2024-05-13 11:46:42 -04:00
|
|
|
# frozen_string_literal: true
|
|
|
|
|
|
|
|
Fabricator(:llm_model) do
|
|
|
|
display_name "A good model"
|
|
|
|
name "gpt-4-turbo"
|
|
|
|
provider "open_ai"
|
2024-06-27 03:27:40 -04:00
|
|
|
tokenizer "DiscourseAi::Tokenizer::OpenAiTokenizer"
|
2024-07-30 12:44:57 -04:00
|
|
|
api_key "123"
|
|
|
|
url "https://api.openai.com/v1/chat/completions"
|
|
|
|
max_prompt_tokens 131_072
|
|
|
|
end
|
|
|
|
|
|
|
|
Fabricator(:anthropic_model, from: :llm_model) do
|
|
|
|
display_name "Claude 3 Opus"
|
|
|
|
name "claude-3-opus"
|
|
|
|
max_prompt_tokens 200_000
|
|
|
|
url "https://api.anthropic.com/v1/messages"
|
|
|
|
tokenizer "DiscourseAi::Tokenizer::AnthropicTokenizer"
|
|
|
|
provider "anthropic"
|
|
|
|
end
|
|
|
|
|
|
|
|
Fabricator(:hf_model, from: :llm_model) do
|
|
|
|
display_name "Llama 3.1"
|
|
|
|
name "meta-llama/Meta-Llama-3.1-70B-Instruct"
|
|
|
|
max_prompt_tokens 64_000
|
|
|
|
tokenizer "DiscourseAi::Tokenizer::Llama3Tokenizer"
|
|
|
|
url "https://test.dev/v1/chat/completions"
|
|
|
|
provider "hugging_face"
|
|
|
|
end
|
|
|
|
|
|
|
|
Fabricator(:vllm_model, from: :llm_model) do
|
|
|
|
display_name "Llama 3.1 vLLM"
|
|
|
|
name "meta-llama/Meta-Llama-3.1-70B-Instruct"
|
|
|
|
max_prompt_tokens 64_000
|
|
|
|
tokenizer "DiscourseAi::Tokenizer::Llama3Tokenizer"
|
|
|
|
url "https://test.dev/v1/chat/completions"
|
|
|
|
provider "vllm"
|
|
|
|
end
|
|
|
|
|
|
|
|
Fabricator(:fake_model, from: :llm_model) do
|
|
|
|
display_name "Fake model"
|
|
|
|
name "fake"
|
|
|
|
provider "fake"
|
|
|
|
tokenizer "DiscourseAi::Tokenizer::OpenAiTokenizer"
|
2024-05-13 11:46:42 -04:00
|
|
|
max_prompt_tokens 32_000
|
2024-08-06 13:35:35 -04:00
|
|
|
api_key "fake"
|
|
|
|
url "https://fake.test/"
|
2024-05-13 11:46:42 -04:00
|
|
|
end
|
2024-07-30 12:44:57 -04:00
|
|
|
|
|
|
|
Fabricator(:gemini_model, from: :llm_model) do
|
|
|
|
display_name "Gemini"
|
|
|
|
name "gemini-1.5-pro"
|
|
|
|
provider "google"
|
|
|
|
tokenizer "DiscourseAi::Tokenizer::OpenAiTokenizer"
|
|
|
|
max_prompt_tokens 800_000
|
|
|
|
url "https://generativelanguage.googleapis.com/v1beta/models/gemini-1.5-pro-latest"
|
|
|
|
end
|
|
|
|
|
|
|
|
Fabricator(:bedrock_model, from: :anthropic_model) do
|
|
|
|
url ""
|
|
|
|
provider "aws_bedrock"
|
|
|
|
api_key "asd-asd-asd"
|
|
|
|
name "claude-3-sonnet"
|
|
|
|
provider_params { { region: "us-east-1", access_key_id: "123456" } }
|
|
|
|
end
|
|
|
|
|
|
|
|
Fabricator(:cohere_model, from: :llm_model) do
|
|
|
|
display_name "Cohere Command R+"
|
|
|
|
name "command-r-plus"
|
|
|
|
provider "cohere"
|
|
|
|
api_key "ABC"
|
|
|
|
url "https://api.cohere.ai/v1/chat"
|
|
|
|
end
|
2024-09-11 21:28:08 -04:00
|
|
|
|
|
|
|
Fabricator(:samba_nova_model, from: :llm_model) do
|
|
|
|
display_name "Samba Nova"
|
|
|
|
name "samba-nova"
|
|
|
|
provider "samba_nova"
|
|
|
|
api_key "ABC"
|
|
|
|
url "https://api.sambanova.ai/v1/chat/completions"
|
|
|
|
end
|
2024-09-30 20:45:03 -04:00
|
|
|
|
|
|
|
Fabricator(:ollama_model, from: :llm_model) do
|
|
|
|
display_name "Ollama llama 3.1"
|
|
|
|
name "llama-3.1"
|
|
|
|
provider "ollama"
|
|
|
|
api_key "ABC"
|
|
|
|
tokenizer "DiscourseAi::Tokenizer::Llama3Tokenizer"
|
|
|
|
url "http://api.ollama.ai/api/chat"
|
2024-10-10 16:25:53 -04:00
|
|
|
provider_params { { enable_native_tool: true } }
|
2024-09-30 20:45:03 -04:00
|
|
|
end
|