423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-12-17 04:10:36 +08:00
Code Issues Packages Projects Releases Wiki Activity

add template, modify datasets

Browse Source
This commit is contained in:
hiyouga
2023-11-09 15:53:23 +08:00
parent 7ca32d8e69
commit 386f590209
7 changed files with 65 additions and 38 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
src/llmtuner/dsets/preprocess.py
View File

@@ -51,9 +51,6 @@ def preprocess_dataset(
setattr(tokenizer, "add_eos_token", True)
tokenized_examples = tokenizer(examples["prompt"], **kwargs)
# Make sure the saved tokenizer is the same as the original
if hasattr(tokenizer, "add_eos_token"): # for Baichuan2 tokenizer
setattr(tokenizer, "add_eos_token", add_eos_token_flag)
concatenated_examples = {k: list(chain(*tokenized_examples[k])) for k in tokenized_examples.keys()}
total_length = len(concatenated_examples[list(concatenated_examples.keys())[0]])
block_size = data_args.cutoff_len
@@ -64,6 +61,9 @@ def preprocess_dataset(
k: [t[i: i + block_size] for i in range(0, total_length, block_size)]
for k, t in concatenated_examples.items()
}
# make sure the saved tokenizer is the same as the original one
if hasattr(tokenizer, "add_eos_token"):
setattr(tokenizer, "add_eos_token", add_eos_token_flag)
return result
def preprocess_supervised_dataset(examples: Dict[str, List[Any]]) -> Dict[str, List[List[int]]]: