gendataGUI
Attributes
Functions
|
Generates text based on a given prompt using an existing pre-trained model. |
|
Generates text using a Monte Carlo aligned model, adjusting output based on lambda-weighted rewards. |
Module Contents
- gendataGUI.device
- gendataGUI.generateGUI_from_original_model(prompt, temperature=1.0, top_k=50, num_beams=1, max_new_tokens=50)
Generates text based on a given prompt using an existing pre-trained model.
- Parameters:
prompt (str) – Input text prompt to generate continuation.
temperature (float, optional) – Sampling temperature for text generation. Defaults to 1.0.
top_k (int, optional) – Limits the number of high-probability tokens to sample from. Defaults to 50.
num_beams (int, optional) – Number of beams for beam search. Defaults to 1.
max_new_tokens (int, optional) – Maximum number of new tokens to generate. Defaults to 50.
- Returns:
Generated text based on the input prompt.
- Return type:
str
- gendataGUI.generateGUI_from_MC_aligned_model(prompt, model_choice, value_list, lam_list, MC_nsamples, temperature, top_k, num_beams, max_new_tokens)
Generates text using a Monte Carlo aligned model, adjusting output based on lambda-weighted rewards.
- Parameters:
prompt (str) – Input text prompt for generation.
model_choice (str) – Indicates whether to use the original or aligned model.
value_list (str) – Comma-separated values indicating alignment criteria.
lam_list (str) – Comma-separated lambda values for reward weighting.
MC_nsamples (int, optional) – Number of Monte Carlo samples for each prompt.
temperature (float, optional) – Sampling temperature for text generation.
top_k (int, optional) – Limits the number of high-probability tokens to sample from.
num_beams (int, optional) – Number of beams for beam search.
max_new_tokens (int, optional) – Maximum number of new tokens to generate.
- Returns:
Selected generated text based on alignment.
- Return type:
str
- gendataGUI.iface