cpt core 4

scripts/cpt_core_model_4.py

@@ -4,6 +4,7 @@ from unsloth import FastLanguageModel
 import torch
 from transformers import AutoTokenizer
 
+
 os.environ['WANDB_PROJECT'] = 'tangled-alpha-0.9-core'
 run_name = 'cpt-core-4'
 
@@ -15,6 +16,7 @@ load_in_4bit = False
 model_name = '../out/pretrain-core-3/hf'
 output_dir = '../out/cpt-core-4'
 
+
 #
 # model
 #
@@ -53,6 +55,7 @@ model = FastLanguageModel.get_peft_model(
 )
 # print(f'{model=}')
 
+
 #
 # dataset
 #
@@ -76,6 +79,7 @@ def unlsoth_generator():
 train_dataset = Dataset.from_generator(unlsoth_generator)
 dataset = train_dataset.train_test_split(test_size=0.01)
 
+
 #
 # trainer
 #