我是用阿里灵积平台的deepseek-v3的配置，该怎么写呢？ OPENAI_API_KEY=sk-5523020cf22bf5a OPENAI_BASE_URL=https://dashscope.aliyuncs.com/compatible-mode/v1 DEEPSEEK_API_KEY= COMPLETION_MODEL=deepseek-v3

我是这么写的，之后litellm就报错没有提供llm_provider.我写成openai/deepseek-v3,报错找不到模型。

Feb 18 '25 03:02 DAAworld

写成openai/deepseek-v3报错 File "/mnt/data/rz/project/AutoAgent/autoagent/core.py", line 108, in get_chat_completion assert litellm.supports_function_calling(model = create_model) == True, f"Model {create_model} does not support function calling, please set FN_CALL=False to use non-function calling mode" File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/litellm/utils.py", line 1833, in supports_function_calling raise Exception( Exception: Model not found or error in checking function calling support. You passed model=deepseek-v3, custom_llm_provider=openai. Error: This model isn't mapped yet. model=deepseek-v3, custom_llm_provider=openai. Add it here - https://github.com/BerriAI/litellm/blob/main/model_prices_and_context_window.json.

Feb 18 '25 03:02 DAAworld

写成deepseek-v3报错 File "/mnt/data/rz/project/AutoAgent/autoagent/cli.py", line 206, in main user_mode(model, context_variables, False) File "/mnt/data/rz/project/AutoAgent/autoagent/cli.py", line 269, in user_mode response = client.run(agent, messages, context_variables, debug=debug) File "/mnt/data/rz/project/AutoAgent/autoagent/core.py", line 384, in run completion = self.get_chat_completion( File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/tenacity/init.py", line 336, in wrapped_f return copy(f, *args, **kw) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/tenacity/init.py", line 475, in call do = self.iter(retry_state=retry_state) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/tenacity/init.py", line 376, in iter result = action(retry_state) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/tenacity/init.py", line 398, in self._add_action_func(lambda rs: rs.outcome.result()) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/concurrent/futures/_base.py", line 451, in result return self.__get_result() File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/concurrent/futures/_base.py", line 403, in __get_result raise self._exception File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/tenacity/init.py", line 478, in call result = fn(*args, **kwargs) File "/mnt/data/rz/project/AutoAgent/autoagent/core.py", line 108, in get_chat_completion assert litellm.supports_function_calling(model = create_model) == True, f"Model {create_model} does not support function calling, please set FN_CALL=False to use non-function calling mode" File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/litellm/utils.py", line 1833, in supports_function_calling raise Exception( Exception: Model not found or error in checking function calling support. You passed model=deepseek-v3, custom_llm_provider=None. Error: litellm.BadRequestError: LLM Provider NOT provided. Pass in the LLM provider you are trying to call. You passed model=deepseek-v3 Pass model as E.g. For 'Huggingface' inference endpoints pass in completion(model='huggingface/starcoder',..) Learn more: https://docs.litellm.ai/docs/providers

Feb 18 '25 03:02 DAAworld

你好，建议使用openai兼容endpoints，注意openai兼容endpoints不支持原生function calling，需要设置FN_CALL=False，以下是正确的使用命令。我们的repos又更新了，您可以pull一下最新更新。

COMPLETION_MODEL=openai/deepseek-v3 API_BASE_URL=https://dashscope.aliyuncs.com/compatible-mode/v1 FN_CALL=False OPENAI_API_KEY=sk-your-api-key auto main

Feb 18 '25 04:02 tjb-tech

我更新代码之后再执行命令就报这个错误

(autoagent) rz@rz-v100:/mnt/data/rz/project/AutoAgent$ auto main Failed to create and switch to new branch. Error: fatal: 一个分支名 'autoagent_mirror_None' 已经存在。

Successfully switched to new branch: autoagent_mirror_None Container 'auto_agent' has been created and started. ⠙ Creating environment... Browser Env [2025-02-18 15:30:37] ⠼ Creating environment... Browser Env [2025-02-18 15:31:42] ⠇ Creating environment... Browser Env [2025-02-18 15:31:43] Starting browser env... Browser Env [2025-02-18 15:31:43] Traceback (most recent call last): File "/mnt/data/rz/miniconda3/envs/autoagent/bin/auto", line 8, in sys.exit(cli()) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/click/core.py", line 1161, in call return self.main(*args, **kwargs) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/click/core.py", line 1082, in main rv = self.invoke(ctx) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/click/core.py", line 1697, in invoke return _process_result(sub_ctx.command.invoke(sub_ctx)) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/click/core.py", line 1443, in invoke return ctx.invoke(self.callback, **ctx.params) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/click/core.py", line 788, in invoke return __callback(*args, **kwargs) File "/mnt/data/rz/project/AutoAgent/autoagent/cli.py", line 191, in main code_env, web_env, file_env = create_environment(docker_config) File "/mnt/data/rz/project/AutoAgent/autoagent/cli.py", line 151, in create_environment web_env = BrowserEnv(browsergym_eval_env = None, local_root=docker_config.local_root, workplace_name=docker_config.workplace_name) File "/mnt/data/rz/project/AutoAgent/autoagent/environment/browser_env.py", line 368, in init self.init_browser() File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/tenacity/init.py", line 336, in wrapped_f return copy(f, *args, **kw) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/tenacity/init.py", line 475, in call do = self.iter(retry_state=retry_state) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/tenacity/init.py", line 376, in iter result = action(retry_state) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/tenacity/init.py", line 398, in self._add_action_func(lambda rs: rs.outcome.result()) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/concurrent/futures/_base.py", line 451, in result return self.__get_result() File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/concurrent/futures/_base.py", line 403, in __get_result raise self._exception File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/tenacity/init.py", line 478, in call result = fn(*args, **kwargs) File "/mnt/data/rz/project/AutoAgent/autoagent/environment/browser_env.py", line 393, in init_browser self.process.start() File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/multiprocessing/process.py", line 121, in start self._popen = self._Popen(self) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/multiprocessing/context.py", line 224, in _Popen return _default_context.get_context().Process._Popen(process_obj) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/multiprocessing/context.py", line 288, in _Popen return Popen(process_obj) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/multiprocessing/popen_spawn_posix.py", line 32, in init super().init(process_obj) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/multiprocessing/popen_fork.py", line 19, in init self._launch(process_obj) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/multiprocessing/popen_spawn_posix.py", line 47, in _launch reduction.dump(process_obj, fp) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/multiprocessing/reduction.py", line 60, in dump ForkingPickler(file, protocol).dump(obj) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/multiprocessing/connection.py", line 968, in reduce_connection df = reduction.DupFd(conn.fileno()) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/multiprocessing/connection.py", line 170, in fileno self._check_closed() File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/multiprocessing/connection.py", line 136, in _check_closed raise OSError("handle is closed") OSError: handle is closed (autoagent) rz@rz-v100:/mnt/data/rz/project/AutoAgent$

Feb 18 '25 07:02 DAAworld

看上去启动浏览器环境的时候出了问题，请问您之前可以正常启动浏览器环境嘛（即进入AutoAgent的起始界面）？您可以检查一下是否是内存不足导致的

Feb 18 '25 07:02 tjb-tech

更新之前是可以进入到模式选择界面的。之后就不行了。

Feb 18 '25 07:02 DAAworld

跟命令行中启动有头浏览器有关吗？

Feb 18 '25 07:02 DAAworld

您可以尝试关闭有头浏览器，在https://github.com/HKUDS/AutoAgent/blob/main/autoagent/environment/browser_env.py中设置

env = gym.make(
                'browsergym/openended',
                task_kwargs={'start_url': 'about:blank', 'goal': 'PLACEHOLDER_GOAL'},
                wait_for_user_message=False,
                headless=True,
                disable_env_checker=True,
                tags_to_mark='all',
                action_mapping = action_mapping
            )

因为有头浏览器所需要的计算资源较多

Feb 18 '25 07:02 tjb-tech

请问您解决了嘛

Feb 18 '25 08:02 tjb-tech

我能进入模式选择了，但是按照上面您的建议写的

Required Github Tokens

GITHUB_AI_TOKEN=ghp_xdsrGualJxvi1nNXfj

Optional API Keys

OPENAI_API_KEY=sk-556b15aa9623020cf22bf5a OPENAI_BASE_URL=https://dashscope.aliyuncs.com/compatible-mode/v1 DEEPSEEK_API_KEY= COMPLETION_MODEL=openai/deepseek-v3 FN_CALL=False ANTHROPIC_API_KEY= GEMINI_API_KEY= HUGGINGFACE_API_KEY= GROQ_API_KEY= XAI_API_KEY=

没有响应，过一段时间，就在重试，

Feb 18 '25 08:02 DAAworld

你好这个可能和您LLM的rate limit有关，您可以设置DEBUG=True MC_MODE=Fasle来看到详细的消息记录来供我们debug。期待您的进一步feedback

Feb 18 '25 09:02 tjb-tech

这个结果测试起来跟FN_CALL=False有关，如果设置=False,base_url正常，设置成True，base——url就是openai的。我想请问下国内有什么模型可以复现readme里面的对话吗？我目前用国内阿里的灵积平台是不能的。

Feb 18 '25 11:02 DAAworld

请问您现在的问题是可以正常运行，但是结果不尽人意嘛？Agent的运行和base model的性能确实有关，您可以试下deepseek v3，deepseek r1，和qwen max这种大参数模型。Claude-3.5 和gpt-4o当然是首选。

Feb 18 '25 12:02 tjb-tech

Tell me what do you want to create with Agent Chain? (type "exit" to quit, press "Enter" to continue): I want to createFinancial Agent that can help me to search the financial information online.You may help me to - get balance sheets for a given ticker over a given period.-get cash flow statements for a given ticker over a given period.-get income statements for a given ticker over a given period. Your request: I want to createFinancial Agent that can help me to search the financial information online.You may help me to - get balance sheets for a given ticker over a given period.-get cash flow statements for a given ticker over a given period.-get income statements for a given ticker over a given period. @Agent Former Agent will help you, be patient... Traceback (most recent call last): File "/mnt/data/rz/miniconda3/envs/autoagent/bin/auto", line 8, in sys.exit(cli()) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/click/core.py", line 1161, in call return self.main(*args, **kwargs) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/click/core.py", line 1082, in main rv = self.invoke(ctx) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/click/core.py", line 1697, in invoke return _process_result(sub_ctx.command.invoke(sub_ctx)) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/click/core.py", line 1443, in invoke return ctx.invoke(self.callback, **ctx.params) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/click/core.py", line 788, in invoke return __callback(*args, **kwargs) File "/mnt/data/rz/project/AutoAgent/autoagent/cli.py", line 213, in main meta_agent(model, context_variables, False) File "/mnt/data/rz/project/AutoAgent/autoagent/cli_utils/metachain_meta_agent.py", line 216, in meta_agent agent_form, output_xml_form, messages = agent_profiling(agent_former, client, messages, context_variables, requirements, debug) File "/mnt/data/rz/project/AutoAgent/autoagent/cli_utils/metachain_meta_agent.py", line 29, in agent_profiling response = client.run(agent_former, messages, context_variables, debug=debug) File "/mnt/data/rz/project/AutoAgent/autoagent/core.py", line 384, in run completion = self.get_chat_completion( File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/tenacity/init.py", line 336, in wrapped_f return copy(f, *args, **kw) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/tenacity/init.py", line 475, in call do = self.iter(retry_state=retry_state) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/tenacity/init.py", line 376, in iter result = action(retry_state) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/tenacity/init.py", line 398, in self._add_action_func(lambda rs: rs.outcome.result()) File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/concurrent/futures/_base.py", line 451, in result return self.__get_result() File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/concurrent/futures/_base.py", line 403, in __get_result raise self._exception File "/mnt/data/rz/miniconda3/envs/autoagent/lib/python3.10/site-packages/tenacity/init.py", line 478, in call result = fn(*args, **kwargs) File "/mnt/data/rz/project/AutoAgent/autoagent/core.py", line 135, in get_chat_completion assert agent.tool_choice == "required", f"Non-function calling mode MUST use tool_choice = 'required' rather than {agent.tool_choice}" AssertionError: Non-function calling mode MUST use tool_choice = 'required' rather than None

这个报错是什么原因呢？设置的灵积平台端点，用的deepseek-v3

Feb 19 '25 01:02 DAAworld

Hi, 你好，这个是因为deepseek-v3使用是采用了tool use转换操作，而agent profiling不需要调用tools。我们会修复这个bug，您有使用其他功能嘛，例如user mode

Feb 20 '25 10:02 tjb-tech

感谢您的回复，是用usermode同样也会报错误，分两种情况，一种是大模型直接响应的比如“帮我做首诗”，这不会报错。当我询问“帮我查询天气”的时候会报错没有playwright的down，page—down。。。的tool，这个错误我没有截屏出来。

Feb 20 '25 11:02 DAAworld

siliconflow api not used，how to use siliconflow api，for example?

Mar 10 '25 10:03 zhongmiyu99

deepseek的官网被攻击，填写deepseek的api_key无法使用报错。

Required Github Tokens

Optional API Keys