Concepts
Handling variable input array
This will help you understand how to handle multiple varrible inputs sizes on Inferless
Often time you will want to send 1 or more inputs to your model. Inferless allows you to send multiple inputs to your model. In this guide, we’ll walk you through the steps to handle multiple inputs in Inferless.
To handle multiple inputs you can define the input_schema.py in your GitHub/GitLab repository with ‘shape’ as -1 which will allow dynamic input sizes.
In input_schema.py
Example
INPUT_SCHEMA = {
"prompt": {
'datatype': 'STRING',
'required': True,
'shape': [-1] ,
'example': ["There is a fine house in the forest"]
},
}
In code app.py
def infer(self, inputs):
prompt = inputs["prompt"]
# If while making the API call you send shape as [1] you will get the variable, if the array > 1 you will get an array
if isinstance(prompt, list):
for each in prompt:
print(each)
else:
print(prompt)
You can also flatten it to get all the values in a single list in app.py
def infer(self, inputs):
prompts = []
# If while making the API call you send shape as [1] you will get the variable, if the array > 1 you will get an array
if isinstance(inputs["prompt"], list):
prompts = inputs["prompt"]
else:
prompts.append(inputs["prompt"])
This is how the API call will look like
curl --location 'https://<>.default.model-v1-dev.inferless.com/v2/models/<model_name>/versions/1/infer' \
--header 'Content-Type: application/json' \
--header 'Authorization: Bearer 432424.....3432423' \
--data '{
"inputs": [
{
"name": "prompt",
"shape": [
3
],
"data": [
"There is a fine house in the forest", "There is a house in the forest", "There is a tree in the forest"
],
"datatype": "BYTES"
}
]
}'