Module: RubyLLM::Providers::GPUStack::Transcription
- Included in:
- ChatCompletions
- Defined in:
- lib/ruby_llm/providers/gpustack/transcription.rb
Overview
GPUStack's vLLM transcription stream uses Chat Completions deltas.
Class Method Summary collapse
Class Method Details
.build_transcription_chunk(data) ⇒ Object
14 15 16 17 18 19 20 21 22 23 24 25 |
# File 'lib/ruby_llm/providers/gpustack/transcription.rb', line 14 def build_transcription_chunk(data) return super unless data.key?('choices') choice = data['choices'].first || {} finished = choice['finish_reason'] || data['choices'].empty? RubyLLM::TranscriptionChunk.new( type: finished ? RubyLLM::TranscriptionChunk::DONE : RubyLLM::TranscriptionChunk::DELTA, delta: choice.dig('delta', 'content'), raw: data ) end |
.stream_transcription(payload, model:) ⇒ Object
10 11 12 |
# File 'lib/ruby_llm/providers/gpustack/transcription.rb', line 10 def stream_transcription(payload, model:, &) super({ stream_include_usage: 'true' }.merge(payload), model:, &) end |