CODEWITHSUNDEEP
google-cloud-platformgoogle-cloud-vertex-aigoogle-ai-platform
user3689720
user3689720

Reputation: 311

Vertex AI Create Tuned Model Internal Error With No Context

I am using the "Create a tuned model" feature in Google Vertex AI dashboard.

I've been able to upload my jsonl file, but when the tuning job runs, it just fails with no indication of what the error is:

google error

I can't find any logs or anything that give me any more information.

I've tried running the same via CLI, but having authentication issues there, so feeling stuck.

I've tried using a simplified mock jsonl file to be sure it's not my formatting that is breaking it. Here's my training file:

{ "input_text": "input1", "output_text": "output1" }
{ "input_text": "input2", "output_text": "output2" }
{ "input_text": "input3", "output_text": "output3" }
{ "input_text": "input4", "output_text": "output4" }
{ "input_text": "input5", "output_text": "output5" }
{ "input_text": "input6", "output_text": "output6" }
{ "input_text": "input7", "output_text": "output7" }
{ "input_text": "input8", "output_text": "output8" }
{ "input_text": "input9", "output_text": "output9" }
{ "input_text": "input10", "output_text": "output10" }

Is there something wrong with my file, or is there any known way to dig deeper into the error?

Any guidance or suggestions much appreciated.

Upvotes: 0

Views: 936

Answers (2)

Santosh Mallah
Santosh Mallah

Reputation: 11

Dataset format is change

{
  "systemInstruction": {
    "role": string,
    "parts": [
      {
        "text": string
      }
    ]
  },
  "contents": [
    {
      "role": string,
      "parts": [
        {
          // Union field data can be only one of the following:
          "text": string,
          "fileData": {
            "mimeType": string,
            "fileUri": string
          }
        }
      ]
    }
  ]
}

Refer document https://cloud.google.com/vertex-ai/generative-ai/docs/models/gemini-supervised-tuning-prepare

Upvotes: 0

user3689720
user3689720

Reputation: 311

In this case, the specific problem was that I did not have the 'Cloud Storage Admin' permission set on the account I was running the job with.

Once I added this, it worked. I hit another error next, but that's unrelated.

Upvotes: 2

Related Questions