Closed
Description
This line here ensures that Temperature values are in range [0,2]
temperature_field = Field(
default=0.8,
ge=0.0,
le=2.0,
description="Adjust the randomness of the generated text.\n\n"
+ "Temperature is a hyperparameter that controls the randomness of the generated text. It affects the probability distribution of the model's output tokens. A higher temperature (e.g., 1.5) makes the output more random and creative, while a lower temperature (e.g., 0.5) makes the output more focused, deterministic, and conservative. The default value is 0.8, which provides a balance between randomness and determinism. At the extreme, a temperature of 0 will always pick the most likely next token, leading to identical outputs in each run.",
)
At the same time temp values less than 0 are expected in _LlamaSamplingContext.
if self.params.temp < 0:
ctx_main.sample_softmax(token_data_array)
id = token_data_array.candidates_data["id"][0]
I suppose that maximum reproducibility of generation results is only possible with temp<0 values, and at the same time I would like to run OpenAI compatible web server for convenience.
abetlen
Metadata
Metadata
Assignees
Labels
New feature or requestNew feature or request