binary-husky/gpt_academic · error · ValueError

Endpoint不正确, 请检查AZURE_ENDPOINT的配置! 当前的Endpoint为:

Error message

Endpoint不正确, 请检查AZURE_ENDPOINT的配置! 当前的Endpoint为:

What it means

The Cohere bridge copies the ChatGPT bridge's verify_endpoint guard: an lru_cache function that raises ValueError if the endpoint string still contains the template placeholder '你亲手写的api名称' from config.py's Azure example. It fires before any request when an azure-* model was selected but the endpoint (or AZURE_CFG_ARRAY entry) was never actually filled in. Purpose: fail fast on unfinished Azure configuration rather than getting a confusing 404 later.

Source

Thrown at request_llms/bridge_cohere.py:68

    try:
        chunkjson = json.loads(chunk_decoded)
        has_choices = 'choices' in chunkjson
        if has_choices: choice_valid = (len(chunkjson['choices']) > 0)
        if has_choices and choice_valid: has_content = ("content" in chunkjson['choices'][0]["delta"])
        if has_content: has_content = (chunkjson['choices'][0]["delta"]["content"] is not None)
        if has_choices and choice_valid: has_role = "role" in chunkjson['choices'][0]["delta"]
    except:
        pass
    return chunk_decoded, chunkjson, has_choices, choice_valid, has_content, has_role

from functools import lru_cache
@lru_cache(maxsize=32)
def verify_endpoint(endpoint):
    """
        检查endpoint是否可用
    """
    if "你亲手写的api名称" in endpoint:
        raise ValueError("Endpoint不正确, 请检查AZURE_ENDPOINT的配置! 当前的Endpoint为:" + endpoint)
    return endpoint

def predict_no_ui_long_connection(inputs:str, llm_kwargs:dict, history:list=[], sys_prompt:str="", observe_window:list=None, console_silence:bool=False):
    """
    发送，等待回复，一次性完成，不显示中间过程。但内部用stream的方法避免中途网线被掐。
    inputs：
        是本次问询的输入
    sys_prompt:
        系统静默prompt
    llm_kwargs：
        内部调优参数
    history：
        是之前的对话列表
    observe_window = None：
        用于负责跨越线程传递已经输出的部分，大部分时候仅仅为了fancy的视觉效果，留空即可。observe_window[0]：观测窗。observe_window[1]：看门狗
    """
    watch_dog_patience = 5 # 看门狗的耐心, 设置5秒即可
    headers, payload = generate_payload(inputs, llm_kwargs, history, system_prompt=sys_prompt, stream=True)

View on GitHub (pinned to d6bde0fa54)

Solutions

1. Fill AZURE_ENDPOINT and the model's AZURE_CFG_ARRAY entry in config_private.py with the real deployment URL.
2. Confirm the deployment name and api-version query parameter match your Azure portal deployment.
3. Grep your config for the literal placeholder '你亲手写的api名称' to catch every unfilled field.
4. Restart gpt_academic so verify_endpoint's lru_cache doesn't serve stale state.

Example fix

# config_private.py — before
AZURE_ENDPOINT = "https://你亲手写的api名称.openai.azure.com/..."

# after
AZURE_ENDPOINT = "https://cohere-resource.openai.azure.com/openai/deployments/cohere/chat/completions?api-version=2023-05-15"

endpoint = AZURE_CFG_ARRAY[model]["AZURE_ENDPOINT"]
assert "你亲手写的api名称" not in endpoint, "Azure endpoint still contains the template placeholder"

try:
    verify_endpoint(endpoint)
except ValueError as e:
    fail_fast_with_config_hint(str(e))

When it happens

Trigger: Selecting an azure- model routed to the Cohere bridge while AZURE_ENDPOINT still contains '你亲手写的api名称'; copying config.py to config_private.py and filling the key but not the endpoint URL; AZURE_CFG_ARRAY retaining placeholder values after a config merge.

Common situations: Incomplete Azure setup for Cohere models via Azure gateways; config template placeholders surviving upgrades; users assuming the key alone is enough for azure- model names.

Related errors

• Endpoint不正确, 请检查AZURE_ENDPOINT的配置! 当前的Endpoint为:
• Endpoint不正确, 请检查AZURE_ENDPOINT的配置! 当前的Endpoint为:{endpoint}
• AZURE_CFG_ARRAY中配置的模型必须以azure开头
• 在线搜索失败，状态码: {response.status_code}\t{response.content.decode
• 用户代理或助理代理未定义

AI-assisted analysis of binary-husky/gpt_academic@d6bde0fa54 (2026-08-14). Data as JSON: /api/errors/9d3a43701173be78. Report an issue: GitHub.