Skip to main content
POST
/
models
# Docs for v2 can be found by changing the above selector ^
from together import Together
import os

client = Together(
    api_key=os.environ.get("TOGETHER_API_KEY"),
)

response = client.models.upload(
    "My-Fine-Tuned-Model",
    "https://ml-models.s3.us-west-2.amazonaws.com/models/my-fine-tuned-model.tar.gz",
)

print(response.job_id)
{
  "data": {
    "job_id": "job-a15dad11-8d8e-4007-97c5-a211304de284",
    "model_name": "necolinehubner/Qwen2.5-72B-Instruct",
    "model_id": "model-c0e32dfc-637e-47b2-bf4e-e9b2e58c9da7",
    "model_source": "huggingface"
  },
  "message": "Processing model weights. Job created."
}

Authorizations

​
Authorization
string
header
default:default
required

Bearer authentication header of the form Bearer <token>, where <token> is your auth token.

Body

application/json
​
model_name
string
required

The name to give to your uploaded model

Example:

"Qwen2.5-72B-Instruct"

​
model_source
string
required

The source location of the model (Hugging Face repo or S3 path)

Example:

"unsloth/Qwen2.5-72B-Instruct"

​
model_type
enum<string>
default:model

Whether the model is a full model or an adapter

Available options:
model,
adapter
Example:

"model"

​
hf_token
string

Hugging Face token (if uploading from Hugging Face)

Example:

"hf_examplehuggingfacetoken"

​
description
string

A description of your model

Example:

"Finetuned Qwen2.5-72B-Instruct by Unsloth"

​
base_model
string

The base model to use for an adapter if setting it to run against a serverless pool. Only used for model_type adapter.

Example:

"Qwen/Qwen2.5-72B-Instruct"

​
lora_model
string

The lora pool to use for an adapter if setting it to run against, say, a dedicated pool. Only used for model_type adapter.

Example:

"my_username/Qwen2.5-72B-Instruct-lora"

Response

200 - application/json

Model / adapter upload job created successfully

​
data
object
required
​
message
string
required
Example:

"Processing model weights. Job created."