SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

[Peft] modules_to_save support for peft integration (#27466)

Browse Source 
* `modules_to_save` support for peft integration

* Update docs/source/en/peft.md

Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* slightly elaborate test

---------

Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
This commit is contained in:
Younes Belkada
2023-11-14 10:32:57 +01:00
committed by GitHub
parent 721d1c8ca6
commit d71fa9f618
3 changed files with 63 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
docs/source/en/peft.md
View File

@@ -98,7 +98,7 @@ You can use [`~peft.PeftModel.add_adapter`] to add a new adapter to a model with
```py
from transformers import AutoModelForCausalLM, OPTForCausalLM, AutoTokenizer
from peft import PeftConfig
from peft import LoraConfig
model_id = "facebook/opt-350m"
model = AutoModelForCausalLM.from_pretrained(model_id)
@@ -208,6 +208,26 @@ model.save_pretrained(save_dir)
model = AutoModelForCausalLM.from_pretrained(save_dir)
```
## Add additional trainable layers to a PEFT adapter
You can also fine-tune additional trainable adapters on top of a model that has adapters attached by passing `modules_to_save` in your PEFT config. For example, if you want to also fine-tune the lm_head on top of a model with a LoRA adapter:
```py
from transformers import AutoModelForCausalLM, OPTForCausalLM, AutoTokenizer
from peft import LoraConfig
model_id = "facebook/opt-350m"
model = AutoModelForCausalLM.from_pretrained(model_id)
lora_config = LoraConfig(
target_modules=["q_proj", "k_proj"],
modules_to_save=["lm_head"],
)
model.add_adapter(lora_config)
```
<!--
TODO: (@younesbelkada @stevhliu)
- Link to PEFT docs for further details