解决数据bug

This commit is contained in:
997146918 2025-08-08 10:16:38 +08:00
parent 1b5c8d64b0
commit 9d081893aa

View File

@ -50,7 +50,8 @@ def create_character_dialogue_samples(character:str, dialogues: List[Dict]) ->Li
tempDialogue = dialogue['dialogue'] tempDialogue = dialogue['dialogue']
elif tempDialogue != '': elif tempDialogue != '':
#确定是提问对话 #确定是提问对话
character_samples.append({"instruction": tempDialogue, character_samples.append({"character": character,
"instruction": tempDialogue,
"input": "", "input": "",
"output": dialogue['dialogue']}) "output": dialogue['dialogue']})
tempDialogue = '' tempDialogue = ''
@ -65,9 +66,9 @@ def main():
for char, convs in character_dialogues.items(): for char, convs in character_dialogues.items():
print(f" {char}: {len(convs)}条对话") print(f" {char}: {len(convs)}条对话")
#获得最终训练数据 #获得最终训练数据
final_samples = {} final_samples = []
for character in characters: for character in characters:
final_samples[character] = create_character_dialogue_samples(character, dialogues) final_samples += create_character_dialogue_samples(character, dialogues)