dat1-cli 0.1.0

No project description provided

dat1-cli

A command line interface for the dat1 platform.

Installation

pip install dat1-cli

Usage

Initialize with your API key:

dat1 login

To initialize a new model project, run in the root directory of your project:

dat1 init

To upload your model to the platform:

dat1 deploy

A good starting point for your model is using the example model.

Otherwise, the platform expects a handler.py file in the root directory of your project that contains a FastAPI app with two endpoints: GET / for healthchecks and POST /infer for inference. An example handler is shown below:

from fastapi import Request, FastAPI
from vllm import LLM, SamplingParams
import os

llm = LLM(model=os.path.expanduser('./'), load_format="safetensors", enforce_eager=True)

app = FastAPI()

@app.get("/")
async def root():
    return "OK"

@app.post("/infer")
async def infer(request: Request):
    request = await request.json()
    prompts = request["prompt"]
    sampling_params = SamplingParams(temperature=0.8, top_p=0.95)
    outputs = llm.generate(prompts, sampling_params)
    return { "response" : outputs[0].outputs[0].text }

Launching Locally

Pre-requisites

• Docker
• CUDA-compatible GPU
• NVIDIA Container Toolkit

To launch your model locally, run:

dat1 serve

License

MIT