ptq.yaml

#
# Copyright (c) 2021 Intel Corporation
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#   http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

version: 1.0                                         # optional. reserved for future use. if not specified, a supported version would be written back to user yaml.

model:                                               # mandatory. used to specify model specific information.
  name: ssd_mobilenet_v1                             # mandatory. the model name.
  framework: tensorflow                              # mandatory. supported values are tensorflow, pytorch, pytorch_fx, pytorch_ipex, onnxrt_integer, onnxrt_qlinear or mxnet; allow new framework backend extension.
  inputs: image_tensor                               # optional. inputs and outputs fields are only required in tensorflow.
  outputs: num_detections,detection_boxes,detection_scores,detection_classes

device: cpu                                          # optional. default value is cpu. other value is gpu.

quantization:                                        # optional. tuning constraints on model-wise for advance user to reduce tuning space.
  approach: post_training_static_quant               # optional. default value is post_training_static_quant.
  recipes:                                           # optional. used to switch neural_compressor int8 receipts ON or OFF.
    scale_propagation_max_pooling: True              # optional. default value is True.
    scale_propagation_concat: True                   # optional. default value is True.
    first_conv_or_matmul_quantization: True          # optional. default value is True.
  calibration:                                       # optional. used to specify calibration behavior of post-training-static-quant. other quantization approachs are not necessary.
    sampling_size: 1000, 2000                        # optional. default value is 100. used to set how many samples should be used in calibration.
    dataloader:                                      # optional. if not specified, user need construct a calib_dataloader in code for neural_compressor.experimental.Quantization.
      dataset:                                       # optional. if not specified, user need construct a calibration dataset in code for calib_dataloader of neural_compressor.experimental.Quantization.
        TFRecordDataset:
          root: /path/to/tf_record
      transform:
        Resize:
          size: 256
        CenterCrop:
          size: 224
  model_wise:                                        # optional. tuning constraints on model-wise for advance user to reduce tuning space.
    weight:
      granularity: per_channel
      scheme: asym
      dtype: int8
      algorithm: minmax
    activation:
      granularity: per_tensor
      scheme: asym
      dtype: int8, fp32
      algorithm: minmax, kl
  op_wise: {                                         # optional. tuning constraints on op-wise for advance user to reduce tuning space. 
         'conv1': {
           'activation':  {'dtype': ['uint8', 'fp32'], 'algorithm': ['minmax', 'kl'], 'scheme':['sym']},
           'weight': {'dtype': ['int8', 'fp32'], 'algorithm': ['minmax']}
         },
         'pool1': {
           'activation': {'dtype': ['int8'], 'scheme': ['sym'], 'granularity': ['per_tensor'], 'algorithm': ['minmax', 'kl']},
         },
         'conv2': {
           'activation':  {'dtype': ['fp32']},
           'weight': {'dtype': ['fp32']}
         }
         'default_qconfig': {                       # optional. set default qconfig to fp32 for FX model
           'activation':  {'dtype': ['fp32']},
           'weight': {'dtype': ['fp32']}
         }
       }

evaluation:                                          # optional. used to config evaluation process.
  accuracy:                                          # optional. required if user doesn't provide eval_func in neural_compressor.Quantization.
    metric:                                          # optional. used to evaluate accuracy of passing model.
      topk: 1                                        # built-in metrics are topk, map, f1, allow user to register new metric.
    configs:                                         # optional. if not specified, use all cores in 1 socket.
      cores_per_instance: 28
      num_of_instance: 1
      inter_num_of_threads: 4
      intra_num_of_threads: 28
      kmp_blocktime: 1
    dataloader:                                      # optional. if not specified, user need construct a q_dataloader in code for neural_compressor.Quantization.
      batch_size: 256
      dataset:
        TFRecordDataset:
          root: /path/to/tf_record
      transform:
        Resize:
          size: 256
        CenterCrop:
          size: 224
  performance:                                       # optional. used to benchmark performance of passing model.
    warmup: 10
    iteration: 100
    configs:
      cores_per_instance: 4
      num_of_instance: 7
      inter_num_of_threads: 1
      intra_num_of_threads: 4
      kmp_blocktime: 1
    dataloader:
      dataset:
        dummy:
          shape: [[128, 3, 224, 224], [128, 1, 1, 1]]

tuning:
  strategy:
    name: basic                                      # optional. default value is basic. other values are bayesian, mse, sigopt.
    sigopt_api_token: YOUR-ACCOUNT-API-TOKEN             # optional. Necessary if strategy name is sigopt.
    sigopt_project_id: PROJECT-ID                    # optional. Necessary if strategy name is sigopt.
    sigopt_experiment_name: nc-tune                # optional. default is nc-tune if strategy name is sigopt.
  accuracy_criterion:
    relative:  0.01                                  # optional. default value is relative, other value is absolute. this example allows relative accuracy loss: 1%.
  objective: performance                             # optional. objective with accuracy constraint guaranteed. default value is performance. other values are modelsize and footprint.

  exit_policy:
    timeout: 0                                       # optional. tuning timeout (seconds). default value is 0 which means early stop. combine with max_trials field to decide when to exit.
    max_trials: 100                                  # optional. max tune times. default value is 100. combine with timeout field to decide when to exit.
    performance_only: False                          # optional. max tune times. default value is False which means only generate fully quantized model.
  random_seed: 9527                                  # optional. random seed for deterministic tuning.
  tensorboard: True                                  # optional. dump tensor distribution in evaluation phase for debug purpose. default value is False.

  workspace:
    path: /path/to/saving/directory                  # optional. default workspace is ./nc_workspace/current_time_stamp, saving tuning history and deploy yaml.
    resume: /path/to/a/specified/snapshot/file       # optional. if specified, resume from tuning history.