using with deepseek api problem

[ivan-vilches]

Hello i installed using pip install
then exported the api key of deepseek :
export DEEPSEEK_API_KEY=sk-mysecretkey

after that i run
(venv) ivanvilches@MacBook-Air-de-Ivan gpt engineer rails % gpte --model deepseek/deepseek-coder --improve proyecto
╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮
│ /Users/ivanvilches/proyectos mios/gpt engineer/gpt engineer │
│ rails/venv/lib/python3.12/site-packages/gpt_engineer/applications/cli/main.py:402 in main │
│ │
│ 399 │ if llm_via_clipboard: │
│ 400 │ │ ai = ClipboardAI() │
│ 401 │ else: │
│ ❱ 402 │ │ ai = AI( │
│ 403 │ │ │ model_name=model, │
│ 404 │ │ │ temperature=temperature, │
│ 405 │ │ │ azure_endpoint=azure_endpoint, │
│ │
│ ╭────────────────────── locals ──────────────────────╮ │
│ │ azure_endpoint = '' │ │
│ │ clarify_mode = False │ │
│ │ debug = False │ │
│ │ entrypoint_prompt_file = '' │ │
│ │ image_directory = '' │ │
│ │ improve_mode = True │ │
│ │ lite_mode = False │ │
│ │ llm_via_clipboard = False │ │
│ │ model = 'deepseek/deepseek-coder' │ │
│ │ no_execution = False │ │
│ │ project_path = 'proyecto' │ │
│ │ prompt_file = 'prompt' │ │
│ │ self_heal_mode = False │ │
│ │ temperature = 0.1 │ │
│ │ use_cache = False │ │
│ │ use_custom_preprompts = False │ │
│ │ verbose = False │ │
│ ╰────────────────────────────────────────────────────╯ │
│ │
│ /Users/ivanvilches/proyectos mios/gpt engineer/gpt engineer │
│ rails/venv/lib/python3.12/site-packages/gpt_engineer/core/ai.py:115 in init │
│ │
│ 112 │ │ │ or ("gpt-4-turbo" in model_name and "preview" not in model_name) │
│ 113 │ │ │ or ("claude" in model_name) │
│ 114 │ │ ) │
│ ❱ 115 │ │ self.llm = self._create_chat_model() │
│ 116 │ │ self.token_usage_log = TokenUsageLog(model_name) │
│ 117 │ │ │
│ 118 │ │ logger.debug(f"Using model {self.model_name}") │
│ │
│ ╭───────────────────────────── locals ─────────────────────────────╮ │
│ │ azure_endpoint = '' │ │
│ │ model_name = 'deepseek/deepseek-coder' │ │
│ │ self = <gpt_engineer.core.ai.AI object at 0x11a78f590> │ │
│ │ streaming = True │ │
│ │ temperature = 0.1 │ │
│ │ vision = False │ │
│ ╰──────────────────────────────────────────────────────────────────╯ │
│ │
│ /Users/ivanvilches/proyectos mios/gpt engineer/gpt engineer │
│ rails/venv/lib/python3.12/site-packages/gpt_engineer/core/ai.py:372 in _create_chat_model │
│ │
│ 369 │ │ │ │ max_tokens=4096, # vision models default to low max token limits │
│ 370 │ │ │ ) │
│ 371 │ │ else: │
│ ❱ 372 │ │ │ return ChatOpenAI( │
│ 373 │ │ │ │ model=self.model_name, │
│ 374 │ │ │ │ temperature=self.temperature, │
│ 375 │ │ │ │ streaming=self.streaming, │
│ │
│ ╭──────────────────────── locals ────────────────────────╮ │
│ │ self = <gpt_engineer.core.ai.AI object at 0x11a78f590> │ │
│ ╰────────────────────────────────────────────────────────╯ │
│ │
│ /Users/ivanvilches/proyectos mios/gpt engineer/gpt engineer │
│ rails/venv/lib/python3.12/site-packages/pydantic/v1/main.py:341 in init │
│ │
│ 338 │ │ # Uses something other than self the first arg to allow "self" as a settable a │
│ 339 │ │ values, fields_set, validation_error = validate_model(pydantic_self._class │
│ 340 │ │ if validation_error: │
│ ❱ 341 │ │ │ raise validation_error │
│ 342 │ │ try: │
│ 343 │ │ │ object_setattr(pydantic_self, 'dict', values) │
│ 344 │ │ except TypeError as e: │
│ │
│ ╭─────────────────────────────────────────── locals ───────────────────────────────────────────╮ │
│ │ data = { │ │
│ │ │ 'model': 'deepseek/deepseek-coder', │ │
│ │ │ 'temperature': 0.1, │ │
│ │ │ 'streaming': True, │ │
│ │ │ 'callbacks': [ │ │
│ │ │ │ │ │
│ │ <langchain_core.callbacks.streaming_stdout.StreamingStdOutCallbackHandler │ │
│ │ object at 0x11a78eed0> │ │
│ │ │ ], │ │
│ │ │ 'model_kwargs': {} │ │
│ │ } │ │
│ │ fields_set = {'model_kwargs', 'temperature', 'model_name', 'streaming', 'callbacks'} │ │
│ │ validation_error = ValidationError( │ │
│ │ │ model='ChatOpenAI', │ │
│ │ │ errors=[ │ │
│ │ │ │ { │ │
│ │ │ │ │ 'loc': ('root',), │ │
│ │ │ │ │ 'msg': 'Did not find openai_api_key, please add an │ │
│ │ environment variable OPENAI_API_KEY'+66, │ │
│ │ │ │ │ 'type': 'value_error' │ │
│ │ │ │ } │ │
│ │ │ ] │ │
│ │ ) │ │
│ │ values = { │ │
│ │ │ 'name': None, │ │
│ │ │ 'cache': None, │ │
│ │ │ 'verbose': False, │ │
│ │ │ 'callbacks': [ │ │
│ │ │ │ │ │
│ │ <langchain_core.callbacks.streaming_stdout.StreamingStdOutCallbackHandler │ │
│ │ object at 0x11a78eed0> │ │
│ │ │ ], │ │
│ │ │ 'tags': None, │ │
│ │ │ 'metadata': None, │ │
│ │ │ 'custom_get_token_ids': None, │ │
│ │ │ 'callback_manager': None, │ │
│ │ │ 'rate_limiter': None, │ │
│ │ │ 'client': None, │ │
│ │ │ ... +29 │ │
│ │ } │ │
│ ╰──────────────────────────────────────────────────────────────────────────────────────────────╯ │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
ValidationError: 1 validation error for ChatOpenAI
root
Did not find openai_api_key, please add an environment variable OPENAI_API_KEY which contains it, or pass openai_api_key as a named parameter.
(type=value_error)

No sure what i do wrong it keep asking me for the open ai key and i want to use deepseek

[zigabrencic]

Hey

Try to set:

export OPENAI_API_BASE="https://api.deepseek.com/v1"
export OPENAI_API_KEY="sk-xxx"

You don't need DEEPSEEK_API_KEY since gpte uses Open AI API interface and DeepSeek supports that.

For more see DeepSeek docs.

Let me know if this resolves the issue.

[viborc]

Hey @ivan-vilches, can you let us know if @zigabrencic's comment solved your issue?

[viborc]

Hey @ivan-vilches, I'll be closing this issue now due to inactivity, but feel free to reopen if needed.