fix(quant): fix code and add quantized weights (#38)

hubconf.py

@@ -47,3 +47,5 @@ from official.vision.keypoints.models import (
 )
 
 from official.vision.keypoints.inference import KeypointEvaluator
+
+from official.quantization.models import quantized_resnet18

official/quantization/config.py

@@ -48,7 +48,7 @@ def get_config(arch: str):
 
 class ShufflenetFinetuneConfig(ShufflenetConfig):
     BATCH_SIZE = 128 // 2
-    LEARNING_RATE = 0.003125 // 2
+    LEARNING_RATE = 0.003125 / 2
     EPOCHS = 30
 
 

official/quantization/inference.py

@@ -17,6 +17,7 @@ import megengine.functional as F
 import megengine.jit as jit
 import megengine.quantization as Q
 import numpy as np
+from megengine.quantization.quantize import quantize, quantize_qat
 
 import models
 
@@ -45,7 +46,10 @@ def main():
     model = models.__dict__[args.arch]()
 
     if args.mode != "normal":
-        Q.quantize_qat(model, Q.ema_fakequant_qconfig)
+        quantize_qat(model, Q.ema_fakequant_qconfig)
+
+    if args.mode == "quantized":
+        quantize(model)
 
     if args.checkpoint:
         logger.info("Load pretrained weights from %s", args.checkpoint)
@@ -53,9 +57,6 @@ def main():
         ckpt = ckpt["state_dict"] if "state_dict" in ckpt else ckpt
         model.load_state_dict(ckpt, strict=False)
 
-    if args.mode == "quantized":
-        Q.quantize(model)
-
     if args.image is None:
         path = "../assets/cat.jpg"
     else:

official/quantization/models/resnet.py

@@ -46,6 +46,7 @@ import math
 import megengine.functional as F
 import megengine.hub as hub
 import megengine.module as M
+from megengine.quantization.quantize import quantize_qat, quantize
 
 
 class BasicBlock(M.Module):
@@ -292,58 +293,23 @@ def resnet18(**kwargs):
     r"""ResNet-18 model from
     `"Deep Residual Learning for Image Recognition" <https://arxiv.org/pdf/1512.03385.pdf>`_
     """
-    return ResNet(BasicBlock, [2, 2, 2, 2], **kwargs)
-
-
-def resnet34(**kwargs):
-    r"""ResNet-34 model from
-    `"Deep Residual Learning for Image Recognition" <https://arxiv.org/pdf/1512.03385.pdf>`_
-    """
-    return ResNet(BasicBlock, [3, 4, 6, 3], **kwargs)
+    m = ResNet(BasicBlock, [2, 2, 2, 2], **kwargs)
+    m.fc.disable_quantize()
+    return m
 
 
 def resnet50(**kwargs):
     r"""ResNet-50 model from
     `"Deep Residual Learning for Image Recognition" <https://arxiv.org/pdf/1512.03385.pdf>`_
     """
-    return ResNet(Bottleneck, [3, 4, 6, 3], **kwargs)
-
-
-def resnet101(**kwargs):
-    r"""ResNet-101 model from
-    `"Deep Residual Learning for Image Recognition" <https://arxiv.org/pdf/1512.03385.pdf>`_
-    """
-    return ResNet(Bottleneck, [3, 4, 23, 3], **kwargs)
-
-
-def resnet152(**kwargs):
-    r"""ResNet-152 model from
-    `"Deep Residual Learning for Image Recognition" <https://arxiv.org/pdf/1512.03385.pdf>`_
-    """
-    return ResNet(Bottleneck, [3, 8, 36, 3], **kwargs)
+    m = ResNet(Bottleneck, [3, 4, 6, 3], **kwargs)
+    m.fc.disable_quantize()
+    return m
 
 
-def resnext50_32x4d(**kwargs):
-    r"""ResNeXt-50 32x4d model from
-    `"Aggregated Residual Transformation for Deep Neural Networks" <https://arxiv.org/pdf/1611.05431.pdf>`_
-
-    Args:
-        pretrained (bool): If True, returns a model pre-trained on ImageNet
-        progress (bool): If True, displays a progress bar of the download to stderr
-    """
-    kwargs["groups"] = 32
-    kwargs["width_per_group"] = 4
-    return ResNet(Bottleneck, [3, 4, 6, 3], **kwargs)
-
-
-def resnext101_32x8d(**kwargs):
-    r"""ResNeXt-101 32x8d model from
-    `"Aggregated Residual Transformation for Deep Neural Networks" <https://arxiv.org/pdf/1611.05431.pdf>`_
-
-    Args:
-        pretrained (bool): If True, returns a model pre-trained on ImageNet
-        progress (bool): If True, displays a progress bar of the download to stderr
-    """
-    kwargs["groups"] = 32
-    kwargs["width_per_group"] = 8
-    return ResNet(Bottleneck, [3, 4, 23, 3], **kwargs)
+@hub.pretrained("https://data.megengine.org.cn/models/weights/resnet18.quantized.pkl")
+def quantized_resnet18(**kwargs):
+    model = resnet18(**kwargs)
+    quantize_qat(model)
+    quantize(model)
+    return model