how to fine tuning model with with dataset (file json/csv..)

[luongnv275]

[jorditost]

+1

is it possible?

[strikoder]

I'm interested as well...

[ianscrivener]

IMO - yes it is possible with prompt tuning

The next questions (which I am not so clear on) are;
(1) exactly how to achieve that (a specific use case and dataset would be required)
(2) how good/accurate the results would be....

Of course the tuning dataset would need to fit into context

[sb-child]

hi, i just found this article about fine tuning model with llama.cpp

This is an example of dataset format in appendix B of that article:

<s>Below is an instruction that describes a task. Write a response that appropriately completes the request.

### Instruction:
{prompt}

### Response:
{response}

And add a --sample-start "<s>" to the training command

[redoC-A2k]

Hey there did you tried fine tuning ? I want to fine tune this model https://huggingface.co/TheBloke/TinyLlama-1.1B-Chat-v1.0-GGUF .
with finetune utiltiy and wondering about how to make th training txt file .
If you could help that will be very helpful