Fine-tuning don't work / bad results

Beginners
1

Hello,
I need your advice.

I’m trying to fine-tune Llama2 7B (base Model) for a single task. My goal is the following: for a given text it should replace some words by others if they are present in the text, otherwise it should only return the text.

However, when I make the inference, the task is not understood, the model simply repeats the instruction and returns a bit of nonsense.

I used the huggingface trainer and formulated my prompt as ###INSTRUCTION: {} ###INPUT: {} ###RESPONSE: {}.

I used the basic Lora parameters found in all the tutorials. I have about 2000 prompt data.

Does anyone have any advice or explanation for my problem?

Thanks for your help.

2

Facing the same issue here.

3

Same issue here.

4

You can use fine-tune-ai dot com

5

why does this thread have 1.3k views and no solutions? are we finetuning it badly?

6

I am facing the same issue and find this site, maybe it could be beneficial for you. Top 7 Large Language Model Fine-tuning Mistakes to Avoid | Entry Point AI