Improve the separation between instruct/chat modes (#1896)

This commit is contained in:
oobabooga 2023-05-07 23:47:02 -03:00 committed by GitHub
parent 9754d6a811
commit 56a5969658
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
3 changed files with 23 additions and 19 deletions

View file

@ -36,7 +36,7 @@ def generate_chat_prompt(user_input, state, **kwargs):
_continue = kwargs['_continue'] if '_continue' in kwargs else False
also_return_rows = kwargs['also_return_rows'] if 'also_return_rows' in kwargs else False
is_instruct = state['mode'] == 'instruct'
rows = [state['context'] if is_instruct else f"{state['context'].strip()}\n"]
rows = [state['context_instruct'] if is_instruct else f"{state['context'].strip()}\n"]
min_rows = 3
# Finding the maximum prompt size
@ -47,17 +47,17 @@ def generate_chat_prompt(user_input, state, **kwargs):
max_length = min(get_max_prompt_length(state), chat_prompt_size)
# Building the turn templates
if 'turn_template' not in state or state['turn_template'] == '':
if is_instruct:
if is_instruct:
if 'turn_template' not in state or state['turn_template'] == '':
template = '<|user|>\n<|user-message|>\n<|bot|>\n<|bot-message|>\n'
else:
template = '<|user|>: <|user-message|>\n<|bot|>: <|bot-message|>\n'
template = state['turn_template'].replace(r'\n', '\n')
else:
template = state['turn_template'].replace(r'\n', '\n')
template = '<|user|>: <|user-message|>\n<|bot|>: <|bot-message|>\n'
replacements = {
'<|user|>': state['name1'].strip(),
'<|bot|>': state['name2'].strip(),
'<|user|>': state['name1_instruct' if is_instruct else 'name1'].strip(),
'<|bot|>': state['name2_instruct' if is_instruct else 'name2'].strip(),
}
user_turn = replace_all(template.split('<|bot|>')[0], replacements)
@ -102,7 +102,7 @@ def generate_chat_prompt(user_input, state, **kwargs):
def get_stopping_strings(state):
if state['mode'] == 'instruct':
stopping_strings = [f"\n{state['name1']}", f"\n{state['name2']}"]
stopping_strings = [f"\n{state['name1_instruct']}", f"\n{state['name2_instruct']}"]
else:
stopping_strings = [f"\n{state['name1']}:", f"\n{state['name2']}:"]
@ -472,8 +472,6 @@ def load_character(character, name1, name2, mode):
if k in data and data[k] != '':
name1 = data[k]
break
else:
name1 = shared.settings['name1']
for field in ['context', 'greeting', 'example_dialogue', 'char_persona', 'char_greeting', 'world_scenario']:
if field in data: