主页 / server / 问题 / 1179143
Accepted
Franck Dernoncourt
Asked: 2025-04-15 15:45:26 +0800 CST

出现“此模型在所选 Azure OpenAI 服务资源上不可用”的错误，但我认为它可用。为什么我漏掉了？

  • 772

我在 Azure 区域部署了一个经过精细调整的 GPT 4o 迷你模型northcentralus。

当我尝试编辑它时，我在 Azure 门户中收到此错误（我想更改最大命中率）：

此模型在所选的 Azure OpenAI 服务资源上不可用。详细了解模型可用性。

在此处输入图片描述

我在 Azure 门户中选择的资源位于northcentralus：

在此处输入图片描述

但是，https ://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models ?tabs=global-standard%2Cstandard-chat-c​​ompletions#fine-tuning-models 指出，经过微调的 GPT 4o 迷你模型可在 Azure 区域使用northcentralus：

在此处输入图片描述

我错过了什么？为什么我收到“此模型在所选 Azure OpenAI 服务资源上不可用”的错误？

我按照 Azure 的GPT 微调教程部署了经过微调的 GPT 4o mini 模型。部署阶段的代码如下：

# Deploy fine-tuned model

import json
import requests

token = '[redacted]'
subscription = '[redacted]'
resource_group = "[redacted]"
resource_name = "[redacted]"
model_deployment_name = "gpt-4o-mini-2024-07-18-ft" # Custom deployment name you chose for your fine-tuning model

deploy_params = {'api-version': "2023-05-01"}
deploy_headers = {'Authorization': 'Bearer {}'.format(token), 'Content-Type': 'application/json'}

deploy_data = {
    "sku": {"name": "standard", "capacity": 1},
    "properties": {
        "model": {
            "format": "OpenAI",
            "name": "gpt-4o-mini-2024-07-18.ft-[redacted]", #retrieve this value from the previous call, it will look like gpt-4o-mini-2024-07-18.ft-[redacted]
            "version": "1"
        }
    }
}
deploy_data = json.dumps(deploy_data)

request_url = f'https://management.azure.com/subscriptions/{subscription}/resourceGroups/{resource_group}/providers/Microsoft.CognitiveServices/accounts/{resource_name}/deployments/{model_deployment_name}'

print('Creating a new deployment...')

r = requests.put(request_url, params=deploy_params, headers=deploy_headers, data=deploy_data)

print(r)
print(r.reason)
print(r.json())

令牌是通过生成的az account get-access-token。

azure

1 个回答

  1. Best Answer

    我已经放弃了那个 Azure UI，这是实现它的 Python 代码。它需要通过 生成的令牌 az account get-access-token。

    import json
import requests

new_capacity = 3 # Change this number to your desired capacity. 3 means 3000 tokens/minute.

# Authentication and resource identification
token = "YOUR_BEARER_TOKEN"  # Replace with your actual token
subscription = ''
resource_group = ""
resource_name = ""
model_deployment_name = ""

# API parameters and headers
update_params = {'api-version': "2023-05-01"}
update_headers = {'Authorization': 'Bearer {}'.format(token), 'Content-Type': 'application/json'}

# First, get the current deployment to preserve its configuration
request_url = f'https://management.azure.com/subscriptions/{subscription}/resourceGroups/{resource_group}/providers/Microsoft.CognitiveServices/accounts/{resource_name}/deployments/{model_deployment_name}'
r = requests.get(request_url, params=update_params, headers=update_headers)

if r.status_code != 200:
    print(f"Failed to get current deployment: {r.status_code}")
    print(r.reason)
    if hasattr(r, 'json'):
        print(r.json())
    exit(1)

# Get the current deployment configuration
current_deployment = r.json()

# Update only the capacity in the configuration
update_data = {
    "sku": {
        "name": current_deployment["sku"]["name"],
        "capacity": new_capacity  
    },
    "properties": current_deployment["properties"]
}

update_data = json.dumps(update_data)

print('Updating deployment capacity...')

# Use PUT to update the deployment
r = requests.put(request_url, params=update_params, headers=update_headers, data=update_data)

print(f"Status code: {r.status_code}")
print(f"Reason: {r.reason}")
if hasattr(r, 'json'):
    print(r.json())

    需要几秒钟才能更新。

    • 0

相关问题