You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
ToolTipService.ToolTip="The top P parameter tells the model to only consider tokens up until a cumulative probability of P. Lower this value for more predictable, focused generation and increase it for more random response."
ToolTipService.ToolTip="The top K parameter tells the model to only consider the K most probable tokens at each generation step. Lower this value for more predictable, focused generation and increase it for more random response."
ToolTipService.ToolTip="The temperature parameter is a scaling factor for the probability distribution of tokens during generation. Values lower than 1 will produce more deterministic response while values higher than 1 will increase ramdomness."
ToolTipService.ToolTip="If Sampling is disabled, a greedy approach will be used and the model will select the most likely token every time. If enabled,tokens will be selected based on the token probability distribution. The Top K, Top P, and Temperature parameters only apply if sampling is enabled." />
0 commit comments