Skip to main content

Together AI

LiteLLM supports all models on Together AI.

API Keys​

import os 
os.environ["TOGETHERAI_API_KEY"] = "your-api-key"

Sample Usage​

from litellm import completion 

os.environ["TOGETHERAI_API_KEY"] = "your-api-key"

messages = [{"role": "user", "content": "Write me a poem about the blue sky"}]

completion(model="together_ai/togethercomputer/Llama-2-7B-32K-Instruct", messages=messages)

Together AI Models​

liteLLM supports non-streaming and streaming requests to all models on https://api.together.xyz/

Example TogetherAI Usage - Note: liteLLM supports all models deployed on TogetherAI

Llama LLMs - Chat​

Model NameFunction CallRequired OS Variables
togethercomputer/llama-2-70b-chatcompletion('together_ai/togethercomputer/llama-2-70b-chat', messages)os.environ['TOGETHERAI_API_KEY']

Llama LLMs - Language / Instruct​

Model NameFunction CallRequired OS Variables
togethercomputer/llama-2-70bcompletion('together_ai/togethercomputer/llama-2-70b', messages)os.environ['TOGETHERAI_API_KEY']
togethercomputer/LLaMA-2-7B-32Kcompletion('together_ai/togethercomputer/LLaMA-2-7B-32K', messages)os.environ['TOGETHERAI_API_KEY']
togethercomputer/Llama-2-7B-32K-Instructcompletion('together_ai/togethercomputer/Llama-2-7B-32K-Instruct', messages)os.environ['TOGETHERAI_API_KEY']
togethercomputer/llama-2-7bcompletion('together_ai/togethercomputer/llama-2-7b', messages)os.environ['TOGETHERAI_API_KEY']

Falcon LLMs​

Model NameFunction CallRequired OS Variables
togethercomputer/falcon-40b-instructcompletion('together_ai/togethercomputer/falcon-40b-instruct', messages)os.environ['TOGETHERAI_API_KEY']
togethercomputer/falcon-7b-instructcompletion('together_ai/togethercomputer/falcon-7b-instruct', messages)os.environ['TOGETHERAI_API_KEY']

Alpaca LLMs​

Model NameFunction CallRequired OS Variables
togethercomputer/alpaca-7bcompletion('together_ai/togethercomputer/alpaca-7b', messages)os.environ['TOGETHERAI_API_KEY']

Other Chat LLMs​

Model NameFunction CallRequired OS Variables
HuggingFaceH4/starchat-alphacompletion('together_ai/HuggingFaceH4/starchat-alpha', messages)os.environ['TOGETHERAI_API_KEY']

Code LLMs​

Model NameFunction CallRequired OS Variables
togethercomputer/CodeLlama-34bcompletion('together_ai/togethercomputer/CodeLlama-34b', messages)os.environ['TOGETHERAI_API_KEY']
togethercomputer/CodeLlama-34b-Instructcompletion('together_ai/togethercomputer/CodeLlama-34b-Instruct', messages)os.environ['TOGETHERAI_API_KEY']
togethercomputer/CodeLlama-34b-Pythoncompletion('together_ai/togethercomputer/CodeLlama-34b-Python', messages)os.environ['TOGETHERAI_API_KEY']
defog/sqlcodercompletion('together_ai/defog/sqlcoder', messages)os.environ['TOGETHERAI_API_KEY']
NumbersStation/nsql-llama-2-7Bcompletion('together_ai/NumbersStation/nsql-llama-2-7B', messages)os.environ['TOGETHERAI_API_KEY']
WizardLM/WizardCoder-15B-V1.0completion('together_ai/WizardLM/WizardCoder-15B-V1.0', messages)os.environ['TOGETHERAI_API_KEY']
WizardLM/WizardCoder-Python-34B-V1.0completion('together_ai/WizardLM/WizardCoder-Python-34B-V1.0', messages)os.environ['TOGETHERAI_API_KEY']

Language LLMs​

Model NameFunction CallRequired OS Variables
NousResearch/Nous-Hermes-Llama2-13bcompletion('together_ai/NousResearch/Nous-Hermes-Llama2-13b', messages)os.environ['TOGETHERAI_API_KEY']
Austism/chronos-hermes-13bcompletion('together_ai/Austism/chronos-hermes-13b', messages)os.environ['TOGETHERAI_API_KEY']
upstage/SOLAR-0-70b-16bitcompletion('together_ai/upstage/SOLAR-0-70b-16bit', messages)os.environ['TOGETHERAI_API_KEY']
WizardLM/WizardLM-70B-V1.0completion('together_ai/WizardLM/WizardLM-70B-V1.0', messages)os.environ['TOGETHERAI_API_KEY']

Prompt Templates​

Using a chat model on Together AI with it's own prompt format?

Using Llama2 Instruct models​

If you're using Together AI's Llama2 variants( model=togethercomputer/llama-2..-instruct), LiteLLM can automatically translate between the OpenAI prompt format and the TogetherAI Llama2 one ([INST]..[/INST]).

from litellm import completion 

# set env variable
os.environ["TOGETHERAI_API_KEY"] = ""

messages = [{"role": "user", "content": "Write me a poem about the blue sky"}]

completion(model="together_ai/togethercomputer/Llama-2-7B-32K-Instruct", messages=messages)

Using another model​

You can create a custom prompt template on LiteLLM (and we welcome PRs to add them to the main repo 🤗)

Let's make one for OpenAssistant/llama2-70b-oasst-sft-v10!

The accepted template format is: Reference

"""
<|im_start|>system
{system_message}<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant
"""

Let's register our custom prompt template: Implementation Code

import litellm 

litellm.register_prompt_template(
model="OpenAssistant/llama2-70b-oasst-sft-v10",
roles={
"system": {
"pre_message": "[<|im_start|>system",
"post_message": "\n"
},
"user": {
"pre_message": "<|im_start|>user",
"post_message": "\n"
},
"assistant": {
"pre_message": "<|im_start|>assistant",
"post_message": "\n"
}
}
)

Let's use it!

from litellm import completion 

# set env variable
os.environ["TOGETHERAI_API_KEY"] = ""

messages=[{"role":"user", "content": "Write me a poem about the blue sky"}]

completion(model="together_ai/OpenAssistant/llama2-70b-oasst-sft-v10", messages=messages)

Complete Code

import litellm 
from litellm import completion

# set env variable
os.environ["TOGETHERAI_API_KEY"] = ""

litellm.register_prompt_template(
model="OpenAssistant/llama2-70b-oasst-sft-v10",
roles={
"system": {
"pre_message": "[<|im_start|>system",
"post_message": "\n"
},
"user": {
"pre_message": "<|im_start|>user",
"post_message": "\n"
},
"assistant": {
"pre_message": "<|im_start|>assistant",
"post_message": "\n"
}
}
)

messages=[{"role":"user", "content": "Write me a poem about the blue sky"}]

response = completion(model="together_ai/OpenAssistant/llama2-70b-oasst-sft-v10", messages=messages)

print(response)

Output

{
"choices": [
{
"finish_reason": "stop",
"index": 0,
"message": {
"content": ".\n\nThe sky is a canvas of blue,\nWith clouds that drift and move,",
"role": "assistant",
"logprobs": null
}
}
],
"created": 1693941410.482018,
"model": "OpenAssistant/llama2-70b-oasst-sft-v10",
"usage": {
"prompt_tokens": 7,
"completion_tokens": 16,
"total_tokens": 23
},
"litellm_call_id": "f21315db-afd6-4c1e-b43a-0b5682de4b06"
}

Advanced Usage​

Instead of using the custom_llm_provider arg to specify which provider you're using (e.g. together ai), you can just pass the provider name as part of the model name, and LiteLLM will parse it out.

Expected format: <custom_llm_provider>/<model_name>

e.g. completion(model="together_ai/togethercomputer/Llama-2-7B-32K-Instruct", ...)

from litellm import completion 

# set env variable
os.environ["TOGETHERAI_API_KEY"] = ""

messages = [{"role": "user", "content": "Write me a poem about the blue sky"}]

completion(model="together_ai/togethercomputer/Llama-2-7B-32K-Instruct", messages=messages)