@basetenlabs/client - v0.2.0
    Preparing search index...

    Interface ModelTRTLLMBuildConfiguration

    interface ModelTRTLLMBuildConfiguration {
        base_model?: ModelTRTLLMModel;
        checkpoint_repository?: CheckpointRepository | null;
        gather_all_token_logits?: boolean;
        lora_adapters?: { [k: string]: CheckpointRepository } | null;
        lora_configuration?: ModelTRTLLMLoraConfiguration | null;
        max_batch_size?: number;
        max_beam_width?: number;
        max_num_tokens?: number;
        max_prompt_embedding_table_size?: number;
        max_seq_len?: number | null;
        moe_expert_parallel_option?: number;
        num_builder_gpus?: number | null;
        pipeline_parallel_count?: number;
        plugin_configuration?: ModelTRTLLMPluginConfiguration;
        quantization_config?: ModelTRTQuantizationConfiguration;
        quantization_type?: ModelTRTLLMQuantizationType;
        sequence_parallel_count?: number;
        skip_build_result?: boolean;
        speculator?: ModelSpeculatorConfiguration | null;
        strongly_typed?: boolean;
        tensor_parallel_count?: number;
        [k: string]: unknown;
    }

    Indexable

    • [k: string]: unknown
    Index
    base_model?: ModelTRTLLMModel
    checkpoint_repository?: CheckpointRepository | null
    gather_all_token_logits?: boolean
    lora_adapters?: { [k: string]: CheckpointRepository } | null
    lora_configuration?: ModelTRTLLMLoraConfiguration | null
    max_batch_size?: number
    max_beam_width?: number
    max_num_tokens?: number
    max_prompt_embedding_table_size?: number
    max_seq_len?: number | null
    moe_expert_parallel_option?: number
    num_builder_gpus?: number | null
    pipeline_parallel_count?: number
    plugin_configuration?: ModelTRTLLMPluginConfiguration
    quantization_config?: ModelTRTQuantizationConfiguration
    quantization_type?: ModelTRTLLMQuantizationType
    sequence_parallel_count?: number
    skip_build_result?: boolean
    speculator?: ModelSpeculatorConfiguration | null
    strongly_typed?: boolean
    tensor_parallel_count?: number