Index

_

__call__() (mii.backend.client.MIIClient method) (mii.batching.ragged_batching.MIIPipeline method)

C

client() (in module mii)

D

deployment_name (mii.config.MIIConfig attribute) deployment_type (mii.config.MIIConfig attribute)

device_map (mii.config.ModelConfig attribute) do_sample (mii.config.GenerateParamsConfig attribute)

E

enable_restful_api (mii.config.MIIConfig attribute)

F

finish_reason (mii.batching.data_classes.Response attribute)

G

generate() (mii.backend.client.MIIClient method) generate_replica_configs() (mii.config.MIIConfig method)

generated_length (mii.batching.data_classes.Response attribute) generated_text (mii.batching.data_classes.Response attribute) GenerationFinishReason (class in mii.constants)

H

hostfile (mii.config.MIIConfig attribute)

I

ignore_eos (mii.config.GenerateParamsConfig attribute)

inference_engine_config (mii.config.ModelConfig attribute) instance_type (mii.config.MIIConfig attribute)

L

LENGTH (mii.constants.GenerationFinishReason attribute)

M

max_length (mii.config.GenerateParamsConfig attribute) (mii.config.ModelConfig attribute) max_new_tokens (mii.config.GenerateParamsConfig attribute) MIIClient (class in mii.backend.client)

MIIPipeline (class in mii.batching.ragged_batching) min_new_tokens (mii.config.GenerateParamsConfig attribute) model_conf (mii.config.MIIConfig attribute) model_name_or_path (mii.config.ModelConfig attribute)

P

pipeline() (in module mii) port_number (mii.config.MIIConfig attribute)

profile_model_time (mii.config.ModelConfig attribute) prompt_length (mii.batching.data_classes.Response attribute) provider (mii.config.ModelConfig property)

Q

quantization_mode (mii.config.ModelConfig attribute)

R

replica_configs (mii.config.ModelConfig attribute) replica_num (mii.config.ModelConfig attribute) Response (class in mii.batching.data_classes)

restful_api_host (mii.config.MIIConfig attribute) restful_api_port (mii.config.MIIConfig attribute) restful_processes (mii.config.MIIConfig attribute) return_full_text (mii.config.GenerateParamsConfig attribute)

S

serve() (in module mii) stop (mii.config.GenerateParamsConfig attribute)

STOP (mii.constants.GenerationFinishReason attribute) stream (mii.config.GenerateParamsConfig attribute) sync_debug (mii.config.ModelConfig attribute)

T

task (mii.config.ModelConfig attribute) temperature (mii.config.GenerateParamsConfig attribute) tensor_parallel (mii.config.ModelConfig attribute) terminate_server() (mii.backend.client.MIIClient method)

tokenizer (mii.config.ModelConfig attribute) top_k (mii.config.GenerateParamsConfig attribute) top_p (mii.config.GenerateParamsConfig attribute) torch_dist_port (mii.config.ModelConfig attribute)

V

version (mii.config.MIIConfig attribute)

Z

zmq_port_number (mii.config.ModelConfig attribute)