Find-tune jobs always produce invalid JSON output & input

[mkultracreative]

Description of the bug:

From the ‘get code’ snip:

import os

import google.generativeai as genai

genai.configure(api_key=os.environ["GEMINI_API_KEY"])

# Create the model
# See https://ai.google.dev/api/python/google/generativeai/GenerativeModel
generation_config = {
  "temperature": 0.5,
  "top_p": 1,
  "max_output_tokens": 1000000 // doesn’t work with a fingernail nor centillion tokens that don’t get propery attention anyway 
  "response_mime_type": "text/plain",
}

model = genai.GenerativeModel(
  model_name="tunedModels/vercel-000000000", // this slash notation is also causing many apps that use no-code major issues. Just stick with a normal thing guys. Cmon. Fine tuned is redundant af
  generation_config=generation_config,
  # safety_settings = Adjust safety settings
  # See https://ai.google.dev/gemini-api/docs/safety-settings
)

response = model.generate_content([
  "input: “, // RIGHT HERE!
  "output: ", // AND HERE! Never seen this in my life. 
])

print(response.text)

I’ve tested on all frameworks including python and curl and Gemini entirely blocks their entire product from being used. Even with an example prompt, it provides a broken trailing output key value parameter:

// the rest is the same…
    {text: "input: why in the world would someone deliberately propagate two separate objects in a JSON file for one single input and output? Google hates me."},
    {text: "output: "},

Actual vs expected behavior:

Doesn’t make me pull my face off for literally 7 mo

Any other information you'd like to share?

That should be about it.