A Flowchart based automation platform to run deep learning workloads with minimal to no code.
otto-m8 (automate) lets users spin up a wide range of AI models, starting from Traditional deep learning models to large language models, all through a flowchart like user interface. At its core, otto-m8 will deploy a Docker container of your workflow that you can use as an API to integrate with your existing workflows, build a AI assistant chatbot or use it as a standalone API/application.
The idea is simple-provide a easy-to-use user interface to spin up AI models. A lot of code needed to run AI models(both LLMs and traditional deep learning models) are boilerplate code blocks, including the deployments which is more often than not a REST API serving the model. The goal of otto-m8 is not only to abstract that through code but to abstract the entire process into a UI. Otto-m8 operates with a Input, Process, Output paradigm, where every flow has some form of input, that gets processed via a series of processes, and then an output.
This is currently an MVP and made source available, which is to say this is not an Open Source software.
- Pre-req: Make sure to have Docker or Docker Desktop Installed on your computer, and in order to run Ollama blocks, make sure you have the Ollama server running in the background.
- Run the following command to make
run.shexecutable
chmod +x run.sh- Then launch the application:
./run.shThis should launch both the dashboard and the server. To access the dashboard, head over to http://localhost:3000/. Use the default login credentials to access the dashboard, and get started on your first workflow.
Below is an example of a workflow that incorporates Langchain's PDF Parser to build a workflow:
To run it as an API:
import requests
import base64
import json
# Find the deployment URL on the Template page
deployment_url = "http://localhost:8001/workflow_run"
path_to_pdf = "./AMZN-Q1-2024-Earnings-Release.pdf"
# Any kind of upload documents expect a base64 encoded string.
with open(path_to_pdf, "rb") as f:
data = f.read()
data_base64 = base64.b64encode(data).decode("utf-8")
# Based on the Block's displayed name(`id` on Block config tab), append your data:
payload = {
"Langchain_PDF_Parser": data_base64,
"Input_Block": "What was amazon's net sales?"
}
request = requests.post(
deployment_url,
json={"data": payload}
)
response = request.json()['message']
response = json.loads(response)
print(response)Output:
"""
{
"f92cffae-14d2-43f4-a961-2fcd5829f1bc": {
"id": "chatcmpl-AgOrZExPgef0TzHVGNVJJr1vmrJyR",
"choices": [
{
"finish_reason": "stop",
"index": 0,
"logprobs": null,
"message": {
"content": "In the first quarter of 2024, Amazon's net sales increased by 13% to $143.3 billion, compared with $127.4 billion in the first quarter of 2023.",
"refusal": null,
"role": "assistant",
"function_call": null,
"tool_calls": null
}
}
],
"created": 1734668713,
"model": "gpt-4o-mini-2024-07-18",
"object": "chat.completion",
"service_tier": null,
"system_fingerprint": "fp_0aa8d3e20b",
"usage": {
"completion_tokens": 42,
"prompt_tokens": 13945,
"total_tokens": 13987,
"completion_tokens_details": {
"audio_tokens": 0,
"reasoning_tokens": 0,
"accepted_prediction_tokens": 0,
"rejected_prediction_tokens": 0
},
"prompt_tokens_details": {
"audio_tokens": 0,
"cached_tokens": 13824
}
},
"conversation": [
{
"role": "user",
"content": "What was amazon's net sales?"
},
{
"role": "assistant",
"content": "In the first quarter .."
}
]
}
}
"""Use the Chat Output block to use the chat interface:
You can run almost any Huggingface model(although not really) that can be run via
Huggingface's pipeline abstraction. Below is a simple demo of the Salesforce/blip-image-captioning-base model.
- Basic Chatbot and HF workflows
- Function Calling
- Lambda Functions/ Custom Code blocks
- Multimodality for Huggingface
- Multimodality for OpenAI
- SDK for interacting with deployed workflows
- Observability for every block's output like a Logger
- Memory for Chatbot and RAG. Goal is to not clutter the drawing board.
- Streamline workflow creation, edits and redeployment. Some form of version control
- ML Model Training via UI?