ### + # gemini 'gemini': gemini, 'gemini-pro': gemini_pro, + + # gemma + 'gemma-2-9b-it': gemma_2_9b_it, + 'gemma-2-27b-it': gemma_2_27b_it, + - # anthropic + ### 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 AI ### 'reka': reka_core, - # other - 'blackbox': blackbox, + + ### NVIDIA ### + 'nemotron-4-340b-instruct': nemotron_4_340b_instruct, + + + ### Blackbox ### + 'blackbox': blackbox, + + + ### CohereForAI ### 'command-r+': command_r_plus, + + + ### Databricks ### 'dbrx-instruct': dbrx_instruct, + + + ### GigaChat ### 'gigachat': gigachat, - 'pi': pi + + + # Other + 'pi': pi, + + + + ############# + ### Image ### + ############# + + ### Stability AI ### + 'sdxl': sdxl, + + ### AI Forever ### + 'kandinsky-2.2': kandinsky_2_2, } _all_models = list(ModelUtils.convert.keys()) |