rishiraj commited on
Commit
4a002be
·
1 Parent(s): 1a6dfeb

Create app.py

Browse files
Files changed (1) hide show
  1. app.py +19 -0
app.py ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ###############
2
+ # Load datasets
3
+ ###############
4
+ raw_datasets = get_datasets(data_args, splits=data_args.dataset_splits)
5
+ logger.info(
6
+ f"Training on the following datasets and their proportions: {[split + ' : ' + str(dset.num_rows) for split, dset in raw_datasets.items()]}"
7
+ )
8
+
9
+ ################
10
+ # Load tokenizer
11
+ ################
12
+ tokenizer = get_tokenizer(model_args, data_args)
13
+
14
+ #####################
15
+ # Apply chat template
16
+ #####################
17
+ raw_datasets = raw_datasets.map(apply_chat_template, fn_kwargs={"tokenizer": tokenizer, "task": "sft"})
18
+ train_dataset = raw_datasets["train"]
19
+ eval_dataset = raw_datasets["test"]