New YOLOv5 Classification Models (#8956)
* Update * Logger step fix: Increment step with epochs (#8654) * enhance * revert * allow training from scratch * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Update --img argument from train.py single line * fix image size from 640 to 128 * suport custom dataloader and augmentation * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * format * Update dataloaders.py * Single line return, single line comment, remove unused argument * address PR comments * fix spelling * don't augment eval set * use fstring * update augmentations.py * new maning convention for transforms * reverse if statement, inline ops * reverse if statement, inline ops * updates * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * update dataloaders * Remove additional if statement * Remove is_train as redundant * Cleanup * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Cleanup2 * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Update classifier.py * Update augmentations.py * fix: imshow clip warning * update * Revert ToTensorV2 removal * Update classifier.py * Update normalize values, revert uint8 * normalize image using cv2 * remove dedundant comment * Update classifier.py * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * replace print with logger * commit steps * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Allow logging models from GenericLogger (#8676) * enhance * revert * allow training from scratch * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Update --img argument from train.py single line * fix image size from 640 to 128 * suport custom dataloader and augmentation * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * format * Update dataloaders.py * Single line return, single line comment, remove unused argument * address PR comments * fix spelling * don't augment eval set * use fstring * update augmentations.py * new maning convention for transforms * reverse if statement, inline ops * reverse if statement, inline ops * updates * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * update dataloaders * Remove additional if statement * Remove is_train as redundant * Cleanup * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Cleanup2 * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Update classifier.py * Update augmentations.py * fix: imshow clip warning * update * Revert ToTensorV2 removal * Update classifier.py * Update normalize values, revert uint8 * normalize image using cv2 * remove dedundant comment * Update classifier.py * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * replace print with logger * commit steps * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * support final model logging * update * update * update * update * remove curses * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Update classifier.py * Update __init__.py Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com> * Update * Update * Update * Update * Update dataset download * Update dataset download * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Pass imgsz to classify_transforms() * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Update * Cos scheduler * Cos scheduler * Remove unused args * Update * Add seed * Add seed * Update * Update * Add run(), main() * Merge master * Merge master * Update * Update * Update * Update * Update * Update * Update * Create YOLOv5 BaseModel class (#8829) * Create BaseModel * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * fix * Hub load device fix * Update Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> * Add experiment * Merge master * Attach names * weight decay = 1e-4 * weight decay = 5e-5 * update smart_optimizer console printout * fashion-mnist fix * Merge master * Update Table * Update Table * Remove destroy process group * add kwargs to forward() * fuse fix for resnet50 * nc, names fix for resnet50 * nc, names fix for resnet50 * ONNX CPU inference fix * revert * cuda * if augment or visualize * if augment or visualize * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * New smart_inference_mode() * Update README * Refactor into /classify dir * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * reset defaults * reset defaults * fix gpu predict * warmup * ema half fix * spacing * remove data * remove cache * remove denormalize * save run settings * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * verbose false on initial plots * new save_yaml() function * Update ci-testing.yml * Path(data) CI fix * Separate classification CI * fix val * fix val * fix val * smartCrossEntropyLoss * skip validation on hub load * autodownload with working dir root * str(data) * Dataset usage example * im_show normalize * im_show normalize * add imagenet simple names to multibackend * Add validation speeds * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * 24-space names * Update bash scripts * Update permissions * Add bash script arguments * remove verbose * TRT data fix * names generator fix * optimize if names * update usage * Add local loading * Verbose=False * update names printing * Add Usage examples * Add Usage examples * Add Usage examples * Add Usage examples * named_children * reshape_classifier_outputs * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * update * update * fix CI * fix incorrect class substitution * fix incorrect class substitution * remove denormalize * ravel fix * cleanup * update opt file printing * update opt file printing * update defaults * add opt to checkpoint * Add warning * Add comment * plot half bug fix * Use NotImplementedError * fix export shape report * Fix TRT load * cleanup CI * profile comment * CI fix * Add cls models * avoid inplace error * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Fix usage examples * Update README * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Update README * Update README * Update README * Update README * Update README * Update README * Update README * Update README * Update README * Update README * Update README * Update README * Update README * Update README * Update README * Update README Co-authored-by: Ayush Chaurasia <ayush.chaurarsia@gmail.com> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
This commit is contained in:
22 files changed
+1436
-161
No files matched your search
+83
-60
@@ -90,8 +90,64 @@ class Detect(nn.Module):
|
||||
return grid, anchor_grid
|
||||
|
||||
|
||||
class Model(nn.Module):
|
||||
# YOLOv5 model
|
||||
class BaseModel(nn.Module):
|
||||
# YOLOv5 base model
|
||||
def forward(self, x, profile=False, visualize=False):
|
||||
return self._forward_once(x, profile, visualize) # single-scale inference, train
|
||||
|
||||
def _forward_once(self, x, profile=False, visualize=False):
|
||||
y, dt = [], [] # outputs
|
||||
for m in self.model:
|
||||
if m.f != -1: # if not from previous layer
|
||||
x = y[m.f] if isinstance(m.f, int) else [x if j == -1 else y[j] for j in m.f] # from earlier layers
|
||||
if profile:
|
||||
self._profile_one_layer(m, x, dt)
|
||||
x = m(x) # run
|
||||
y.append(x if m.i in self.save else None) # save output
|
||||
if visualize:
|
||||
feature_visualization(x, m.type, m.i, save_dir=visualize)
|
||||
return x
|
||||
|
||||
def _profile_one_layer(self, m, x, dt):
|
||||
c = m == self.model[-1] # is final layer, copy input as inplace fix
|
||||
o = thop.profile(m, inputs=(x.copy() if c else x,), verbose=False)[0] / 1E9 * 2 if thop else 0 # FLOPs
|
||||
t = time_sync()
|
||||
for _ in range(10):
|
||||
m(x.copy() if c else x)
|
||||
dt.append((time_sync() - t) * 100)
|
||||
if m == self.model[0]:
|
||||
LOGGER.info(f"{'time (ms)':>10s} {'GFLOPs':>10s} {'params':>10s} module")
|
||||
LOGGER.info(f'{dt[-1]:10.2f} {o:10.2f} {m.np:10.0f} {m.type}')
|
||||
if c:
|
||||
LOGGER.info(f"{sum(dt):10.2f} {'-':>10s} {'-':>10s} Total")
|
||||
|
||||
def fuse(self): # fuse model Conv2d() + BatchNorm2d() layers
|
||||
LOGGER.info('Fusing layers... ')
|
||||
for m in self.model.modules():
|
||||
if isinstance(m, (Conv, DWConv)) and hasattr(m, 'bn'):
|
||||
m.conv = fuse_conv_and_bn(m.conv, m.bn) # update conv
|
||||
delattr(m, 'bn') # remove batchnorm
|
||||
m.forward = m.forward_fuse # update forward
|
||||
self.info()
|
||||
return self
|
||||
|
||||
def info(self, verbose=False, img_size=640): # print model information
|
||||
model_info(self, verbose, img_size)
|
||||
|
||||
def _apply(self, fn):
|
||||
# Apply to(), cpu(), cuda(), half() to model tensors that are not parameters or registered buffers
|
||||
self = super()._apply(fn)
|
||||
m = self.model[-1] # Detect()
|
||||
if isinstance(m, Detect):
|
||||
m.stride = fn(m.stride)
|
||||
m.grid = list(map(fn, m.grid))
|
||||
if isinstance(m.anchor_grid, list):
|
||||
m.anchor_grid = list(map(fn, m.anchor_grid))
|
||||
return self
|
||||
|
||||
|
||||
class DetectionModel(BaseModel):
|
||||
# YOLOv5 detection model
|
||||
def __init__(self, cfg='yolov5s.yaml', ch=3, nc=None, anchors=None): # model, input channels, number of classes
|
||||
super().__init__()
|
||||
if isinstance(cfg, dict):
|
||||
@@ -149,19 +205,6 @@ class Model(nn.Module):
|
||||
y = self._clip_augmented(y) # clip augmented tails
|
||||
return torch.cat(y, 1), None # augmented inference, train
|
||||
|
||||
def _forward_once(self, x, profile=False, visualize=False):
|
||||
y, dt = [], [] # outputs
|
||||
for m in self.model:
|
||||
if m.f != -1: # if not from previous layer
|
||||
x = y[m.f] if isinstance(m.f, int) else [x if j == -1 else y[j] for j in m.f] # from earlier layers
|
||||
if profile:
|
||||
self._profile_one_layer(m, x, dt)
|
||||
x = m(x) # run
|
||||
y.append(x if m.i in self.save else None) # save output
|
||||
if visualize:
|
||||
feature_visualization(x, m.type, m.i, save_dir=visualize)
|
||||
return x
|
||||
|
||||
def _descale_pred(self, p, flips, scale, img_size):
|
||||
# de-scale predictions following augmented inference (inverse operation)
|
||||
if self.inplace:
|
||||
@@ -190,19 +233,6 @@ class Model(nn.Module):
|
||||
y[-1] = y[-1][:, i:] # small
|
||||
return y
|
||||
|
||||
def _profile_one_layer(self, m, x, dt):
|
||||
c = isinstance(m, Detect) # is final layer, copy input as inplace fix
|
||||
o = thop.profile(m, inputs=(x.copy() if c else x,), verbose=False)[0] / 1E9 * 2 if thop else 0 # FLOPs
|
||||
t = time_sync()
|
||||
for _ in range(10):
|
||||
m(x.copy() if c else x)
|
||||
dt.append((time_sync() - t) * 100)
|
||||
if m == self.model[0]:
|
||||
LOGGER.info(f"{'time (ms)':>10s} {'GFLOPs':>10s} {'params':>10s} module")
|
||||
LOGGER.info(f'{dt[-1]:10.2f} {o:10.2f} {m.np:10.0f} {m.type}')
|
||||
if c:
|
||||
LOGGER.info(f"{sum(dt):10.2f} {'-':>10s} {'-':>10s} Total")
|
||||
|
||||
def _initialize_biases(self, cf=None): # initialize biases into Detect(), cf is class frequency
|
||||
# https://arxiv.org/abs/1708.02002 section 3.3
|
||||
# cf = torch.bincount(torch.tensor(np.concatenate(dataset.labels, 0)[:, 0]).long(), minlength=nc) + 1.
|
||||
@@ -213,41 +243,34 @@ class Model(nn.Module):
|
||||
b[:, 5:] += math.log(0.6 / (m.nc - 0.999999)) if cf is None else torch.log(cf / cf.sum()) # cls
|
||||
mi.bias = torch.nn.Parameter(b.view(-1), requires_grad=True)
|
||||
|
||||
def _print_biases(self):
|
||||
m = self.model[-1] # Detect() module
|
||||
for mi in m.m: # from
|
||||
b = mi.bias.detach().view(m.na, -1).T # conv.bias(255) to (3,85)
|
||||
LOGGER.info(
|
||||
('%6g Conv2d.bias:' + '%10.3g' * 6) % (mi.weight.shape[1], *b[:5].mean(1).tolist(), b[5:].mean()))
|
||||
|
||||
# def _print_weights(self):
|
||||
# for m in self.model.modules():
|
||||
# if type(m) is Bottleneck:
|
||||
# LOGGER.info('%10.3g' % (m.w.detach().sigmoid() * 2)) # shortcut weights
|
||||
Model = DetectionModel # retain YOLOv5 'Model' class for backwards compatibility
|
||||
|
||||
def fuse(self): # fuse model Conv2d() + BatchNorm2d() layers
|
||||
LOGGER.info('Fusing layers... ')
|
||||
for m in self.model.modules():
|
||||
if isinstance(m, (Conv, DWConv)) and hasattr(m, 'bn'):
|
||||
m.conv = fuse_conv_and_bn(m.conv, m.bn) # update conv
|
||||
delattr(m, 'bn') # remove batchnorm
|
||||
m.forward = m.forward_fuse # update forward
|
||||
self.info()
|
||||
return self
|
||||
|
||||
def info(self, verbose=False, img_size=640): # print model information
|
||||
model_info(self, verbose, img_size)
|
||||
class ClassificationModel(BaseModel):
|
||||
# YOLOv5 classification model
|
||||
def __init__(self, cfg=None, model=None, nc=1000, cutoff=10): # yaml, model, number of classes, cutoff index
|
||||
super().__init__()
|
||||
self._from_detection_model(model, nc, cutoff) if model is not None else self._from_yaml(cfg)
|
||||
|
||||
def _apply(self, fn):
|
||||
# Apply to(), cpu(), cuda(), half() to model tensors that are not parameters or registered buffers
|
||||
self = super()._apply(fn)
|
||||
m = self.model[-1] # Detect()
|
||||
if isinstance(m, Detect):
|
||||
m.stride = fn(m.stride)
|
||||
m.grid = list(map(fn, m.grid))
|
||||
if isinstance(m.anchor_grid, list):
|
||||
m.anchor_grid = list(map(fn, m.anchor_grid))
|
||||
return self
|
||||
def _from_detection_model(self, model, nc=1000, cutoff=10):
|
||||
# Create a YOLOv5 classification model from a YOLOv5 detection model
|
||||
if isinstance(model, DetectMultiBackend):
|
||||
model = model.model # unwrap DetectMultiBackend
|
||||
model.model = model.model[:cutoff] # backbone
|
||||
m = model.model[-1] # last layer
|
||||
ch = m.conv.in_channels if hasattr(m, 'conv') else m.cv1.conv.in_channels # ch into module
|
||||
c = Classify(ch, nc) # Classify()
|
||||
c.i, c.f, c.type = m.i, m.f, 'models.common.Classify' # index, from, type
|
||||
model.model[-1] = c # replace
|
||||
self.model = model.model
|
||||
self.stride = model.stride
|
||||
self.save = []
|
||||
self.nc = nc
|
||||
|
||||
def _from_yaml(self, cfg):
|
||||
# Create a YOLOv5 classification model from a *.yaml file
|
||||
self.model = None
|
||||
|
||||
|
||||
def parse_model(d, ch): # model_dict, input_channels(3)
|
||||
@@ -321,7 +344,7 @@ if __name__ == '__main__':
|
||||
|
||||
# Options
|
||||
if opt.line_profile: # profile layer by layer
|
||||
_ = model(im, profile=True)
|
||||
model(im, profile=True)
|
||||
|
||||
elif opt.profile: # profile forward-backward
|
||||
results = profile(input=im, ops=[model], n=3)
|
||||
|
||||
Reference in new issue
Block a user