Module: RubyLLM::Protocols::GPUStack::Tokenization

Included in:
RubyLLM::Providers::GPUStack::ChatCompletions
Defined in:
lib/ruby_llm/protocols/gpustack/tokenization.rb

Overview

vLLM's text tokenizer through GPUStack's model proxy.

Instance Method Summary collapse

Instance Method Details

#parse_tokenization_response(response, model:) ⇒ Object 



16
17
18
 
# File 'lib/ruby_llm/protocols/gpustack/tokenization.rb', line 16

def parse_tokenization_response(response, model:)
  RubyLLM::Tokenization.new(ids: response.body.fetch('tokens'), model: model, raw: response.body)
end

#render_tokenization_payload(text, model:) ⇒ Object 



12
13
14
 
# File 'lib/ruby_llm/protocols/gpustack/tokenization.rb', line 12

def render_tokenization_payload(text, model:)
  { model: model, prompt: text }
end

#tokenization_url ⇒ Object 



8
9
10
 
# File 'lib/ruby_llm/protocols/gpustack/tokenization.rb', line 8

def tokenization_url
  "#{@provider.backend_api_base}/tokenize"
end