Grayscale v7

Train.ipynb

@@ -0,0 +1,35 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "4fbd6fa3-a31c-4ba1-a917-6669fafcd904",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "!python -m torch.distributed.launch --nproc_per_node 2 train.py --img <img size> --batch <number of batches> --epochs <number of epochs> --data <yaml with labels path> --hyp data/hyps/hyp.scratch-low.yaml --weights yolov5m --cache --device 0,1 --name <experiment name>"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.8.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

data/hyps/hyp.Objects365.yaml

@@ -32,3 +32,10 @@ fliplr: 0.5
 mosaic: 1.0
 mixup: 0.0
 copy_paste: 0.0
+
+# Custom Params
+area_threshold: 0.3 # area threshold for random perspective
+maximum_mistakes_size: 7680 # image size for mistakes mosaic
+maximum_mistakes_subplots: 64 # subimages in mistakes mosaic
+minimum_mistakes_iou: 0.5 # mistakes minimum iou for match
+minimum_mistakes_confidence: 0.5 # mistakes minimum confidence for checking

data/hyps/hyp.scratch-high.yaml

@@ -32,3 +32,10 @@ fliplr: 0.5  # image flip left-right (probability)
 mosaic: 1.0  # image mosaic (probability)
 mixup: 0.1  # image mixup (probability)
 copy_paste: 0.1  # segment copy-paste (probability)
+
+# Custom Params
+area_threshold: 0.3 # area threshold for random perspective
+maximum_mistakes_size: 7680 # image size for mistakes mosaic
+maximum_mistakes_subplots: 64 # subimages in mistakes mosaic
+minimum_mistakes_iou: 0.5 # mistakes minimum iou for match
+minimum_mistakes_confidence: 0.5 # mistakes minimum confidence for checking

data/hyps/hyp.scratch-low.yaml

@@ -32,3 +32,10 @@ fliplr: 0.5  # image flip left-right (probability)
 mosaic: 1.0  # image mosaic (probability)
 mixup: 0.0  # image mixup (probability)
 copy_paste: 0.0  # segment copy-paste (probability)
+
+# Custom Params
+area_threshold: 0.3 # area threshold for random perspective
+maximum_mistakes_size: 7680 # image size for mistakes mosaic
+maximum_mistakes_subplots: 64 # subimages in mistakes mosaic
+minimum_mistakes_iou: 0.5 # mistakes minimum iou for match
+minimum_mistakes_confidence: 0.5 # mistakes minimum confidence for checking

data/hyps/hyp.scratch-med.yaml

@@ -32,3 +32,10 @@ fliplr: 0.5  # image flip left-right (probability)
 mosaic: 1.0  # image mosaic (probability)
 mixup: 0.1  # image mixup (probability)
 copy_paste: 0.0  # segment copy-paste (probability)
+
+# Custom Params
+area_threshold: 0.3 # area threshold for random perspective
+maximum_mistakes_size: 7680 # image size for mistakes mosaic
+maximum_mistakes_subplots: 64 # subimages in mistakes mosaic
+minimum_mistakes_iou: 0.5 # mistakes minimum iou for match
+minimum_mistakes_confidence: 0.5 # mistakes minimum confidence for checking

models/yolo.py

@@ -164,7 +164,7 @@ def _apply(self, fn):
 
 class DetectionModel(BaseModel):
     # YOLOv5 detection model
-    def __init__(self, cfg='yolov5s.yaml', ch=3, nc=None, anchors=None):  # model, input channels, number of classes
+    def __init__(self, cfg='yolov5s.yaml', ch=1, nc=None, anchors=None):  # model, input channels, number of classes
         super().__init__()
         if isinstance(cfg, dict):
             self.yaml = cfg  # model dict
@@ -188,6 +188,9 @@ def __init__(self, cfg='yolov5s.yaml', ch=3, nc=None, anchors=None):  # model, i
 
         # Build strides, anchors
         m = self.model[-1]  # Detect()
+
+        ch = 1
+
         if isinstance(m, (Detect, Segment)):
             s = 256  # 2x min stride
             m.inplace = self.inplace
@@ -298,11 +301,13 @@ def _from_yaml(self, cfg):
 
 def parse_model(d, ch):  # model_dict, input_channels(3)
     # Parse a YOLOv5 model.yaml dictionary
+    ch = [1] #!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
     LOGGER.info(f"\n{'':>3}{'from':>18}{'n':>3}{'params':>10}  {'module':<40}{'arguments':<30}")
     anchors, nc, gd, gw, act = d['anchors'], d['nc'], d['depth_multiple'], d['width_multiple'], d.get('activation')
     if act:
         Conv.default_act = eval(act)  # redefine default activation, i.e. Conv.default_act = nn.SiLU()
         LOGGER.info(f"{colorstr('activation:')} {act}")  # print
+
     na = (len(anchors[0]) // 2) if isinstance(anchors, list) else anchors  # number of anchors
     no = na * (nc + 5)  # number of outputs = anchors * (classes + 5)
 

models/yolov5m.yaml

@@ -46,3 +46,4 @@ head:
 
    [[17, 20, 23], 1, Detect, [nc, anchors]],  # Detect(P3, P4, P5)
   ]
+ch: 1

requirements.txt

@@ -12,8 +12,8 @@ psutil  # system resources
 PyYAML>=5.3.1
 requests>=2.23.0
 scipy>=1.4.1
-thop>=0.1.1  # FLOPs computation
-torch>=1.7.0  # see https://pytorch.org/get-started/locally (recommended)
+thop>=0.1.1 # FLOPs computation
+torch>=1.7.0
 torchvision>=0.8.1
 tqdm>=4.64.0
 # protobuf<=3.20.1  # https://github.com/ultralytics/yolov5/issues/8012