def get_config(is_train): class General: log_frequency = 10 name = __name__.rsplit("/")[-1].rsplit(".")[-1] batch_image = 2 if is_train else 1 fp16 = False class KvstoreParam: kvstore = "nccl" batch_image = General.batch_image gpus = [0, 1, 2, 3, 4, 5, 6, 7] fp16 = General.fp16 class NormalizeParam: normalizer = normalizer_factory(type="fixbn") class BackboneParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer class NeckParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer class RpnParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer batch_image = General.batch_image class anchor_generate: scale = (8, ) ratio = (0.5, 1.0, 2.0) stride = (4, 8, 16, 32, 64) image_anchor = 256 class head: conv_channel = 256 mean = (0, 0, 0, 0) std = (1, 1, 1, 1) class proposal: pre_nms_top_n = 2000 if is_train else 1000 post_nms_top_n = 2000 if is_train else 1000 nms_thr = 0.7 min_bbox_side = 0 class subsample_proposal: proposal_wo_gt = False image_roi = 512 fg_fraction = 0.25 fg_thr = 0.5 bg_thr_hi = 0.5 bg_thr_lo = 0.0 class bbox_target: num_reg_class = 81 class_agnostic = False weight = (1.0, 1.0, 1.0, 1.0) mean = (0.0, 0.0, 0.0, 0.0) std = (0.1, 0.1, 0.2, 0.2) class BboxParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer num_class = 1 + 80 image_roi = 512 batch_image = General.batch_image class regress_target: class_agnostic = False mean = (0.0, 0.0, 0.0, 0.0) std = (0.1, 0.1, 0.2, 0.2) class RoiParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer out_size = 7 stride = (4, 8, 16, 32) roi_canonical_scale = 224 roi_canonical_level = 4 class DatasetParam: if is_train: image_set = ("coco_train2014", "coco_valminusminival2014") else: image_set = ("coco_minival2014", ) backbone = Backbone(BackboneParam) neck = Neck(NeckParam) rpn_head = RpnHead(RpnParam) roi_extractor = RoiExtractor(RoiParam) bbox_head = BboxHead(BboxParam) detector = Detector() if is_train: train_sym = detector.get_train_symbol(backbone, neck, rpn_head, roi_extractor, bbox_head) rpn_test_sym = None test_sym = None else: train_sym = None rpn_test_sym = detector.get_rpn_test_symbol(backbone, neck, rpn_head) test_sym = detector.get_test_symbol(backbone, neck, rpn_head, roi_extractor, bbox_head) class ModelParam: train_symbol = train_sym test_symbol = test_sym rpn_test_symbol = rpn_test_sym from_scratch = False random = True memonger = False memonger_until = "stage3_unit21_plus" class pretrain: prefix = "pretrain_model/resnet-v1-101" epoch = 0 fixed_param = ["conv0", "stage1", "gamma", "beta"] class OptimizeParam: class optimizer: type = "sgd" lr = 0.01 / 8 * len(KvstoreParam.gpus) * KvstoreParam.batch_image momentum = 0.9 wd = 0.0001 clip_gradient = None class schedule: begin_epoch = 0 end_epoch = 6 lr_iter = [ 60000 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image), 80000 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image) ] class warmup: type = "gradual" lr = 0.01 / 8 * len( KvstoreParam.gpus) * KvstoreParam.batch_image / 3.0 iter = 500 class TestParam: min_det_score = 0.05 max_det_per_image = 100 process_roidb = lambda x: x process_output = lambda x, y: x class model: prefix = "experiments/{}/checkpoint".format(General.name) epoch = OptimizeParam.schedule.end_epoch class nms: type = "nms" thr = 0.5 class coco: annotation = "data/coco/annotations/instances_minival2014.json" # data processing class NormParam: mean = (122.7717, 115.9465, 102.9801) # RGB order std = (1.0, 1.0, 1.0) # data processing class ResizeParam: short = 800 long = 1333 class PadParam: short = 800 long = 1333 max_num_gt = 100 class AnchorTarget2DParam: def __init__(self): self.generate = self._generate() class _generate: def __init__(self): self.stride = (4, 8, 16, 32, 64) self.short = (200, 100, 50, 25, 13) self.long = (334, 167, 84, 42, 21) scales = (8) aspects = (0.5, 1.0, 2.0) class assign: allowed_border = 0 pos_thr = 0.7 neg_thr = 0.3 min_pos_thr = 0.0 class sample: image_anchor = 256 pos_fraction = 0.5 class RenameParam: mapping = dict(image="data") from core.detection_input import ReadRoiRecord, Resize2DImageBbox, \ ConvertImageFromHwcToChw, Flip2DImageBbox, Pad2DImageBbox, \ RenameRecord, Norm2DImage from models.FPN.input import PyramidAnchorTarget2D if is_train: transform = [ ReadRoiRecord(None), Norm2DImage(NormParam), Resize2DImageBbox(ResizeParam), Flip2DImageBbox(), Pad2DImageBbox(PadParam), ConvertImageFromHwcToChw(), PyramidAnchorTarget2D(AnchorTarget2DParam()), RenameRecord(RenameParam.mapping) ] data_name = ["data", "im_info", "gt_bbox"] label_name = ["rpn_cls_label", "rpn_reg_target", "rpn_reg_weight"] else: transform = [ ReadRoiRecord(None), Norm2DImage(NormParam), Resize2DImageBbox(ResizeParam), ConvertImageFromHwcToChw(), RenameRecord(RenameParam.mapping) ] data_name = ["data", "im_info", "im_id", "rec_id"] label_name = [] import core.detection_metric as metric rpn_acc_metric = metric.AccWithIgnore("RpnAcc", ["rpn_cls_loss_output"], ["rpn_cls_label"]) rpn_l1_metric = metric.L1("RpnL1", ["rpn_reg_loss_output"], ["rpn_cls_label"]) # for bbox, the label is generated in network so it is an output box_acc_metric = metric.AccWithIgnore( "RcnnAcc", ["bbox_cls_loss_output", "bbox_label_blockgrad_output"], []) box_l1_metric = metric.L1( "RcnnL1", ["bbox_reg_loss_output", "bbox_label_blockgrad_output"], []) metric_list = [ rpn_acc_metric, rpn_l1_metric, box_acc_metric, box_l1_metric ] return General, KvstoreParam, RpnParam, RoiParam, BboxParam, DatasetParam, \ ModelParam, OptimizeParam, TestParam, \ transform, data_name, label_name, metric_list
def get_config(is_train): class General: log_frequency = 10 name = __name__.rsplit("/")[-1].rsplit(".")[-1] batch_image = 2 if is_train else 1 fp16 = False class KvstoreParam: kvstore = "local" batch_image = General.batch_image gpus = [0, 1, 2, 3, 4, 5, 6, 7] fp16 = General.fp16 class NormalizeParam: # normalizer = normalizer_factory(type="syncbn", ndev=len(KvstoreParam.gpus)) normalizer = normalizer_factory(type="fixbn") class BackboneParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer class NeckParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer class RpnParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer batch_image = General.batch_image class anchor_generate: scale = (2, 4, 8, 16, 32) ratio = (0.5, 1.0, 2.0) stride = 16 image_anchor = 256 class head: conv_channel = 512 mean = (0, 0, 0, 0) std = (1, 1, 1, 1) class proposal: pre_nms_top_n = 12000 if is_train else 6000 post_nms_top_n = 2000 if is_train else 1000 nms_thr = 0.7 min_bbox_side = 0 class subsample_proposal: proposal_wo_gt = True image_roi = 256 fg_fraction = 0.25 fg_thr = 0.5 bg_thr_hi = 0.5 bg_thr_lo = 0.0 class bbox_target: num_reg_class = 2 class_agnostic = True weight = (1.0, 1.0, 1.0, 1.0) mean = (0.0, 0.0, 0.0, 0.0) std = (0.1, 0.1, 0.2, 0.2) class BboxParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer num_class = 1 + 80 image_roi = 256 batch_image = General.batch_image class regress_target: class_agnostic = True mean = (0.0, 0.0, 0.0, 0.0) std = (0.1, 0.1, 0.2, 0.2) class RoiParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer out_size = 7 stride = 16 class DatasetParam: if is_train: image_set = ("coco_train2014", "coco_valminusminival2014") else: image_set = ("coco_minival2014", ) backbone = Backbone(BackboneParam) neck = Neck(NeckParam) rpn_head = RpnHead(RpnParam) roi_extractor = RoiExtractor(RoiParam) bbox_head = BboxHead(BboxParam) detector = Detector() if is_train: train_sym = detector.get_train_symbol(backbone, neck, rpn_head, roi_extractor, bbox_head) test_sym = None else: train_sym = None test_sym = detector.get_test_symbol(backbone, neck, rpn_head, roi_extractor, bbox_head) class ModelParam: train_symbol = train_sym test_symbol = test_sym from_scratch = False random = True memonger = False memonger_until = "stage3_unit21_plus" class pretrain: prefix = "pretrain_model/resnet-50" epoch = 0 fixed_param = ["conv0", "stage1", "gamma", "beta"] class OptimizeParam: class optimizer: type = "sgd" lr = 0.01 / 8 * len(KvstoreParam.gpus) * KvstoreParam.batch_image momentum = 0.9 wd = 0.0001 clip_gradient = 35 class schedule: begin_epoch = 0 end_epoch = 12 lr_iter = [ 120000 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image), 160000 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image) ] class warmup: type = "gradual" lr = 0.0 iter = 3000 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image) class TestScaleParam: short_ranges = [600, 800, 1000, 1200] long_ranges = [2000, 2000, 2000, 2000] @staticmethod def add_resize_info(roidb): ms_roidb = [] for r_ in roidb: for short, long in zip(TestScaleParam.short_ranges, TestScaleParam.long_ranges): r = r_.copy() r["resize_long"] = long r["resize_short"] = short ms_roidb.append(r) return ms_roidb class TestParam: min_det_score = 0.001 max_det_per_image = 0 process_roidb = TestScaleParam.add_resize_info process_output = lambda x, y: x class model: prefix = "experiments/{}/checkpoint".format(General.name) epoch = OptimizeParam.schedule.end_epoch class nms: type = "nms" thr = 0.5 class coco: annotation = "data/coco/annotations/instances_minival2014.json" # data processing class ResizeParam: short = 800 long = 1200 if is_train else 2000 class RandResizeParam: short = None # generate on the fly long = None short_ranges = [600, 800, 1000, 1200] long_ranges = [2000, 2000, 2000, 2000] class RandCropParam: mode = "center" # random or center short = 800 long = 1200 class PadParam: short = 800 long = 1200 if is_train else 2000 max_num_gt = 100 class AnchorTarget2DParam: class generate: short = 800 // 16 long = 1200 // 16 stride = 16 scales = (2, 4, 8, 16, 32) aspects = (0.5, 1.0, 2.0) class assign: allowed_border = 0 pos_thr = 0.7 neg_thr = 0.3 min_pos_thr = 0.0 class sample: image_anchor = 256 pos_fraction = 0.5 class RenameParam: mapping = dict(image="data") from core.detection_input import ReadRoiRecord, RandResize2DImageBbox, RandCrop2DImageBbox, \ Resize2DImageBboxByRoidb, ConvertImageFromHwcToChw, Flip2DImageBbox, Pad2DImageBbox, \ RenameRecord, AnchorTarget2D if is_train: transform = [ ReadRoiRecord(None), RandResize2DImageBbox(RandResizeParam), RandCrop2DImageBbox(RandCropParam), Flip2DImageBbox(), Pad2DImageBbox(PadParam), ConvertImageFromHwcToChw(), AnchorTarget2D(AnchorTarget2DParam), RenameRecord(RenameParam.mapping) ] data_name = ["data", "im_info", "gt_bbox"] label_name = ["rpn_cls_label", "rpn_reg_target", "rpn_reg_weight"] else: transform = [ ReadRoiRecord(None), #Resize2DImageBbox(ResizeParam), Resize2DImageBboxByRoidb(), ConvertImageFromHwcToChw(), RenameRecord(RenameParam.mapping) ] data_name = ["data", "im_info", "im_id", "rec_id"] label_name = [] import core.detection_metric as metric rpn_acc_metric = metric.AccWithIgnore("RpnAcc", ["rpn_cls_loss_output"], ["rpn_cls_label"]) rpn_l1_metric = metric.L1("RpnL1", ["rpn_reg_loss_output"], ["rpn_cls_label"]) # for bbox, the label is generated in network so it is an output box_acc_metric = metric.AccWithIgnore( "RcnnAcc", ["bbox_cls_loss_output", "bbox_label_blockgrad_output"], []) box_l1_metric = metric.L1( "RcnnL1", ["bbox_reg_loss_output", "bbox_label_blockgrad_output"], []) metric_list = [ rpn_acc_metric, rpn_l1_metric, box_acc_metric, box_l1_metric ] return General, KvstoreParam, RpnParam, RoiParam, BboxParam, DatasetParam, \ ModelParam, OptimizeParam, TestParam, \ transform, data_name, label_name, metric_list
def get_config(is_train): class General: log_frequency = 10 name = __name__.rsplit("/")[-1].rsplit(".")[-1] batch_image = 8 if is_train else 1 fp16 = True loader_worker = 8 class KvstoreParam: kvstore = "nccl" batch_image = General.batch_image gpus = [0, 1, 2, 3, 4, 5, 6, 7] fp16 = General.fp16 class NormalizeParam: normalizer = normalizer_factory(type="localbn", ndev=len(KvstoreParam.gpus)) # normalizer = normalizer_factory(type="gn") class BackboneParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer class NeckParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer class RpnParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer batch_image = General.batch_image nnvm_proposal = True nnvm_rpn_target = False class anchor_generate: scale = (4, ) ratio = (0.5, 1.0, 2.0) stride = (4, 8, 16, 32, 64) image_anchor = 256 max_side = 700 class anchor_assign: allowed_border = 0 pos_thr = 0.7 neg_thr = 0.3 min_pos_thr = 0.0 image_anchor = 256 pos_fraction = 0.5 class head: conv_channel = 256 mean = (0, 0, 0, 0) std = (1, 1, 1, 1) class proposal: pre_nms_top_n = 2000 if is_train else 1000 post_nms_top_n = 2000 if is_train else 1000 nms_thr = 0.7 min_bbox_side = 0 class subsample_proposal: proposal_wo_gt = False image_roi = 512 fg_fraction = 0.25 fg_thr = 0.5 bg_thr_hi = 0.5 bg_thr_lo = 0.0 class bbox_target: num_reg_class = 81 class_agnostic = False weight = (1.0, 1.0, 1.0, 1.0) mean = (0.0, 0.0, 0.0, 0.0) std = (0.1, 0.1, 0.2, 0.2) class BboxParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer num_class = 1 + 80 image_roi = 512 batch_image = General.batch_image class regress_target: class_agnostic = False mean = (0.0, 0.0, 0.0, 0.0) std = (0.1, 0.1, 0.2, 0.2) class RoiParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer out_size = 7 stride = (4, 8, 16, 32) roi_canonical_scale = 224 roi_canonical_level = 4 class DatasetParam: if is_train: image_set = ("coco_train2017", ) total_image = 82783 + 35504 else: image_set = ("coco_val2017", ) total_image = 5000 backbone = Backbone(BackboneParam) neck = Neck(NeckParam) rpn_head = RpnHead(RpnParam) roi_extractor = RoiExtractor(RoiParam) bbox_head = BboxHead(BboxParam) detector = Detector() if is_train: train_sym = detector.get_train_symbol(backbone, neck, rpn_head, roi_extractor, bbox_head) rpn_test_sym = None test_sym = None else: train_sym = None rpn_test_sym = detector.get_rpn_test_symbol(backbone, neck, rpn_head) test_sym = detector.get_test_symbol(backbone, neck, rpn_head, roi_extractor, bbox_head) class ModelParam: train_symbol = train_sym test_symbol = test_sym rpn_test_symbol = rpn_test_sym from_scratch = True random = True memonger = False memonger_until = "stage3_unit21_plus" class pretrain: prefix = None epoch = 0 fixed_param = [] def process_weight(sym, arg, aux): for stride in RpnParam.anchor_generate.stride: add_anchor_to_arg(sym, arg, aux, RpnParam.anchor_generate.max_side, stride, RpnParam.anchor_generate.scale, RpnParam.anchor_generate.ratio) class OptimizeParam: class optimizer: type = "sgd" lr = 0.01 / 8 * len(KvstoreParam.gpus) * KvstoreParam.batch_image momentum = 0.9 wd = 1e-4 clip_gradient = None class schedule: mult = 6 begin_epoch = 0 end_epoch = 6 * mult if mult <= 2: lr_iter = [ 60000 * mult * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image), 80000 * mult * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image) ] else: # follow the setting in Rethinking ImageNet Pre-training # reduce the lr in the last 60k and 20k iterations lr_iter = [ (DatasetParam.total_image * 2 // 16 * end_epoch - 60000) * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image), (DatasetParam.total_image * 2 // 16 * end_epoch - 20000) * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image) ] class warmup: type = "gradual" lr = 0 iter = 500 class TestParam: min_det_score = 0.05 max_det_per_image = 100 process_roidb = lambda x: x process_output = lambda x, y: x class model: prefix = "experiments/{}/checkpoint".format(General.name) epoch = OptimizeParam.schedule.end_epoch class nms: type = "nms" thr = 0.5 class coco: annotation = "data/coco/annotations/instances_minival2014.json" # data processing class NormParam: mean = tuple(i * 255 for i in (0.485, 0.456, 0.406)) # RGB order std = tuple(i * 255 for i in (0.229, 0.224, 0.225)) # data processing class ResizeParam: short = 400 long = 600 class PadParam: short = 400 long = 600 max_num_gt = 100 class AnchorTarget2DParam: def __init__(self): self.generate = self._generate() class _generate: def __init__(self): self.stride = (4, 8, 16, 32, 64) self.short = (100, 50, 25, 13, 7) self.long = (150, 75, 38, 19, 10) scales = (4) aspects = (0.5, 1.0, 2.0) class assign: allowed_border = 0 pos_thr = 0.7 neg_thr = 0.3 min_pos_thr = 0.0 class sample: image_anchor = 256 pos_fraction = 0.5 class RenameParam: mapping = dict(image="data") from core.detection_input import ReadRoiRecord, Resize2DImageBbox, \ ConvertImageFromHwcToChw, Flip2DImageBbox, Pad2DImageBbox, \ RenameRecord, Norm2DImage from models.FPN.input import PyramidAnchorTarget2D if is_train: transform = [ ReadRoiRecord(None), Norm2DImage(NormParam), Resize2DImageBbox(ResizeParam), Flip2DImageBbox(), Pad2DImageBbox(PadParam), ConvertImageFromHwcToChw(), RenameRecord(RenameParam.mapping) ] data_name = ["data"] label_name = ["gt_bbox", "im_info"] if not RpnParam.nnvm_rpn_target: transform.append(PyramidAnchorTarget2D(AnchorTarget2DParam())) label_name += ["rpn_cls_label", "rpn_reg_target", "rpn_reg_weight"] else: transform = [ ReadRoiRecord(None), Norm2DImage(NormParam), Resize2DImageBbox(ResizeParam), Pad2DImageBbox(PadParam), ConvertImageFromHwcToChw(), RenameRecord(RenameParam.mapping) ] data_name = ["data", "im_info", "im_id", "rec_id"] label_name = [] import core.detection_metric as metric rpn_acc_metric = metric.AccWithIgnore( "RpnAcc", ["rpn_cls_loss_output", "rpn_cls_label_blockgrad_output"], []) rpn_l1_metric = metric.L1( "RpnL1", ["rpn_reg_loss_output", "rpn_cls_label_blockgrad_output"], []) # for bbox, the label is generated in network so it is an output box_acc_metric = metric.AccWithIgnore( "RcnnAcc", ["bbox_cls_loss_output", "bbox_label_blockgrad_output"], []) box_l1_metric = metric.L1( "RcnnL1", ["bbox_reg_loss_output", "bbox_label_blockgrad_output"], []) metric_list = [ rpn_acc_metric, rpn_l1_metric, box_acc_metric, box_l1_metric ] return General, KvstoreParam, RpnParam, RoiParam, BboxParam, DatasetParam, \ ModelParam, OptimizeParam, TestParam, \ transform, data_name, label_name, metric_list
def get_config(is_train): class General: use_groupsoftmax = True log_frequency = 20 name = __name__.rsplit("/")[-1].rsplit(".")[-1] batch_image = 2 if is_train else 1 fp16 = True class KvstoreParam: kvstore = "local" batch_image = General.batch_image gpus = [0, 1, 2, 3, 4, 5, 6, 7] fp16 = General.fp16 class NormalizeParam: if is_train: normalizer = normalizer_factory(type="syncbn", ndev=len(KvstoreParam.gpus)) else: normalizer = normalizer_factory(type="fixbn") class BackboneParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer class NeckParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer class RpnParam: fp16 = General.fp16 normalizer = normalizer_factory( type="fixbn") # old model does not use BN in RPN head batch_image = General.batch_image use_groupsoftmax = General.use_groupsoftmax num_class = (1 + 2) if use_groupsoftmax else 2 class anchor_generate: scale = (2, 4, 8, 16, 32) ratio = (0.5, 1.0, 2.0) stride = 16 image_anchor = 256 class head: conv_channel = 512 mean = (0, 0, 0, 0) std = (1, 1, 1, 1) class proposal: pre_nms_top_n = 12000 if is_train else 6000 post_nms_top_n = 2000 if is_train else 1000 nms_thr = 0.7 min_bbox_side = 0 class subsample_proposal: proposal_wo_gt = True image_roi = 256 fg_fraction = 0.25 fg_thr = 0.5 bg_thr_hi = 0.5 bg_thr_lo = 0.0 class bbox_target: num_reg_class = 2 class_agnostic = True weight = (1.0, 1.0, 1.0, 1.0) mean = (0.0, 0.0, 0.0, 0.0) std = (0.1, 0.1, 0.2, 0.2) class BboxParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer num_class = 1 + 83 image_roi = 256 batch_image = General.batch_image use_groupsoftmax = General.use_groupsoftmax class regress_target: class_agnostic = True mean = (0.0, 0.0, 0.0, 0.0) std = (0.1, 0.1, 0.2, 0.2) class RoiParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer out_size = 7 stride = 16 class DatasetParam: if is_train: image_set = ("coco_train2014", "coco_valminusminival2014", "cctsdb_train") else: image_set = ("coco_minival2014", ) backbone = Backbone(BackboneParam) neck = Neck(NeckParam) rpn_head = RpnHead(RpnParam) roi_extractor = RoiExtractor(RoiParam) bbox_head = BboxHead(BboxParam) detector = Detector() if is_train: train_sym = detector.get_train_symbol(backbone, neck, rpn_head, roi_extractor, bbox_head) rpn_test_sym = None test_sym = None else: train_sym = None rpn_test_sym = detector.get_rpn_test_symbol(backbone, neck, rpn_head) test_sym = detector.get_test_symbol(backbone, neck, rpn_head, roi_extractor, bbox_head) class ModelParam: train_symbol = train_sym test_symbol = test_sym rpn_test_symbol = rpn_test_sym from_scratch = False random = True memonger = False memonger_until = "stage3_unit21_plus" class pretrain: prefix = "pretrain_model/resnet-101" epoch = 0 fixed_param = [] class OptimizeParam: class optimizer: type = "sgd" lr = 0.01 / 8 * len(KvstoreParam.gpus) * KvstoreParam.batch_image momentum = 0.9 wd = 0.0001 clip_gradient = 5 class schedule: begin_epoch = 0 end_epoch = 6 lr_iter = [ 60000 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image), 80000 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image) ] class warmup: type = "gradual" lr = 0.0 iter = 3000 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image) class TestParam: min_det_score = 0.05 max_det_per_image = 100 process_roidb = lambda x: x process_output = lambda x, y: x class model: prefix = "experiments/{}/checkpoint".format(General.name) epoch = OptimizeParam.schedule.end_epoch class nms: type = "nms" thr = 0.5 class coco: annotation = "/ws/data/opendata/coco/annotations/instances_minival2014.json" # data processing class GroupParam: # box 83 classes boxv0 = np.array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, \ 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, \ 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83], dtype=np.float32) #COCO benchmark boxv1 = np.array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, \ 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, \ 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 0, 0, 0 ], dtype=np.float32) #CCTSDB benchmark boxv2 = np.array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, \ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, \ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 81, 82, 83], dtype=np.float32) rpnv0 = np.array([0, 1, 2], dtype=np.float32) # rpn 3 classes rpnv1 = np.array([0, 1, 0], dtype=np.float32) # COCO benchmark rpnv2 = np.array([0, 0, 2], dtype=np.float32) # CCTSDB benchmark rpn_groups = [rpnv0, rpnv1, rpnv2] box_groups = [boxv0, boxv1, boxv2] class ResizeParam: short = 800 long = 1200 if is_train else 2000 class PadParam: short = 800 long = 1200 max_num_gt = 100 class AnchorTarget2DParam: class generate: short = 800 // 16 long = 1200 // 16 stride = 16 scales = (2, 4, 8, 16, 32) aspects = (0.5, 1.0, 2.0) use_groupsoftmax = General.use_groupsoftmax class assign: allowed_border = 0 pos_thr = 0.7 neg_thr = 0.3 min_pos_thr = 0.0 class sample: image_anchor = 256 pos_fraction = 0.5 def gtclass2rpn(gtclass): class_gap = 80 gtclass[gtclass > class_gap] = -1 gtclass[gtclass > 0] = 1 gtclass[gtclass < 0] = 2 return gtclass class RenameParam: mapping = dict(image="data") from core.detection_input import ReadRoiRecord, Resize2DImageBbox, \ ConvertImageFromHwcToChw, Flip2DImageBbox, Pad2DImageBbox, \ RenameRecord, AnchorTarget2D, GroupRead if is_train: transform = [ ReadRoiRecord(None), Resize2DImageBbox(ResizeParam), Flip2DImageBbox(), Pad2DImageBbox(PadParam), ConvertImageFromHwcToChw(), AnchorTarget2D(AnchorTarget2DParam), RenameRecord(RenameParam.mapping) ] data_name = ["data", "im_info", "gt_bbox"] label_name = ["rpn_cls_label", "rpn_reg_target", "rpn_reg_weight"] if General.use_groupsoftmax: data_name.append("rpn_group") data_name.append("box_group") transform.append(GroupRead(GroupParam)) else: transform = [ ReadRoiRecord(None), Resize2DImageBbox(ResizeParam), ConvertImageFromHwcToChw(), RenameRecord(RenameParam.mapping) ] data_name = ["data", "im_info", "im_id", "rec_id"] label_name = [] import core.detection_metric as metric rpn_acc_metric = metric.AccWithIgnore("RpnAcc", ["rpn_cls_loss_output"], ["rpn_cls_label"]) rpn_l1_metric = metric.L1("RpnL1", ["rpn_reg_loss_output"], ["rpn_cls_label"]) # for bbox, the label is generated in network so it is an output box_acc_metric = metric.AccWithIgnore( "RcnnAcc", ["bbox_cls_loss_output", "bbox_label_blockgrad_output"], []) box_l1_metric = metric.L1( "RcnnL1", ["bbox_reg_loss_output", "bbox_label_blockgrad_output"], []) metric_list = [ rpn_acc_metric, rpn_l1_metric, box_acc_metric, box_l1_metric ] return General, KvstoreParam, RpnParam, RoiParam, BboxParam, DatasetParam, \ ModelParam, OptimizeParam, TestParam, \ transform, data_name, label_name, metric_list
def get_config(is_train): class General: log_frequency = 20 name = __name__.rsplit("/")[-1].rsplit(".")[-1] batch_image = 2 if is_train else 1 fp16 = False loader_worker = 4 loader_collector = 2 profile = False class KvstoreParam: kvstore = "nccl" batch_image = General.batch_image gpus = [0, 1, 2, 3, 4, 5, 6, 7] fp16 = General.fp16 class NormalizeParam: normalizer = normalizer_factory(type="fix") class BboxParam: pass class RoiParam: pass class RpnParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer batch_image = General.batch_image class head: conv_channel = 256 mean = (0, 0, 0, 0) std = (1, 1, 1, 1) class proposal: pre_nms_thresh = 0.05 pre_nms_top_n = 1000 post_nms_top_n = 1000 fpn_box_max_n = 100 nms_thr = 0.6 min_bbox_side = 0 class subsample_proposal: proposal_wo_gt = False image_roi = 512 fg_fraction = 0.25 fg_thr = 0.5 bg_thr_hi = 0.5 bg_thr_lo = 0.0 class loss_setting: focal_loss_alpha = 0.25 focal_loss_gamma = 2.0 ignore_label = -1 ignore_offset = -1 class FCOSParam: num_classifier = 81 - 1 # COCO: 80 object + 1 background stride = (8, 16, 32, 64, 128) # data processing class NormParam: mean = (122.7717, 115.9465, 102.9801) # RGB order std = (1.0, 1.0, 1.0) class ResizeParam: short = 800 long = 1333 class PadParam: short = 800 long = 1333 max_num_gt = 100 class FCOSFPNAssignParam: stages = [ [-1, 64], [64, 128], [128, 256], [256, 512], [512, INF], ] stride = (8, 16, 32, 64, 128) num_classifier = 81 - 1 # COCO: 80 object + 1 background ignore_label = RpnParam.loss_setting.ignore_label ignore_offset = RpnParam.loss_setting.ignore_offset data_size = [PadParam.short, PadParam.long] class RenameParam: mapping = dict(image="data") class BackboneParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer class NeckParam: fp16 = General.fp16 normalizer = normalizer_factory(type="dummy") class DatasetParam: if is_train: image_set = ("coco_train2014", "coco_valminusminival2014") else: image_set = ("coco_minival2014", ) # throw out param used as custom op's input global throwout_param throwout_param = FCOSFPNAssignParam # This line MUST be in front of rpn_head = RpnHead(RpnParam) backbone = Backbone(BackboneParam) neck = Neck(NeckParam) rpn_head = RpnHead(RpnParam) detector = Detector() if is_train: train_sym = detector.get_train_symbol(backbone, neck, rpn_head) rpn_test_sym = None test_sym = None else: rpn_test_sym = detector.get_rpn_test_symbol(backbone, neck, rpn_head) train_sym = None class ModelParam: train_symbol = train_sym test_symbol = rpn_test_sym rpn_test_symbol = rpn_test_sym from_scratch = False random = True memonger = False memonger_until = "stage3_unit21_plus" class pretrain: prefix = "pretrain_model/resnet-101" epoch = 0 fixed_param = ["conv0", "stage1", "gamma", "beta"] class OptimizeParam: class optimizer: type = "sgd" lr = 0.01 / 16 * len(KvstoreParam.gpus) * KvstoreParam.batch_image momentum = 0.9 wd = 0.0001 clip_gradient = None class schedule: begin_epoch = 0 end_epoch = 6 lr_iter = [ 60000 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image), 80000 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image) ] class warmup: type = "gradual" lr = 0.01 / 16 * len( KvstoreParam.gpus) * KvstoreParam.batch_image / 3.0 iter = 500 class TestParam: min_det_score = 0 max_det_per_image = 100 process_roidb = lambda x: x process_output = lambda x, y: x class model: prefix = "experiments/{}/checkpoint".format(General.name) epoch = OptimizeParam.schedule.end_epoch class nms: type = "nms" thr = 0.6 class coco: annotation = "data/coco/annotations/instances_minival2014.json" from core.detection_input import ReadRoiRecord, Resize2DImageBbox, \ ConvertImageFromHwcToChw, Flip2DImageBbox, Pad2DImageBbox, \ RenameRecord, Norm2DImage if is_train: transform = [ ReadRoiRecord(None), Norm2DImage(NormParam), Resize2DImageBbox(ResizeParam), Flip2DImageBbox(), Pad2DImageBbox(PadParam), ConvertImageFromHwcToChw(), RenameRecord(RenameParam.mapping) ] data_name = ["data", "im_info"] label_name = ["gt_bbox"] else: transform = [ ReadRoiRecord(None), Norm2DImage(NormParam), Resize2DImageBbox(ResizeParam), ConvertImageFromHwcToChw(), RenameRecord(RenameParam.mapping) ] data_name = ["data", "im_info", "im_id", "rec_id"] label_name = [] import models.FCOS.metric as metric centerness_loss_metric = metric.LossMeter(RpnParam.FCOSParam.stride, pred_id_start=0, pred_id_end=1, name='centernessloss_meter') cls_loss_metric = metric.LossMeter(RpnParam.FCOSParam.stride, pred_id_start=1, pred_id_end=2, name='clsloss_meter') reg_loss_metric = metric.LossMeter(RpnParam.FCOSParam.stride, pred_id_start=2, pred_id_end=3, name='offsetloss_meter') metric_list = [centerness_loss_metric, cls_loss_metric, reg_loss_metric] return General, KvstoreParam, RpnParam, RoiParam, BboxParam, DatasetParam, \ ModelParam, OptimizeParam, TestParam, \ transform, data_name, label_name, metric_list
def get_config(is_train): class General: log_frequency = 10 name = __name__.rsplit("/")[-1].rsplit(".")[-1] batch_image = 2 if is_train else 1 fp16 = True loader_worker = 24 class KvstoreParam: kvstore = "nccl" batch_image = General.batch_image gpus = [0, 1, 2, 3, 4, 5, 6, 7] fp16 = General.fp16 class NormalizeParam: normalizer = normalizer_factory(type="fixbn") class BackboneParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer depth = 50 class NeckParam: fp16 = General.fp16 normalizer = normalizer_factory(type="localbn") class RpnParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer batch_image = General.batch_image nnvm_proposal = True nnvm_rpn_target = False class anchor_generate: scale = (8, ) ratio = (0.5, 1, 2) stride = (4, 8, 16, 32, 64) max_side = 1450 class anchor_assign: allowed_border = 1000 pos_thr = 0.7 neg_thr = 0.3 min_pos_thr = 0.0 image_anchor = 256 pos_fraction = 0.5 class head: conv_channel = 256 mean = (0, 0, 0, 0) std = (1, 1, 1, 1) class proposal: pre_nms_top_n = 3000 if is_train else 2000 post_nms_top_n = 2000 if is_train else 1000 nms_thr = 0.7 min_bbox_side = 2 class subsample_proposal: proposal_wo_gt = False image_roi = 512 fg_fraction = 0.5 fg_thr = 0.5 bg_thr_hi = 0.5 bg_thr_lo = 0.0 class bbox_target: num_reg_class = 2 class_agnostic = False weight = (1.0, 1.0, 1.0, 1.0) mean = (0.0, 0.0, 0.0, 0.0) std = (0.1, 0.1, 0.2, 0.2) class BboxParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer num_class = 1 + 1 image_roi = 512 batch_image = General.batch_image class regress_target: class_agnostic = False mean = (0.0, 0.0, 0.0, 0.0) std = (0.1, 0.1, 0.2, 0.2) class RoiParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer out_size = 7 stride = (4, 8, 16, 32) roi_canonical_scale = 224 roi_canonical_level = 4 class DatasetParam: if is_train: image_set = ("crowdhuman_train", ) else: image_set = ("crowdhuman_val", ) backbone = Backbone(BackboneParam) neck = Neck(NeckParam) rpn_head = RpnHead(RpnParam) roi_extractor = RoiExtractor(RoiParam) bbox_head = BboxHead(BboxParam) detector = Detector() if is_train: train_sym = detector.get_train_symbol(backbone, neck, rpn_head, roi_extractor, bbox_head) rpn_test_sym = None test_sym = None else: train_sym = None rpn_test_sym = detector.get_rpn_test_symbol(backbone, neck, rpn_head) test_sym = detector.get_test_symbol(backbone, neck, rpn_head, roi_extractor, bbox_head) class ModelParam: train_symbol = train_sym test_symbol = test_sym rpn_test_symbol = rpn_test_sym from_scratch = False random = True memonger = False memonger_until = "stage3_unit21_plus" class pretrain: prefix = "pretrain_model/resnet%s_v1b" % BackboneParam.depth epoch = 0 fixed_param = [ "conv0", "stage1", "bn_gamma", "bn_beta", "bn0", "bn1", "bn2", "bn3", "bn4" ] def process_weight(sym, arg, aux): for stride in RpnParam.anchor_generate.stride: add_anchor_to_arg(sym, arg, aux, RpnParam.anchor_generate.max_side, stride, RpnParam.anchor_generate.scale, RpnParam.anchor_generate.ratio) class OptimizeParam: class optimizer: type = "sgd" lr = 0.01 / 8 * len(KvstoreParam.gpus) * KvstoreParam.batch_image momentum = 0.9 wd = 0.0001 clip_gradient = None class schedule: begin_epoch = 0 # end_epoch = 15 # lr_iter = [18750 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image), # 24375 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image)] end_epoch = 10 lr_iter = [ 14960 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image), 17765 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image) ] class warmup: type = "gradual" lr = 0.01 / 8 * len( KvstoreParam.gpus) * KvstoreParam.batch_image / 3.0 iter = 500 class TestParam: min_det_score = 0.05 max_det_per_image = 300 process_roidb = lambda x: x process_output = lambda x, y: x class model: prefix = "experiments/{}/checkpoint".format(General.name) epoch = OptimizeParam.schedule.end_epoch class nms: type = "nms" thr = 0.5 class coco: annotation = "" # data processing class NormParam: mean = tuple(i * 255 for i in (0.485, 0.456, 0.406)) # RGB order std = tuple(i * 255 for i in (0.229, 0.224, 0.225)) # data processing class ResizeParam: short = 800 long = 1400 class PadParam: short = 800 long = 1400 max_num_gt = 500 class AnchorTarget2DParam: def __init__(self): self.generate = self._generate() class _generate: def __init__(self): self.stride = (4, 8, 16, 32, 64) self.short = (200, 100, 50, 25, 13) self.long = (350, 175, 88, 44, 22) scales = (8) aspects = RpnParam.anchor_generate.ratio class assign: allowed_border = 1000 pos_thr = 0.7 neg_thr = 0.3 min_pos_thr = 0.0 class sample: image_anchor = 256 pos_fraction = 0.5 class RenameParam: mapping = dict(image="data") from core.detection_input import ReadRoiRecord, Resize2DImageBbox, \ ConvertImageFromHwcToChw, Flip2DImageBbox, Pad2DImageBbox, \ RenameRecord, Norm2DImage from models.crowdhuman.input import PyramidAnchorTarget2D if is_train: transform = [ ReadRoiRecord(None), Norm2DImage(NormParam), Resize2DImageBbox(ResizeParam), Flip2DImageBbox(), Pad2DImageBbox(PadParam), ConvertImageFromHwcToChw(), RenameRecord(RenameParam.mapping) ] data_name = ["data"] label_name = ["gt_bbox", "im_info"] if not RpnParam.nnvm_rpn_target: transform.append(PyramidAnchorTarget2D(AnchorTarget2DParam())) label_name += ["rpn_cls_label", "rpn_reg_target", "rpn_reg_weight"] else: transform = [ ReadRoiRecord(None), Norm2DImage(NormParam), Resize2DImageBbox(ResizeParam), ConvertImageFromHwcToChw(), RenameRecord(RenameParam.mapping) ] data_name = ["data", "im_info", "im_id", "rec_id"] label_name = [] import core.detection_metric as metric rpn_acc_metric = metric.AccWithIgnore( "RpnAcc", ["rpn_cls_loss_output", "rpn_cls_label_blockgrad_output"], []) rpn_l1_metric = metric.L1( "RpnL1", ["rpn_reg_loss_output", "rpn_cls_label_blockgrad_output"], []) # for bbox, the label is generated in network so it is an output box_acc_metric = metric.AccWithIgnore( "RcnnAcc", ["bbox_cls_loss_output", "bbox_label_blockgrad_output"], []) box_l1_metric = metric.L1( "RcnnL1", ["bbox_reg_loss_output", "bbox_label_blockgrad_output"], []) metric_list = [ rpn_acc_metric, rpn_l1_metric, box_acc_metric, box_l1_metric ] return General, KvstoreParam, RpnParam, RoiParam, BboxParam, DatasetParam, \ ModelParam, OptimizeParam, TestParam, \ transform, data_name, label_name, metric_list
def get_config(is_train): class General: log_frequency = 10 name = __name__.rsplit("/")[-1].rsplit(".")[-1] batch_image = 2 if is_train else 1 fp16 = False class KvstoreParam: kvstore = "nccl" batch_image = General.batch_image gpus = [0, 1, 2, 3, 4, 5, 6, 7] fp16 = General.fp16 class NormalizeParam: # normalizer = normalizer_factory(type="syncbn", ndev=len(KvstoreParam.gpus)) normalizer = normalizer_factory(type="fixbn") class BackboneParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer class NeckParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer class RpnParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer batch_image = General.batch_image class anchor_generate: scale = (2, 4, 8, 16, 32) ratio = (0.5, 1.0, 2.0) stride = 16 image_anchor = 256 class head: conv_channel = 512 mean = (0, 0, 0, 0) std = (1, 1, 1, 1) class proposal: pre_nms_top_n = 12000 if is_train else 6000 post_nms_top_n = 2000 if is_train else 1000 nms_thr = 0.7 min_bbox_side = 0 class BboxParam: pass class RoiParam: pass class DatasetParam: if is_train: image_set = ("coco_train2014", "coco_valminusminival2014") else: image_set = ("coco_minival2014", ) backbone = Backbone(BackboneParam) neck = Neck(NeckParam) rpn_head = RpnHead(RpnParam) detector = Detector() if is_train: train_sym = detector.get_train_symbol(backbone, neck, rpn_head) rpn_test_sym = None test_sym = None else: train_sym = None rpn_test_sym = detector.get_rpn_test_symbol(backbone, neck, rpn_head) test_sym = None class ModelParam: train_symbol = train_sym test_symbol = test_sym rpn_test_symbol = rpn_test_sym from_scratch = False random = True memonger = False memonger_until = "stage3_unit21_plus" class pretrain: prefix = "pretrain_model/resnet-50" epoch = 0 fixed_param = ["conv0", "stage1", "gamma", "beta"] class OptimizeParam: class optimizer: type = "sgd" lr = 0.01 / 8 * len(KvstoreParam.gpus) * KvstoreParam.batch_image momentum = 0.9 wd = 0.0001 clip_gradient = 35 class schedule: begin_epoch = 0 end_epoch = 6 lr_iter = [ 60000 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image), 80000 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image) ] class warmup: type = "gradual" lr = 0.0 iter = 3000 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image) class TestParam: min_det_score = 0.05 max_det_per_image = 100 process_roidb = lambda x: x process_output = lambda x, y: x class model: prefix = "experiments/{}/checkpoint".format(General.name) epoch = OptimizeParam.schedule.end_epoch class nms: type = "nms" thr = 0.5 class coco: annotation = "data/coco/annotations/instances_minival2014.json" # data processing class ResizeParam: short = 800 long = 1200 if is_train else 2000 class PadParam: short = 800 long = 1200 if is_train else 2000 max_num_gt = 100 class AnchorTarget2DParam: class generate: short = 800 // 16 long = 1200 // 16 stride = 16 scales = (2, 4, 8, 16, 32) aspects = (0.5, 1.0, 2.0) class assign: allowed_border = 0 pos_thr = 0.7 neg_thr = 0.3 min_pos_thr = 0.0 class sample: image_anchor = 256 pos_fraction = 0.5 class RenameParam: mapping = dict(image="data") from core.detection_input import ReadRoiRecord, Resize2DImageBbox, \ ConvertImageFromHwcToChw, Flip2DImageBbox, Pad2DImageBbox, \ RenameRecord, AnchorTarget2D if is_train: transform = [ ReadRoiRecord(None), Resize2DImageBbox(ResizeParam), Flip2DImageBbox(), Pad2DImageBbox(PadParam), ConvertImageFromHwcToChw(), AnchorTarget2D(AnchorTarget2DParam), RenameRecord(RenameParam.mapping) ] data_name = ["data"] label_name = ["rpn_cls_label", "rpn_reg_target", "rpn_reg_weight"] else: transform = [ ReadRoiRecord(None), Resize2DImageBbox(ResizeParam), ConvertImageFromHwcToChw(), RenameRecord(RenameParam.mapping) ] data_name = ["data", "im_info", "im_id", "rec_id"] label_name = [] import core.detection_metric as metric rpn_acc_metric = metric.AccWithIgnore("RpnAcc", ["rpn_cls_loss_output"], ["rpn_cls_label"]) rpn_l1_metric = metric.L1("RpnL1", ["rpn_reg_loss_output"], ["rpn_cls_label"]) metric_list = [rpn_acc_metric, rpn_l1_metric] return General, KvstoreParam, RpnParam, RoiParam, BboxParam, DatasetParam, \ ModelParam, OptimizeParam, TestParam, \ transform, data_name, label_name, metric_list
def get_config(is_train): class General: log_frequency = 10 name = __name__.rsplit("/")[-1].rsplit(".")[-1] batch_image = 2 if is_train else 1 fp16 = False long_side = 1200 short_side = 800 class KvstoreParam: kvstore = "local" batch_image = General.batch_image gpus = [0, 1, 2, 3, 4, 5, 6, 7] fp16 = General.fp16 class NormalizeParam: # normalizer = normalizer_factory(type="syncbn", ndev=len(KvstoreParam.gpus)) normalizer = normalizer_factory(type="fixbn") class BackboneParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer depth = 50 class NeckParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer class RpnParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer batch_image = General.batch_image use_symbolic_proposal = None class anchor_generate: scale = (2, 4, 8, 16, 32) ratio = (0.5, 1.0, 2.0) stride = 16 max_side = General.long_side image_anchor = 256 class head: conv_channel = 512 mean = (0, 0, 0, 0) std = (1, 1, 1, 1) class proposal: pre_nms_top_n = 12000 if is_train else 6000 post_nms_top_n = 2000 if is_train else 1000 nms_thr = 0.7 min_bbox_side = 0 class subsample_proposal: proposal_wo_gt = False image_roi = 512 fg_fraction = 0.25 fg_thr = 0.5 bg_thr_hi = 0.5 bg_thr_lo = 0.0 class bbox_target: num_reg_class = 2 class_agnostic = True weight = (1.0, 1.0, 1.0, 1.0) mean = (0.0, 0.0, 0.0, 0.0) std = (0.1, 0.1, 0.2, 0.2) class BboxParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer num_class = 1 + 80 image_roi = 512 batch_image = General.batch_image class regress_target: class_agnostic = True mean = (0.0, 0.0, 0.0, 0.0) std = (0.1, 0.1, 0.2, 0.2) class RoiParam: fp16 = General.fp16 normalizer = NormalizeParam.normalizer out_size = 7 stride = 16 class DatasetParam: if is_train: image_set = ("coco_train2017", ) else: image_set = ("coco_minival2017", ) backbone = Backbone(BackboneParam) neck = Neck(NeckParam) rpn_head = RpnHead(RpnParam) roi_extractor = RoiExtractor(RoiParam) bbox_head = BboxHead(BboxParam) detector = Detector() if is_train: train_sym = detector.get_train_symbol(backbone, neck, rpn_head, roi_extractor, bbox_head) rpn_test_sym = None test_sym = None else: train_sym = None rpn_test_sym = detector.get_rpn_test_symbol(backbone, neck, rpn_head) test_sym = detector.get_test_symbol(backbone, neck, rpn_head, roi_extractor, bbox_head) class ModelParam: train_symbol = train_sym test_symbol = test_sym rpn_test_symbol = rpn_test_sym from_scratch = False random = True memonger = False memonger_until = "stage3_unit21_plus" process_weight = lambda sym, arg, aux: \ add_anchor_to_arg( sym, arg, aux, RpnParam.anchor_generate.max_side, RpnParam.anchor_generate.stride,RpnParam.anchor_generate.scale, RpnParam.anchor_generate.ratio) class pretrain: prefix = "pretrain_model/resnet%s_v1b" % BackboneParam.depth epoch = 0 fixed_param = ["conv0", "stage1", "gamma", "beta"] class QuantizeTrainingParam: quantize_flag = False # quantized_op = ("Convolution", "FullyConnected", "Deconvolution","Concat", "Pooling", "add_n", "elemwise_add") quantized_op = ("Convolution", "FullyConnected", "Deconvolution") class WeightQuantizeParam: delay_quant = 0 ema_decay = 0.99 grad_mode = "ste" is_weight = True is_weight_perchannel = False quant_mode = "minmax" class ActQuantizeParam: delay_quant = 0 ema_decay = 0.99 grad_mode = "ste" is_weight = False is_weight_perchannel = False quant_mode = "minmax" class OptimizeParam: class optimizer: type = "sgd" lr = 0.01 / 8 * len(KvstoreParam.gpus) * KvstoreParam.batch_image momentum = 0.9 wd = 0.0001 clip_gradient = 35 class schedule: begin_epoch = 0 end_epoch = 6 lr_iter = [60000 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image), 80000 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image)] class warmup: type = "gradual" lr = 0.0 iter = 750 * 16 // (len(KvstoreParam.gpus) * KvstoreParam.batch_image) class TestParam: min_det_score = 0.05 max_det_per_image = 100 process_roidb = lambda x: x process_output = lambda x, y: x class model: prefix = "experiments/{}/checkpoint".format(General.name) epoch = OptimizeParam.schedule.end_epoch class nms: type = "nms" thr = 0.5 class coco: annotation = "data/coco/annotations/instances_minival2014.json" # data processing class NormParam: mean = tuple(i * 255 for i in (0.485, 0.456, 0.406)) # RGB order std = tuple(i * 255 for i in (0.229, 0.224, 0.225)) class ResizeParam: short = General.short_side long = General.long_side class PadParam: short = ResizeParam.short long = ResizeParam.long max_num_gt = 100 class AnchorTarget2DParam: class generate: short = ResizeParam.short // RpnParam.anchor_generate.stride long = ResizeParam.long // RpnParam.anchor_generate.stride stride = RpnParam.anchor_generate.stride scales = RpnParam.anchor_generate.scale aspects = RpnParam.anchor_generate.ratio class assign: allowed_border = 0 pos_thr = 0.7 neg_thr = 0.3 min_pos_thr = 0.0 class sample: image_anchor = RpnParam.anchor_generate.image_anchor pos_fraction = 0.5 class RenameParam: mapping = dict(image="data") from core.detection_input import ReadRoiRecord, Resize2DImageBbox, \ ConvertImageFromHwcToChw, Flip2DImageBbox, Pad2DImageBbox, \ RenameRecord, AnchorTarget2D, Norm2DImage, Pad2DImage if is_train: transform = [ ReadRoiRecord(None), Norm2DImage(NormParam), Resize2DImageBbox(ResizeParam), Flip2DImageBbox(), Pad2DImageBbox(PadParam), ConvertImageFromHwcToChw(), AnchorTarget2D(AnchorTarget2DParam), RenameRecord(RenameParam.mapping) ] data_name = ["data", "im_info", "gt_bbox"] label_name = ["rpn_cls_label", "rpn_reg_target", "rpn_reg_weight"] else: transform = [ ReadRoiRecord(None), Norm2DImage(NormParam), Resize2DImageBbox(ResizeParam), Pad2DImage(PadParam), ConvertImageFromHwcToChw(), RenameRecord(RenameParam.mapping) ] data_name = ["data", "im_info", "im_id", "rec_id"] label_name = [] import core.detection_metric as metric rpn_acc_metric = metric.AccWithIgnore( "RpnAcc", ["rpn_cls_loss_output"], ["rpn_cls_label"] ) rpn_l1_metric = metric.L1( "RpnL1", ["rpn_reg_loss_output"], ["rpn_cls_label"] ) # for bbox, the label is generated in network so it is an output box_acc_metric = metric.AccWithIgnore( "RcnnAcc", ["bbox_cls_loss_output", "bbox_label_blockgrad_output"], [] ) box_l1_metric = metric.L1( "RcnnL1", ["bbox_reg_loss_output", "bbox_label_blockgrad_output"], [] ) metric_list = [rpn_acc_metric, rpn_l1_metric, box_acc_metric, box_l1_metric] return General, KvstoreParam, RpnParam, RoiParam, BboxParam, DatasetParam, \ ModelParam, OptimizeParam, TestParam, \ transform, data_name, label_name, metric_list