Fine-tuning Babbage-002 Model Fails After Running for a Day

API
1

Hello everyone,

I’ve encountered an issue while trying to fine-tune a babbage-002 model. The process begins as expected—changing status from ‘pending’ to ‘running’—but eventually fails after about a day. Here’s the API response I get:

"data": [
    {
      "object": "fine_tuning.job",
      "id": "ftjob-**************",
      "model": "babbage-002",
      "created_at": 1693551811,
      "finished_at": 1693590642,
      "fine_tuned_model": null,
      "organization_id": "org-**************",
      "result_files": [],
      "status": "failed",
      "validation_file": null,
      "training_file": "file-**************",
      "hyperparameters": {
        "n_epochs": 1
      },
      "trained_tokens": null
    }

Using the old fine-tune models everything works as a charm.
Am I the only one experiencing this?

Thanks in advance.

1 Like
2

I’m having the same issue, were you able to solve it?

3

Other forum users have fine-tuned babbage-002 (and gotten different behavior than prior models, as one might expect). So it is not completely broken.

4

Nope, I have tried with multiple files, behaviour has stayed the same :confused:

5

Yesterday i was able to fine-tune a babbage-002 model. Funny thing is, fundamentally I did not change anything, so maybe some bug was fixed in the meantime.