New model editions (GPT4)

[deep-diver]

Hi @tloen

I have trained the following models on GPT4 generated Alpaca dataset(from the one in this repo), and they are available through Hugging Face Model hub.

• chansung/gpt4-alpaca-lora-7b
• chansung/gpt4-alpaca-lora-13b
• chansung/gpt4-alpaca-lora-30b

You can also find out the link for the training logs on each Model repository.
I hope this might be useful for someone, and I also hope these could be included in the list in this repo.

[T-Atlas]

Hi @deep-diver
I tried using GPT-4 data to train the adapter myself, but I found compared to models trained with original data, adapter models trained with GPT-4 data will output instructions and inputs during generation.
python generate.py --load_8bit --base_model 'decapoda-research/llama-7b-hf' --lora_weights 'gpt4-alpaca- lora-7b'
I want to know if this is normal?

The following is an example of raw data training
python generate.py --load_8bit --base_model='decapoda-research/llama-7b-hf'

[deep-diver]

I think so. You need to trim to get after Response

[T-Atlas]

I think so. You need to trim to get after Response

But I think the format and prompt template of these two pieces of data are the same. Do you have any understanding of why there is such a difference?

[su-park]

Hello.
I am experiencing the same issue as the one @T-Atlas posted above.
I have prepared a benchmark set and compared the performance of Alpaca-7b with regards to the same prompt.
The instruction and inputs are attached to the generated output like an echo.

[JianqiaoLu]

I think so. You need to trim to get after Response

I think so. You need to trim to get after Response

But I think the format and prompt template of these two pieces of data are the same. Do you have any understanding of why there is such a difference?

It looks like the loss not only applies to model genereated output but also such template such as "instruction:" and "input：{input}"

[T-Atlas]

I think so. You need to trim to get after Response

I think so. You need to trim to get after Response

But I think the format and prompt template of these two pieces of data are the same. Do you have any understanding of why there is such a difference?

It looks like the loss not only applies to model genereated output but also such template such as "instruction:" and "input：{input}"

Sounds reasonable, do you have any attempts to correct it?

[JianqiaoLu]

I think so. You need to trim to get after Response

I think so. You need to trim to get after Response

But I think the format and prompt template of these two pieces of data are the same. Do you have any understanding of why there is such a difference?

It looks like the loss not only applies to model genereated output but also such template such as "instruction:" and "input：{input}"

Sounds reasonable, do you have any attempts to correct it?

The only way that comes to my mind is to re fine-tune the model and set labels of "instruction, input, etc" to -100.