Instructions to use philschmid/gemma-7b-dolly-chatml with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use philschmid/gemma-7b-dolly-chatml with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("google/gemma-7b") model = PeftModel.from_pretrained(base_model, "philschmid/gemma-7b-dolly-chatml") - Notebooks
- Google Colab
- Kaggle
Update trl-lora.py
Browse files- trl-lora.py +1 -1
trl-lora.py
CHANGED
|
@@ -64,7 +64,7 @@ trainer = SFTTrainer(
|
|
| 64 |
tokenizer=tokenizer,
|
| 65 |
packing=True,
|
| 66 |
dataset_kwargs={
|
| 67 |
-
"add_special_tokens":
|
| 68 |
"append_concat_token": False, # make sure to not add additional tokens when packing
|
| 69 |
}
|
| 70 |
)
|
|
|
|
| 64 |
tokenizer=tokenizer,
|
| 65 |
packing=True,
|
| 66 |
dataset_kwargs={
|
| 67 |
+
"add_special_tokens": False, # <bos> and <eos> should be part of the dataset.
|
| 68 |
"append_concat_token": False, # make sure to not add additional tokens when packing
|
| 69 |
}
|
| 70 |
)
|