Refactor several function calls and the API
This commit is contained in:
oobabooga
GitHub
parent
378d21e80c
commit
3f3e42e26c
8 changed files with 147 additions and 118 deletions
51
server.py
51
server.py
|
|
@ -15,7 +15,7 @@ import gradio as gr
|
|||
from PIL import Image
|
||||
|
||||
import modules.extensions as extensions_module
|
||||
from modules import chat, shared, training, ui
|
||||
from modules import chat, shared, training, ui, api
|
||||
from modules.html_generator import chat_html_wrapper
|
||||
from modules.LoRA import add_lora_to_model
|
||||
from modules.models import load_model, load_soft_prompt
|
||||
|
|
@ -85,7 +85,7 @@ def load_lora_wrapper(selected_lora):
|
|||
add_lora_to_model(selected_lora)
|
||||
return selected_lora
|
||||
|
||||
def load_preset_values(preset_menu, return_dict=False):
|
||||
def load_preset_values(preset_menu, state, return_dict=False):
|
||||
generate_params = {
|
||||
'do_sample': True,
|
||||
'temperature': 1,
|
||||
|
|
@ -107,13 +107,13 @@ def load_preset_values(preset_menu, return_dict=False):
|
|||
i = i.rstrip(',').strip().split('=')
|
||||
if len(i) == 2 and i[0].strip() != 'tokens':
|
||||
generate_params[i[0].strip()] = eval(i[1].strip())
|
||||
|
||||
generate_params['temperature'] = min(1.99, generate_params['temperature'])
|
||||
|
||||
if return_dict:
|
||||
return generate_params
|
||||
else:
|
||||
return generate_params['do_sample'], generate_params['temperature'], generate_params['top_p'], generate_params['typical_p'], generate_params['repetition_penalty'], generate_params['encoder_repetition_penalty'], generate_params['top_k'], generate_params['min_length'], generate_params['no_repeat_ngram_size'], generate_params['num_beams'], generate_params['penalty_alpha'], generate_params['length_penalty'], generate_params['early_stopping']
|
||||
state.update(generate_params)
|
||||
return state, *[generate_params[k] for k in ['do_sample', 'temperature', 'top_p', 'typical_p', 'repetition_penalty', 'encoder_repetition_penalty', 'top_k', 'min_length', 'no_repeat_ngram_size', 'num_beams', 'penalty_alpha', 'length_penalty', 'early_stopping']]
|
||||
|
||||
def upload_soft_prompt(file):
|
||||
with zipfile.ZipFile(io.BytesIO(file)) as zf:
|
||||
|
|
@ -170,7 +170,10 @@ def create_prompt_menus():
|
|||
shared.gradio['save_prompt'].click(save_prompt, [shared.gradio['textbox']], [shared.gradio['status']], show_progress=False)
|
||||
|
||||
def create_settings_menus(default_preset):
|
||||
generate_params = load_preset_values(default_preset if not shared.args.flexgen else 'Naive', return_dict=True)
|
||||
generate_params = load_preset_values(default_preset if not shared.args.flexgen else 'Naive', {}, return_dict=True)
|
||||
for k in ['max_new_tokens', 'seed', 'stop_at_newline', 'chat_prompt_size', 'chat_generation_attempts']:
|
||||
generate_params[k] = shared.settings[k]
|
||||
shared.gradio['generate_state'] = gr.State(generate_params)
|
||||
|
||||
with gr.Row():
|
||||
with gr.Column():
|
||||
|
|
@ -221,17 +224,16 @@ def create_settings_menus(default_preset):
|
|||
with gr.Row():
|
||||
shared.gradio['upload_softprompt'] = gr.File(type='binary', file_types=['.zip'])
|
||||
|
||||
shared.gradio['model_menu'].change(load_model_wrapper, [shared.gradio['model_menu']], [shared.gradio['model_menu']], show_progress=True)
|
||||
shared.gradio['preset_menu'].change(load_preset_values, [shared.gradio['preset_menu']], [shared.gradio[k] for k in ['do_sample', 'temperature', 'top_p', 'typical_p', 'repetition_penalty', 'encoder_repetition_penalty', 'top_k', 'min_length', 'no_repeat_ngram_size', 'num_beams', 'penalty_alpha', 'length_penalty', 'early_stopping']])
|
||||
shared.gradio['lora_menu'].change(load_lora_wrapper, [shared.gradio['lora_menu']], [shared.gradio['lora_menu']], show_progress=True)
|
||||
shared.gradio['softprompts_menu'].change(load_soft_prompt, [shared.gradio['softprompts_menu']], [shared.gradio['softprompts_menu']], show_progress=True)
|
||||
shared.gradio['upload_softprompt'].upload(upload_soft_prompt, [shared.gradio['upload_softprompt']], [shared.gradio['softprompts_menu']])
|
||||
shared.gradio['model_menu'].change(load_model_wrapper, shared.gradio['model_menu'], shared.gradio['model_menu'], show_progress=True)
|
||||
shared.gradio['preset_menu'].change(load_preset_values, [shared.gradio[k] for k in ['preset_menu', 'generate_state']], [shared.gradio[k] for k in ['generate_state', 'do_sample', 'temperature', 'top_p', 'typical_p', 'repetition_penalty', 'encoder_repetition_penalty', 'top_k', 'min_length', 'no_repeat_ngram_size', 'num_beams', 'penalty_alpha', 'length_penalty', 'early_stopping']])
|
||||
shared.gradio['lora_menu'].change(load_lora_wrapper, shared.gradio['lora_menu'], shared.gradio['lora_menu'], show_progress=True)
|
||||
shared.gradio['softprompts_menu'].change(load_soft_prompt, shared.gradio['softprompts_menu'], shared.gradio['softprompts_menu'], show_progress=True)
|
||||
shared.gradio['upload_softprompt'].upload(upload_soft_prompt, shared.gradio['upload_softprompt'], shared.gradio['softprompts_menu'])
|
||||
|
||||
def set_interface_arguments(interface_mode, extensions, bool_active):
|
||||
modes = ["default", "notebook", "chat", "cai_chat"]
|
||||
cmd_list = vars(shared.args)
|
||||
bool_list = [k for k in cmd_list if type(cmd_list[k]) is bool and k not in modes]
|
||||
#int_list = [k for k in cmd_list if type(k) is int]
|
||||
|
||||
shared.args.extensions = extensions
|
||||
for k in modes[1:]:
|
||||
|
|
@ -372,11 +374,11 @@ def create_interface():
|
|||
shared.gradio['chat_prompt_size_slider'] = gr.Slider(minimum=shared.settings['chat_prompt_size_min'], maximum=shared.settings['chat_prompt_size_max'], step=1, label='Maximum prompt size in tokens', value=shared.settings['chat_prompt_size'])
|
||||
with gr.Column():
|
||||
shared.gradio['chat_generation_attempts'] = gr.Slider(minimum=shared.settings['chat_generation_attempts_min'], maximum=shared.settings['chat_generation_attempts_max'], value=shared.settings['chat_generation_attempts'], step=1, label='Generation attempts (for longer replies)')
|
||||
shared.gradio['check'] = gr.Checkbox(value=shared.settings['stop_at_newline'], label='Stop generating at new line character?')
|
||||
shared.gradio['stop_at_newline'] = gr.Checkbox(value=shared.settings['stop_at_newline'], label='Stop generating at new line character?')
|
||||
|
||||
create_settings_menus(default_preset)
|
||||
|
||||
shared.input_params = [shared.gradio[k] for k in ['Chat input', 'max_new_tokens', 'do_sample', 'temperature', 'top_p', 'typical_p', 'repetition_penalty', 'encoder_repetition_penalty', 'top_k', 'min_length', 'no_repeat_ngram_size', 'num_beams', 'penalty_alpha', 'length_penalty', 'early_stopping', 'seed', 'name1', 'name2', 'context', 'check', 'chat_prompt_size_slider', 'chat_generation_attempts', 'Chat mode', 'end_of_turn']]
|
||||
shared.input_params = [shared.gradio[k] for k in ['Chat input', 'generate_state', 'name1', 'name2', 'context', 'Chat mode', 'end_of_turn']]
|
||||
|
||||
def set_chat_input(textbox):
|
||||
return textbox, ""
|
||||
|
|
@ -456,9 +458,9 @@ def create_interface():
|
|||
with gr.Tab("Parameters", elem_id="parameters"):
|
||||
create_settings_menus(default_preset)
|
||||
|
||||
shared.input_params = [shared.gradio[k] for k in ['textbox', 'max_new_tokens', 'do_sample', 'temperature', 'top_p', 'typical_p', 'repetition_penalty', 'encoder_repetition_penalty', 'top_k', 'min_length', 'no_repeat_ngram_size', 'num_beams', 'penalty_alpha', 'length_penalty', 'early_stopping', 'seed']]
|
||||
shared.input_params = [shared.gradio[k] for k in ['textbox', 'generate_state']]
|
||||
output_params = [shared.gradio[k] for k in ['textbox', 'markdown', 'html']]
|
||||
gen_events.append(shared.gradio['Generate'].click(generate_reply, shared.input_params, output_params, show_progress=shared.args.no_stream, api_name='textgen'))
|
||||
gen_events.append(shared.gradio['Generate'].click(generate_reply, shared.input_params, output_params, show_progress=shared.args.no_stream))
|
||||
gen_events.append(shared.gradio['textbox'].submit(generate_reply, shared.input_params, output_params, show_progress=shared.args.no_stream))
|
||||
shared.gradio['Stop'].click(stop_everything_event, [], [], queue=False, cancels=gen_events if shared.args.no_stream else None)
|
||||
shared.gradio['interface'].load(None, None, None, _js=f"() => {{{ui.main_js}}}")
|
||||
|
|
@ -489,9 +491,9 @@ def create_interface():
|
|||
with gr.Tab("Parameters", elem_id="parameters"):
|
||||
create_settings_menus(default_preset)
|
||||
|
||||
shared.input_params = [shared.gradio[k] for k in ['textbox', 'max_new_tokens', 'do_sample', 'temperature', 'top_p', 'typical_p', 'repetition_penalty', 'encoder_repetition_penalty', 'top_k', 'min_length', 'no_repeat_ngram_size', 'num_beams', 'penalty_alpha', 'length_penalty', 'early_stopping', 'seed']]
|
||||
shared.input_params = [shared.gradio[k] for k in ['textbox', 'generate_state']]
|
||||
output_params = [shared.gradio[k] for k in ['output_textbox', 'markdown', 'html']]
|
||||
gen_events.append(shared.gradio['Generate'].click(generate_reply, shared.input_params, output_params, show_progress=shared.args.no_stream, api_name='textgen'))
|
||||
gen_events.append(shared.gradio['Generate'].click(generate_reply, shared.input_params, output_params, show_progress=shared.args.no_stream))
|
||||
gen_events.append(shared.gradio['textbox'].submit(generate_reply, shared.input_params, output_params, show_progress=shared.args.no_stream))
|
||||
gen_events.append(shared.gradio['Continue'].click(generate_reply, [shared.gradio['output_textbox']] + shared.input_params[1:], output_params, show_progress=shared.args.no_stream))
|
||||
shared.gradio['Stop'].click(stop_everything_event, [], [], queue=False, cancels=gen_events if shared.args.no_stream else None)
|
||||
|
|
@ -524,6 +526,21 @@ def create_interface():
|
|||
if shared.args.extensions is not None:
|
||||
extensions_module.create_extensions_block()
|
||||
|
||||
def change_dict_value(d, key, value):
|
||||
d[key] = value
|
||||
return d
|
||||
|
||||
for k in ['do_sample', 'temperature', 'top_p', 'typical_p', 'repetition_penalty', 'encoder_repetition_penalty', 'top_k', 'min_length', 'no_repeat_ngram_size', 'num_beams', 'penalty_alpha', 'length_penalty', 'early_stopping', 'max_new_tokens', 'seed', 'stop_at_newline', 'chat_prompt_size_slider', 'chat_generation_attempts']:
|
||||
if k not in shared.gradio:
|
||||
continue
|
||||
if type(shared.gradio[k]) in [gr.Checkbox, gr.Number]:
|
||||
shared.gradio[k].change(lambda state, value, copy=k: change_dict_value(state, copy, value), inputs=[shared.gradio['generate_state'], shared.gradio[k]], outputs=shared.gradio['generate_state'])
|
||||
else:
|
||||
shared.gradio[k].release(lambda state, value, copy=k: change_dict_value(state, copy, value), inputs=[shared.gradio['generate_state'], shared.gradio[k]], outputs=shared.gradio['generate_state'])
|
||||
|
||||
if not shared.is_chat():
|
||||
api.create_apis()
|
||||
|
||||
# Authentication
|
||||
auth = None
|
||||
if shared.args.gradio_auth_path is not None:
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue