Udostępnij za pomocą

Używanie usługi Azure OpenAI w usłudze Fabric z interfejsem API REST (wersja zapoznawcza)

Ważne

Ta funkcja jest dostępna w wersji zapoznawczej.

W tym dokumencie przedstawiono przykłady używania usługi Azure OpenAI w usłudze Fabric przy użyciu interfejsu API REST.

Inicjowanie

from synapse.ml.mlflow import get_mlflow_env_config
from trident_token_library_wrapper import PyTridentTokenLibrary

mlflow_env_configs = get_mlflow_env_config()
mwc_token = PyTridentTokenLibrary.get_mwc_token(mlflow_env_configs.workspace_id, mlflow_env_configs.artifact_id, 2)

auth_headers = {
    "Authorization" : "MwcToken {}".format(mwc_token)
}

Czat

GPT-4o i GPT-4o-mini to modele językowe zoptymalizowane pod kątem interfejsów konwersacyjnych.

import requests

def print_chat_result(messages, response_code, response):
    print("==========================================================================================")
    print("| OpenAI Input    |")
    for msg in messages:
        if msg["role"] == "system":
            print("[System] ", msg["content"])
        elif msg["role"] == "user":
            print("Q: ", msg["content"])
        else:
            print("A: ", msg["content"])
    print("------------------------------------------------------------------------------------------")
    print("| Response Status |", response_code)
    print("------------------------------------------------------------------------------------------")
    print("| OpenAI Output   |")
    if response.status_code == 200:
        print(response.json()["choices"][0]["message"]["content"])
    else:
        print(response.content)
    print("==========================================================================================")


deployment_name = "gpt-4o" # deployment_id could be one of {gpt-4o or gpt-4o-mini}
openai_url = mlflow_env_configs.workload_endpoint + f"cognitive/openai/openai/deployments/{deployment_name}/chat/completions?api-version=2025-04-01-preview"
payload = {
    "messages": [
        {"role": "system", "content": "You are an AI assistant that helps people find information."},
        {"role": "user", "content": "Does Azure OpenAI support customer managed keys?"}
    ]
}

response = requests.post(openai_url, headers=auth_headers, json=payload)
print_chat_result(payload["messages"], response.status_code, response)

Wynik

==========================================================================================
| OpenAI Input    |
[System]  You are an AI assistant that helps people find information.
Q:  Does Azure OpenAI support customer managed keys?
------------------------------------------------------------------------------------------
| Response Status | 200
------------------------------------------------------------------------------------------
| OpenAI Output   |
As of my last training cut-off in October 2023, Azure OpenAI Service did not natively support customer-managed keys (CMK) for encryption of data at rest. Data within Azure OpenAI is typically encrypted using Microsoft-managed keys. 

However, you should verify this information on the official Azure documentation or by contacting Microsoft support, as cloud service features and capabilities are frequently updated.
==========================================================================================

Osadzanie

Osadzanie to specjalny format reprezentacji danych, z którego mogą łatwo korzystać modele i algorytmy uczenia maszynowego. Zawiera on semantycznie sformatowane informacje znaczenie tekstu reprezentowane przez wektor liczb zmiennoprzecinkowych. Odległość między dwoma osadzaniami w przestrzeni wektorowej jest powiązana z semantyczną podobieństwem między dwoma oryginalnymi danymi wejściowymi. Jeśli na przykład dwa teksty są podobne, ich reprezentacje wektorowe również powinny być podobne.

Aby uzyskać dostęp do punktu końcowego osadzania usługi Azure OpenAI w sieci szkieletowej, możesz wysłać żądanie interfejsu API przy użyciu następującego formatu:

POST <url_prefix>/openai/deployments/<deployment_name>/embeddings?api-version=2024-02-01

deployment_name może to być text-embedding-ada-002.

import requests

def print_embedding_result(prompts, response_code, response):
    print("==========================================================================================")
    print("| OpenAI Input    |\n\t" + "\n\t".join(prompts))
    print("------------------------------------------------------------------------------------------")
    print("| Response Status |", response_code)
    print("------------------------------------------------------------------------------------------")
    print("| OpenAI Output   |")
    if response_code == 200:
        for data in response.json()['data']:
            print("\t[" + ", ".join([f"{n:.8f}" for n in data["embedding"][:10]]) + ", ... ]")
    else:
        print(response.content)
    print("==========================================================================================")

deployment_name = "text-embedding-ada-002"
openai_url = mlflow_env_configs.workload_endpoint + f"cognitive/openai/openai/deployments/{deployment_name}/embeddings?api-version=2025-04-01-preview"
payload = {
    "input": [
        "empty prompt, need to fill in the content before the request",
        "Once upon a time"
    ]
}

response = requests.post(openai_url, headers=auth_headers, json=payload)
print_embedding_result(payload["input"], response.status_code, response)

Wyjście:

==========================================================================================
| OpenAI Input    |
	empty prompt, need to fill in the content before the request
	Once upon a time
------------------------------------------------------------------------------------------
| Response Status | 200
------------------------------------------------------------------------------------------
| OpenAI Output   |
	[-0.00258819, -0.00449802, -0.01700411, 0.00405622, -0.03064079, 0.01899395, -0.01295485, -0.01426286, -0.03512142, -0.01831212, ... ]
	[0.02129045, -0.02013996, -0.00462094, -0.01146069, -0.01123944, 0.00199124, 0.00228992, -0.01370478, 0.00855917, -0.01470356, ... ]
==========================================================================================

Powiązana zawartość

  • Last updated on