- cache_dir (olive.engine.EngineConfig attribute)
-
calibrate_method
-
calibration_sampling_size
-
checkpoint_path
- clean_cache (olive.engine.EngineConfig attribute)
- clean_evaluation_cache (olive.engine.EngineConfig attribute)
-
command line option
- activation_type, [1], [2]
- ActivationSymmetric, [1], [2], [3]
- AddQDQPairToWeight
- algorithms
- all_tensors_to_one_file, [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15]
- approach, [1], [2]
- backend, [1], [2]
- batch_size, [1], [2], [3], [4], [5], [6]
- calibrate_method, [1], [2]
- calibration_sampling_size, [1]
- checkpoint_path
- cpu_cores
- data_config, [1], [2], [3], [4], [5], [6]
- data_dir, [1], [2], [3], [4], [5], [6], [7]
- dataloader_func, [1], [2], [3], [4], [5], [6], [7]
- device, [1], [2], [3]
- disable_shape_infer
- do_validate
- domain, [1], [2]
- enable_htp
- EnableSubgraph, [1], [2]
- engine_config
- excluded_precisions, [1], [2]
- execution_mode_list
- execution_provider
- external_data_name, [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15]
- extra.Sigmoid.nnapi, [1], [2]
- extra_args, [1]
- extra_config
- extra_options, [1], [2], [3]
- extra_session_config
- float16
- force_fp32_ops
- ForceQuantizeNoInputCheck, [1], [2]
- gpus
- group
- hidden_size
- htp_socs
- input
- input_int32
- input_layouts
- input_names, [1]
- input_nodes
- input_shape
- input_shapes, [1]
- input_types, [1]
- inter_thread_num_list
- intra_thread_num_list
- io_bind
- keep_io_types, [1]
- logger
- MatMulConstBOnly, [1], [2]
- max_finite_val
- metric_func
- min_positive_val
- model_type
- modules_to_fuse
- multivariate
- no_repeat_ngram_size
- node_block_list
- nodes_to_exclude, [1], [2], [3]
- nodes_to_quantize, [1], [2], [3]
- num_epochs
- num_heads
- num_samples, [1]
- num_steps
- only_onnxruntime
- op_block_list, [1]
- op_types_to_quantize, [1], [2], [3]
- opt_level
- opt_level_list
- optimization_options
- optimize_model, [1], [2], [3]
- output_names
- output_nodes
- output_shapes
- per_channel, [1], [2]
- post
- pre
- providers_list
- ptl_data_module
- ptl_module
- qconfig_func
- quant_format, [1], [2], [3], [4]
- quant_level, [1], [2]
- quant_mode, [1], [2], [3]
- quant_preprocess, [1], [2], [3]
- recipes, [1], [2]
- reduce_range, [1], [2], [3], [4], [5]
- save_as_external_data, [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15]
- script_dir, [1], [2], [3], [4], [5], [6], [7], [8], [9], [10]
- seed, [1], [2]
- target_device
- target_opset, [1], [2], [3]
- tool_command
- tool_command_args
- train_batch_size
- train_data_dir
- train_dataloader_func
- training_loop_func
- trt_fp16_enable
- use_distributed_tuning, [1], [2]
- use_enhanced_quantizer
- use_external_data_format
- use_gpu
- user_script, [1], [2], [3], [4], [5], [6], [7], [8], [9], [10]
- val_data_dir
- weight_type, [1], [2], [3]
- WeightSymmetric, [1], [2], [3]
- with_replacement
|
|