mmarch/text-generation-webui
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

Add mirostat parameters for llama.cpp (#2287)

Browse source
This commit is contained in:
oobabooga 2023-05-22 19:37:24 -03:00 • committed by GitHub
parent ec7437f00a
commit c0fd7f3257
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
13 changed files with 80 additions and 15 deletions
Download patch file Download diff file Expand all files Collapse all files

38
server.py
View file

@ -97,7 +97,11 @@ def load_preset_values(preset_menu, state, return_dict=False):
'length_penalty': 1,
'no_repeat_ngram_size': 0,
'early_stopping': False,
'mirostat_mode': 0,
'mirostat_tau': 5.0,
'mirostat_eta': 0.1,
}
with open(Path(f'presets/{preset_menu}.txt'), 'r') as infile:
preset = infile.read()
for i in preset.splitlines():
@ -110,7 +114,7 @@ def load_preset_values(preset_menu, state, return_dict=False):
return generate_params
else:
state.update(generate_params)
return state, *[generate_params[k] for k in ['do_sample', 'temperature', 'top_p', 'typical_p', 'epsilon_cutoff', 'eta_cutoff', 'repetition_penalty', 'encoder_repetition_penalty', 'top_k', 'min_length', 'no_repeat_ngram_size', 'num_beams', 'penalty_alpha', 'length_penalty', 'early_stopping']]
return state, *[generate_params[k] for k in ['do_sample', 'temperature', 'top_p', 'typical_p', 'epsilon_cutoff', 'eta_cutoff', 'repetition_penalty', 'encoder_repetition_penalty', 'top_k', 'min_length', 'no_repeat_ngram_size', 'num_beams', 'penalty_alpha', 'length_penalty', 'early_stopping', 'mirostat_mode', 'mirostat_tau', 'mirostat_eta']]
def upload_soft_prompt(file):
@ -434,27 +438,32 @@ def create_settings_menus(default_preset):
with gr.Row():
with gr.Column():
with gr.Row():
shared.gradio['preset_menu'] = gr.Dropdown(choices=utils.get_available_presets(), value=default_preset if not shared.args.flexgen else 'Naive', label='Generation parameters preset')
ui.create_refresh_button(shared.gradio['preset_menu'], lambda: None, lambda: {'choices': utils.get_available_presets()}, 'refresh-button')
with gr.Column():
shared.gradio['seed'] = gr.Number(value=shared.settings['seed'], label='Seed (-1 for random)')
with gr.Column():
with gr.Row():
shared.gradio['preset_menu'] = gr.Dropdown(choices=utils.get_available_presets(), value=default_preset if not shared.args.flexgen else 'Naive', label='Generation parameters preset')
ui.create_refresh_button(shared.gradio['preset_menu'], lambda: None, lambda: {'choices': utils.get_available_presets()}, 'refresh-button')
with gr.Column():
shared.gradio['seed'] = gr.Number(value=shared.settings['seed'], label='Seed (-1 for random)')
with gr.Row():
with gr.Column():
with gr.Box():
gr.Markdown('Main parameters ([click here to view technical documentation](https://huggingface.co/docs/transformers/main_classes/text_generation#transformers.GenerationConfig))')
gr.Markdown('Main parameters')
with gr.Row():
with gr.Column():
shared.gradio['temperature'] = gr.Slider(0.01, 1.99, value=generate_params['temperature'], step=0.01, label='temperature', info='Primary factor to control randomness of outputs. 0 = deterministic (only the most likely token is used). Higher value = more randomness.')
shared.gradio['top_p'] = gr.Slider(0.0, 1.0, value=generate_params['top_p'], step=0.01, label='top_p', info='If not set to 1, select tokens with probabilities adding up to less than this number. Higher value = higher range of possible random results.')
shared.gradio['top_k'] = gr.Slider(0, 200, value=generate_params['top_k'], step=1, label='top_k', info='Similar to top_p, but select instead only the top_k most likely tokens. Higher value = higher range of possible random results.')
shared.gradio['typical_p'] = gr.Slider(0.0, 1.0, value=generate_params['typical_p'], step=0.01, label='typical_p', info='If not set to 1, select only tokens that are at least this much more likely to appear than random tokens, given the prior text.')
shared.gradio['epsilon_cutoff'] = gr.Slider(0, 9, value=generate_params['epsilon_cutoff'], step=0.01, label='epsilon_cutoff', info='In units of 1e-4')
shared.gradio['eta_cutoff'] = gr.Slider(0, 20, value=generate_params['eta_cutoff'], step=0.01, label='eta_cutoff', info='In units of 1e-4')
with gr.Column():
shared.gradio['repetition_penalty'] = gr.Slider(1.0, 1.5, value=generate_params['repetition_penalty'], step=0.01, label='repetition_penalty', info='Exponential penalty factor for repeating prior tokens. 1 means no penalty, higher value = less repetition, lower value = more repetition.')
shared.gradio['encoder_repetition_penalty'] = gr.Slider(0.8, 1.5, value=generate_params['encoder_repetition_penalty'], step=0.01, label='encoder_repetition_penalty', info='Also known as the "Hallucinations filter". Used to penalize tokens that are *not* in the prior text. Higher value = more likely to stay in context, lower value = more likely to diverge.')
shared.gradio['no_repeat_ngram_size'] = gr.Slider(0, 20, step=1, value=generate_params['no_repeat_ngram_size'], label='no_repeat_ngram_size', info='If not set to 0, specifies the length of token sets that are completely blocked from repeating at all. Higher values = blocks larger phrases, lower values = blocks words or letters from repeating. Only 0 or high values are a good idea in most cases.')
shared.gradio['min_length'] = gr.Slider(0, 2000, step=1, value=generate_params['min_length'], label='min_length', info='Minimum generation length in tokens.')
shared.gradio['do_sample'] = gr.Checkbox(value=generate_params['do_sample'], label='do_sample')
shared.gradio['do_sample'] = gr.Checkbox(value=generate_params['do_sample'], label='do_sample')
with gr.Column():
with gr.Box():
with gr.Row():
@ -468,9 +477,12 @@ def create_settings_menus(default_preset):
shared.gradio['early_stopping'] = gr.Checkbox(value=generate_params['early_stopping'], label='early_stopping')
with gr.Column():
gr.Markdown('Mirostrat (for llama.cpp)')
shared.gradio['mirostat_mode'] = gr.Slider(0, 2, step=1, value=generate_params['mirostat_mode'], label='mirostat_mode')
shared.gradio['mirostat_tau'] = gr.Slider(0, 10, step=0.01, value=generate_params['mirostat_tau'], label='mirostat_tau')
shared.gradio['mirostat_eta'] = gr.Slider(0, 1, step=0.01, value=generate_params['mirostat_eta'], label='mirostat_eta')
gr.Markdown('Other')
shared.gradio['epsilon_cutoff'] = gr.Slider(0, 9, value=generate_params['epsilon_cutoff'], step=0.01, label='epsilon_cutoff', info='In units of 1e-4')
shared.gradio['eta_cutoff'] = gr.Slider(0, 20, value=generate_params['eta_cutoff'], step=0.01, label='eta_cutoff', info='In units of 1e-4')
with gr.Accordion('Soft prompt', open=False):
with gr.Row():
shared.gradio['softprompts_menu'] = gr.Dropdown(choices=utils.get_available_softprompts(), value='None', label='Soft prompt')
@ -492,7 +504,9 @@ def create_settings_menus(default_preset):
shared.gradio['skip_special_tokens'] = gr.Checkbox(value=shared.settings['skip_special_tokens'], label='Skip special tokens', info='Some specific models need this unset.')
shared.gradio['stream'] = gr.Checkbox(value=not shared.args.no_stream, label='Activate text streaming')
shared.gradio['preset_menu'].change(load_preset_values, [shared.gradio[k] for k in ['preset_menu', 'interface_state']], [shared.gradio[k] for k in ['interface_state', 'do_sample', 'temperature', 'top_p', 'typical_p', 'epsilon_cutoff', 'eta_cutoff', 'repetition_penalty', 'encoder_repetition_penalty', 'top_k', 'min_length', 'no_repeat_ngram_size', 'num_beams', 'penalty_alpha', 'length_penalty', 'early_stopping']])
gr.Markdown('[Click here for more information.](https://github.com/oobabooga/text-generation-webui/docs/Generation-parameters.md)')
shared.gradio['preset_menu'].change(load_preset_values, [shared.gradio[k] for k in ['preset_menu', 'interface_state']], [shared.gradio[k] for k in ['interface_state', 'do_sample', 'temperature', 'top_p', 'typical_p', 'epsilon_cutoff', 'eta_cutoff', 'repetition_penalty', 'encoder_repetition_penalty', 'top_k', 'min_length', 'no_repeat_ngram_size', 'num_beams', 'penalty_alpha', 'length_penalty', 'early_stopping', 'mirostat_mode', 'mirostat_tau', 'mirostat_eta']])
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'])