From 81eec9ba70e2b6f754350bf91cbb265bc9d2b99e Mon Sep 17 00:00:00 2001 From: Qingru Zhang Date: Mon, 27 Feb 2023 21:08:55 -0500 Subject: [PATCH] train script --- .../peft_lora_seq2seq_accelerate_ds_zero3_offload.py | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/examples/conditional_generation/peft_lora_seq2seq_accelerate_ds_zero3_offload.py b/examples/conditional_generation/peft_lora_seq2seq_accelerate_ds_zero3_offload.py index cef9773..3c80bfb 100644 --- a/examples/conditional_generation/peft_lora_seq2seq_accelerate_ds_zero3_offload.py +++ b/examples/conditional_generation/peft_lora_seq2seq_accelerate_ds_zero3_offload.py @@ -102,7 +102,8 @@ class TorchTracemalloc: def main(): accelerator = Accelerator() - model_name_or_path = "bigscience/T0_3B" + # model_name_or_path = "bigscience/T0_3B" + model_name_or_path = "facebook/bart-large" dataset_name = "twitter_complaints" peft_config = LoraConfig( task_type=TaskType.SEQ_2_SEQ_LM, inference_mode=False, r=8, lora_alpha=32, lora_dropout=0.1