class
Llama::Sampler::MinP
- Llama::Sampler::MinP
- Llama::Sampler::Base
- Reference
- Object
Overview
Min-P sampler
The Min-P sampler keeps tokens with probability >= P * max_probability. This is similar to Top-P but uses a minimum probability threshold relative to the most likely token, rather than a cumulative probability threshold.
Example:
sampler = Llama::Sampler::MinP.new(0.05, 1) # Keep tokens with prob >= 5% of max probDefined in:
llama/sampler/min_p.crConstructors
-
.new(p : Float32, min_keep : Int32 = 1)
Creates a new Min-P sampler
Instance methods inherited from class Llama::Sampler::Base
to_unsafe : Pointer(Llama::LibLlama::LlamaSampler)
to_unsafe
Constructor methods inherited from class Llama::Sampler::Base
new(handle : Pointer(LibLlama::LlamaSampler))
new
Constructor Detail
def self.new(p : Float32, min_keep : Int32 = 1)
#
Creates a new Min-P sampler
Parameters:
- p: The minimum probability threshold (0.0 to 1.0)
- min_keep: Minimum number of tokens to keep
Raises:
- Llama::Error if the sampler cannot be created