[Python] GPT-3 fine tuning Error: Incorrect API key provided

I'm following this tutorial to fine-tune a GPT-3 model. However, when I run this part of the code:

# Enter credentials
%env OPENAI_API_KEY= "<MY OPENAI KEY>"
!openai api fine_tunes.create \
-t dw_train.jsonl \
-v dw_valid.jsonl \
-m $model \
--n_epochs $n_epochs \
--batch_size $batch_size \
--learning_rate_multiplier $learning_rate_multiplier \
--prompt_loss_weight $prompt_loss_weight


I get this error:

Error: Incorrect API key provided: "sk-czja*****************************************gk0". You can find your API key at https://beta.openai.com. (HTTP status code: 401)
​

The curious thing is that the API key is correct. So much so that, if I use it to make a prompt, it works perfectly. Example:

def GPT_Completion(texts):
response = openai.Completion.create(
engine="text-davinci-002",
prompt = texts,
temperature = 0.6,
top_p = 1,
max_tokens = 64,
frequency_penalty = 0,
presence_penalty = 0
)
return print(response.choices[0].text)

GPT_Completion("My dear friend,")


What could be causing this error? I thought maybe the GPT-3 training could require a paid account. However, I did not find this restriction on the OpenAI website.

The whole code I'm using is here.

Continue reading...