Walkthru 12
walkthru 12
• 10 min read
import os
os.environ['CUDA_VISIBLE_DEVICES'] = "0"
from fastai.vision.all import *
from fastcore.parallel import *
path = Path()
trn_path = path/'train_images'
arch ='convnext_tiny_in22k'
df = pd.read_csv(path/'train.csv')
df
| image_id | label | variety | age | |
|---|---|---|---|---|
| 0 | 100330.jpg | bacterial_leaf_blight | ADT45 | 45 |
| 1 | 100365.jpg | bacterial_leaf_blight | ADT45 | 45 |
| 2 | 100382.jpg | bacterial_leaf_blight | ADT45 | 45 |
| 3 | 100632.jpg | bacterial_leaf_blight | ADT45 | 45 |
| 4 | 101918.jpg | bacterial_leaf_blight | ADT45 | 45 |
| ... | ... | ... | ... | ... |
| 10402 | 107607.jpg | tungro | Zonal | 55 |
| 10403 | 107811.jpg | tungro | Zonal | 55 |
| 10404 | 108547.jpg | tungro | Zonal | 55 |
| 10405 | 110245.jpg | tungro | Zonal | 55 |
| 10406 | 110381.jpg | tungro | Zonal | 55 |
10407 rows × 4 columns
dls = ImageDataLoaders.from_folder(
trn_path, valid_pct=.2, seed=42, item_tfms=Resize(224),batch=aug_transforms(size=224, min_scale= 0.75),bs=32)
learn = vision_learner(dls, arch, metrics=error_rate)
learn.fine_tune(1, 0.02)
| epoch | train_loss | valid_loss | error_rate | time |
|---|---|---|---|---|
| 0 | 1.392303 | 0.888152 | 0.271985 | 00:38 |
| epoch | train_loss | valid_loss | error_rate | time |
|---|---|---|---|---|
| 0 | 0.507126 | 0.295987 | 0.097069 | 01:55 |
m = learn.model
m
Sequential(
(0): TimmBody(
(model): ConvNeXt(
(stem): Sequential(
(0): Conv2d(3, 96, kernel_size=(4, 4), stride=(4, 4))
(1): LayerNorm2d((96,), eps=1e-06, elementwise_affine=True)
)
(stages): Sequential(
(0): ConvNeXtStage(
(downsample): Identity()
(blocks): Sequential(
(0): ConvNeXtBlock(
(conv_dw): Conv2d(96, 96, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3), groups=96)
(norm): LayerNorm((96,), eps=1e-06, elementwise_affine=True)
(mlp): Mlp(
(fc1): Linear(in_features=96, out_features=384, bias=True)
(act): GELU()
(drop1): Dropout(p=0.0, inplace=False)
(fc2): Linear(in_features=384, out_features=96, bias=True)
(drop2): Dropout(p=0.0, inplace=False)
)
(drop_path): Identity()
)
(1): ConvNeXtBlock(
(conv_dw): Conv2d(96, 96, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3), groups=96)
(norm): LayerNorm((96,), eps=1e-06, elementwise_affine=True)
(mlp): Mlp(
(fc1): Linear(in_features=96, out_features=384, bias=True)
(act): GELU()
(drop1): Dropout(p=0.0, inplace=False)
(fc2): Linear(in_features=384, out_features=96, bias=True)
(drop2): Dropout(p=0.0, inplace=False)
)
(drop_path): Identity()
)
(2): ConvNeXtBlock(
(conv_dw): Conv2d(96, 96, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3), groups=96)
(norm): LayerNorm((96,), eps=1e-06, elementwise_affine=True)
(mlp): Mlp(
(fc1): Linear(in_features=96, out_features=384, bias=True)
(act): GELU()
(drop1): Dropout(p=0.0, inplace=False)
(fc2): Linear(in_features=384, out_features=96, bias=True)
(drop2): Dropout(p=0.0, inplace=False)
)
(drop_path): Identity()
)
)
)
(1): ConvNeXtStage(
(downsample): Sequential(
(0): LayerNorm2d((96,), eps=1e-06, elementwise_affine=True)
(1): Conv2d(96, 192, kernel_size=(2, 2), stride=(2, 2))
)
(blocks): Sequential(
(0): ConvNeXtBlock(
(conv_dw): Conv2d(192, 192, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3), groups=192)
(norm): LayerNorm((192,), eps=1e-06, elementwise_affine=True)
(mlp): Mlp(
(fc1): Linear(in_features=192, out_features=768, bias=True)
(act): GELU()
(drop1): Dropout(p=0.0, inplace=False)
(fc2): Linear(in_features=768, out_features=192, bias=True)
(drop2): Dropout(p=0.0, inplace=False)
)
(drop_path): Identity()
)
(1): ConvNeXtBlock(
(conv_dw): Conv2d(192, 192, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3), groups=192)
(norm): LayerNorm((192,), eps=1e-06, elementwise_affine=True)
(mlp): Mlp(
(fc1): Linear(in_features=192, out_features=768, bias=True)
(act): GELU()
(drop1): Dropout(p=0.0, inplace=False)
(fc2): Linear(in_features=768, out_features=192, bias=True)
(drop2): Dropout(p=0.0, inplace=False)
)
(drop_path): Identity()
)
(2): ConvNeXtBlock(
(conv_dw): Conv2d(192, 192, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3), groups=192)
(norm): LayerNorm((192,), eps=1e-06, elementwise_affine=True)
(mlp): Mlp(
(fc1): Linear(in_features=192, out_features=768, bias=True)
(act): GELU()
(drop1): Dropout(p=0.0, inplace=False)
(fc2): Linear(in_features=768, out_features=192, bias=True)
(drop2): Dropout(p=0.0, inplace=False)
)
(drop_path): Identity()
)
)
)
(2): ConvNeXtStage(
(downsample): Sequential(
(0): LayerNorm2d((192,), eps=1e-06, elementwise_affine=True)
(1): Conv2d(192, 384, kernel_size=(2, 2), stride=(2, 2))
)
(blocks): Sequential(
(0): ConvNeXtBlock(
(conv_dw): Conv2d(384, 384, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3), groups=384)
(norm): LayerNorm((384,), eps=1e-06, elementwise_affine=True)
(mlp): Mlp(
(fc1): Linear(in_features=384, out_features=1536, bias=True)
(act): GELU()
(drop1): Dropout(p=0.0, inplace=False)
(fc2): Linear(in_features=1536, out_features=384, bias=True)
(drop2): Dropout(p=0.0, inplace=False)
)
(drop_path): Identity()
)
(1): ConvNeXtBlock(
(conv_dw): Conv2d(384, 384, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3), groups=384)
(norm): LayerNorm((384,), eps=1e-06, elementwise_affine=True)
(mlp): Mlp(
(fc1): Linear(in_features=384, out_features=1536, bias=True)
(act): GELU()
(drop1): Dropout(p=0.0, inplace=False)
(fc2): Linear(in_features=1536, out_features=384, bias=True)
(drop2): Dropout(p=0.0, inplace=False)
)
(drop_path): Identity()
)
(2): ConvNeXtBlock(
(conv_dw): Conv2d(384, 384, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3), groups=384)
(norm): LayerNorm((384,), eps=1e-06, elementwise_affine=True)
(mlp): Mlp(
(fc1): Linear(in_features=384, out_features=1536, bias=True)
(act): GELU()
(drop1): Dropout(p=0.0, inplace=False)
(fc2): Linear(in_features=1536, out_features=384, bias=True)
(drop2): Dropout(p=0.0, inplace=False)
)
(drop_path): Identity()
)
(3): ConvNeXtBlock(
(conv_dw): Conv2d(384, 384, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3), groups=384)
(norm): LayerNorm((384,), eps=1e-06, elementwise_affine=True)
(mlp): Mlp(
(fc1): Linear(in_features=384, out_features=1536, bias=True)
(act): GELU()
(drop1): Dropout(p=0.0, inplace=False)
(fc2): Linear(in_features=1536, out_features=384, bias=True)
(drop2): Dropout(p=0.0, inplace=False)
)
(drop_path): Identity()
)
(4): ConvNeXtBlock(
(conv_dw): Conv2d(384, 384, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3), groups=384)
(norm): LayerNorm((384,), eps=1e-06, elementwise_affine=True)
(mlp): Mlp(
(fc1): Linear(in_features=384, out_features=1536, bias=True)
(act): GELU()
(drop1): Dropout(p=0.0, inplace=False)
(fc2): Linear(in_features=1536, out_features=384, bias=True)
(drop2): Dropout(p=0.0, inplace=False)
)
(drop_path): Identity()
)
(5): ConvNeXtBlock(
(conv_dw): Conv2d(384, 384, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3), groups=384)
(norm): LayerNorm((384,), eps=1e-06, elementwise_affine=True)
(mlp): Mlp(
(fc1): Linear(in_features=384, out_features=1536, bias=True)
(act): GELU()
(drop1): Dropout(p=0.0, inplace=False)
(fc2): Linear(in_features=1536, out_features=384, bias=True)
(drop2): Dropout(p=0.0, inplace=False)
)
(drop_path): Identity()
)
(6): ConvNeXtBlock(
(conv_dw): Conv2d(384, 384, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3), groups=384)
(norm): LayerNorm((384,), eps=1e-06, elementwise_affine=True)
(mlp): Mlp(
(fc1): Linear(in_features=384, out_features=1536, bias=True)
(act): GELU()
(drop1): Dropout(p=0.0, inplace=False)
(fc2): Linear(in_features=1536, out_features=384, bias=True)
(drop2): Dropout(p=0.0, inplace=False)
)
(drop_path): Identity()
)
(7): ConvNeXtBlock(
(conv_dw): Conv2d(384, 384, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3), groups=384)
(norm): LayerNorm((384,), eps=1e-06, elementwise_affine=True)
(mlp): Mlp(
(fc1): Linear(in_features=384, out_features=1536, bias=True)
(act): GELU()
(drop1): Dropout(p=0.0, inplace=False)
(fc2): Linear(in_features=1536, out_features=384, bias=True)
(drop2): Dropout(p=0.0, inplace=False)
)
(drop_path): Identity()
)
(8): ConvNeXtBlock(
(conv_dw): Conv2d(384, 384, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3), groups=384)
(norm): LayerNorm((384,), eps=1e-06, elementwise_affine=True)
(mlp): Mlp(
(fc1): Linear(in_features=384, out_features=1536, bias=True)
(act): GELU()
(drop1): Dropout(p=0.0, inplace=False)
(fc2): Linear(in_features=1536, out_features=384, bias=True)
(drop2): Dropout(p=0.0, inplace=False)
)
(drop_path): Identity()
)
)
)
(3): ConvNeXtStage(
(downsample): Sequential(
(0): LayerNorm2d((384,), eps=1e-06, elementwise_affine=True)
(1): Conv2d(384, 768, kernel_size=(2, 2), stride=(2, 2))
)
(blocks): Sequential(
(0): ConvNeXtBlock(
(conv_dw): Conv2d(768, 768, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3), groups=768)
(norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
(mlp): Mlp(
(fc1): Linear(in_features=768, out_features=3072, bias=True)
(act): GELU()
(drop1): Dropout(p=0.0, inplace=False)
(fc2): Linear(in_features=3072, out_features=768, bias=True)
(drop2): Dropout(p=0.0, inplace=False)
)
(drop_path): Identity()
)
(1): ConvNeXtBlock(
(conv_dw): Conv2d(768, 768, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3), groups=768)
(norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
(mlp): Mlp(
(fc1): Linear(in_features=768, out_features=3072, bias=True)
(act): GELU()
(drop1): Dropout(p=0.0, inplace=False)
(fc2): Linear(in_features=3072, out_features=768, bias=True)
(drop2): Dropout(p=0.0, inplace=False)
)
(drop_path): Identity()
)
(2): ConvNeXtBlock(
(conv_dw): Conv2d(768, 768, kernel_size=(7, 7), stride=(1, 1), padding=(3, 3), groups=768)
(norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
(mlp): Mlp(
(fc1): Linear(in_features=768, out_features=3072, bias=True)
(act): GELU()
(drop1): Dropout(p=0.0, inplace=False)
(fc2): Linear(in_features=3072, out_features=768, bias=True)
(drop2): Dropout(p=0.0, inplace=False)
)
(drop_path): Identity()
)
)
)
)
(norm_pre): Identity()
(head): Sequential(
(global_pool): SelectAdaptivePool2d (pool_type=avg, flatten=Identity())
(norm): LayerNorm2d((768,), eps=1e-06, elementwise_affine=True)
(flatten): Flatten(start_dim=1, end_dim=-1)
(drop): Dropout(p=0.0, inplace=False)
(fc): Identity()
)
)
)
(1): Sequential(
(0): AdaptiveConcatPool2d(
(ap): AdaptiveAvgPool2d(output_size=1)
(mp): AdaptiveMaxPool2d(output_size=1)
)
(1): Flatten(full=False)
(2): BatchNorm1d(1536, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(3): Dropout(p=0.25, inplace=False)
(4): Linear(in_features=1536, out_features=512, bias=False)
(5): ReLU(inplace=True)
(6): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(7): Dropout(p=0.5, inplace=False)
(8): Linear(in_features=512, out_features=10, bias=False)
)
)
learn.summary()
Sequential (Input shape: 32 x 3 x 224 x 224)
============================================================================
Layer (type) Output Shape Param # Trainable
============================================================================
32 x 96 x 56 x 56
Conv2d 4704 True
LayerNorm2d 192 True
Identity
Conv2d 4800 True
LayerNorm 192 True
____________________________________________________________________________
32 x 56 x 56 x 384
Linear 37248 True
GELU
Dropout
____________________________________________________________________________
32 x 56 x 56 x 96
Linear 36960 True
Dropout
Identity
Conv2d 4800 True
LayerNorm 192 True
____________________________________________________________________________
32 x 56 x 56 x 384
Linear 37248 True
GELU
Dropout
____________________________________________________________________________
32 x 56 x 56 x 96
Linear 36960 True
Dropout
Identity
Conv2d 4800 True
LayerNorm 192 True
____________________________________________________________________________
32 x 56 x 56 x 384
Linear 37248 True
GELU
Dropout
____________________________________________________________________________
32 x 56 x 56 x 96
Linear 36960 True
Dropout
Identity
LayerNorm2d 192 True
____________________________________________________________________________
32 x 192 x 28 x 28
Conv2d 73920 True
Conv2d 9600 True
LayerNorm 384 True
____________________________________________________________________________
32 x 28 x 28 x 768
Linear 148224 True
GELU
Dropout
____________________________________________________________________________
32 x 28 x 28 x 192
Linear 147648 True
Dropout
Identity
Conv2d 9600 True
LayerNorm 384 True
____________________________________________________________________________
32 x 28 x 28 x 768
Linear 148224 True
GELU
Dropout
____________________________________________________________________________
32 x 28 x 28 x 192
Linear 147648 True
Dropout
Identity
Conv2d 9600 True
LayerNorm 384 True
____________________________________________________________________________
32 x 28 x 28 x 768
Linear 148224 True
GELU
Dropout
____________________________________________________________________________
32 x 28 x 28 x 192
Linear 147648 True
Dropout
Identity
LayerNorm2d 384 True
____________________________________________________________________________
32 x 384 x 14 x 14
Conv2d 295296 True
Conv2d 19200 True
LayerNorm 768 True
____________________________________________________________________________
32 x 14 x 14 x 1536
Linear 591360 True
GELU
Dropout
____________________________________________________________________________
32 x 14 x 14 x 384
Linear 590208 True
Dropout
Identity
Conv2d 19200 True
LayerNorm 768 True
____________________________________________________________________________
32 x 14 x 14 x 1536
Linear 591360 True
GELU
Dropout
____________________________________________________________________________
32 x 14 x 14 x 384
Linear 590208 True
Dropout
Identity
Conv2d 19200 True
LayerNorm 768 True
____________________________________________________________________________
32 x 14 x 14 x 1536
Linear 591360 True
GELU
Dropout
____________________________________________________________________________
32 x 14 x 14 x 384
Linear 590208 True
Dropout
Identity
Conv2d 19200 True
LayerNorm 768 True
____________________________________________________________________________
32 x 14 x 14 x 1536
Linear 591360 True
GELU
Dropout
____________________________________________________________________________
32 x 14 x 14 x 384
Linear 590208 True
Dropout
Identity
Conv2d 19200 True
LayerNorm 768 True
____________________________________________________________________________
32 x 14 x 14 x 1536
Linear 591360 True
GELU
Dropout
____________________________________________________________________________
32 x 14 x 14 x 384
Linear 590208 True
Dropout
Identity
Conv2d 19200 True
LayerNorm 768 True
____________________________________________________________________________
32 x 14 x 14 x 1536
Linear 591360 True
GELU
Dropout
____________________________________________________________________________
32 x 14 x 14 x 384
Linear 590208 True
Dropout
Identity
Conv2d 19200 True
LayerNorm 768 True
____________________________________________________________________________
32 x 14 x 14 x 1536
Linear 591360 True
GELU
Dropout
____________________________________________________________________________
32 x 14 x 14 x 384
Linear 590208 True
Dropout
Identity
Conv2d 19200 True
LayerNorm 768 True
____________________________________________________________________________
32 x 14 x 14 x 1536
Linear 591360 True
GELU
Dropout
____________________________________________________________________________
32 x 14 x 14 x 384
Linear 590208 True
Dropout
Identity
Conv2d 19200 True
LayerNorm 768 True
____________________________________________________________________________
32 x 14 x 14 x 1536
Linear 591360 True
GELU
Dropout
____________________________________________________________________________
32 x 14 x 14 x 384
Linear 590208 True
Dropout
Identity
LayerNorm2d 768 True
____________________________________________________________________________
32 x 768 x 7 x 7
Conv2d 1180416 True
Conv2d 38400 True
LayerNorm 1536 True
____________________________________________________________________________
32 x 7 x 7 x 3072
Linear 2362368 True
GELU
Dropout
____________________________________________________________________________
32 x 7 x 7 x 768
Linear 2360064 True
Dropout
Identity
Conv2d 38400 True
LayerNorm 1536 True
____________________________________________________________________________
32 x 7 x 7 x 3072
Linear 2362368 True
GELU
Dropout
____________________________________________________________________________
32 x 7 x 7 x 768
Linear 2360064 True
Dropout
Identity
Conv2d 38400 True
LayerNorm 1536 True
____________________________________________________________________________
32 x 7 x 7 x 3072
Linear 2362368 True
GELU
Dropout
____________________________________________________________________________
32 x 7 x 7 x 768
Linear 2360064 True
Dropout
Identity
Identity
____________________________________________________________________________
32 x 768 x 1 x 1
AdaptiveAvgPool2d
AdaptiveMaxPool2d
____________________________________________________________________________
32 x 1536
Flatten
BatchNorm1d 3072 True
Dropout
____________________________________________________________________________
32 x 512
Linear 786432 True
ReLU
BatchNorm1d 1024 True
Dropout
____________________________________________________________________________
Total params: 28,602,496
Total trainable params: 28,602,496
Total non-trainable params: 0
Optimizer used: <function Adam at 0x7fba154d2820>
Loss function: FlattenedLoss of CrossEntropyLoss()
Model unfrozen
Callbacks:
- TrainEvalCallback
- Recorder
- ProgressCallback
h= m[1]
h
Sequential(
(0): AdaptiveConcatPool2d(
(ap): AdaptiveAvgPool2d(output_size=1)
(mp): AdaptiveMaxPool2d(output_size=1)
)
(1): Flatten(full=False)
(2): BatchNorm1d(1536, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(3): Dropout(p=0.25, inplace=False)
(4): Linear(in_features=1536, out_features=512, bias=False)
(5): ReLU(inplace=True)
(6): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(7): Dropout(p=0.5, inplace=False)
(8): Linear(in_features=512, out_features=10, bias=False)
)
ll =h[-1]
ll
Linear(in_features=512, out_features=10, bias=False)
ll.parameters()
<generator object Module.parameters at 0x7f3eb07f6510>
llp=list(ll.parameters())[0]
llp
Parameter containing:
tensor([[-0.0744, 0.0336, 0.0605, ..., -0.0420, 0.0424, -0.0034],
[-0.1795, -0.2076, 0.1650, ..., -0.0621, -0.0006, 0.0082],
[ 0.1246, 0.1805, -0.0646, ..., 0.0427, -0.0914, -0.0676],
...,
[-0.0006, 0.0500, 0.1109, ..., -0.1318, -0.1138, 0.0212],
[-0.0090, 0.0812, -0.0604, ..., 0.1045, 0.2306, 0.1239],
[-0.0352, -0.0684, 0.1508, ..., 0.0299, 0.0247, -0.0539]],
device='cuda:0', requires_grad=True)
llp.shape
torch.Size([10, 512])
ll
Linear(in_features=512, out_features=10, bias=False)
from copy import deepcopy
dls = ImageDataLoaders.from_folder(trn_path, valid_pct=.2, seed=42, item_tfms=Resize(224),batch=aug_transforms(size=224, min_scale= 0.75))
learn = vision_learner(dls, arch)#, metrics=error_rate) # If you miss this part, you'll get an error! metrics=error_rate creates a problem
learn2 = deepcopy(learn)
curr_loss=learn2.loss_func
def dtc_loss(preds,targs):
rice_preds,dis_preds = preds
return curr_loss(dis_preds,targs)
class DiseaseAndTypeClassifier(nn.Module):
def __init__(self,m):
super().__init__()
self.l1 = nn.Linear(in_features=512, out_features=10, bias=False) #rice type
self.l2 = nn.Linear(in_features=512, out_features=10, bias=False) #disease
del(m[1][-1]) #it removes the last layer
self.m = m
def forward (self,x):
x = self.m(x)
x1 = self.l1(x)
x2 = self.l2(x)
return x1,x2
dtc = DiseaseAndTypeClassifier(learn2.model)
learn2.model = dtc
learn2.loss_func=dtc_loss
preds,targs = learn2.get_preds(dl=learn2.dls.valid)
rice_preds, dis_preds = preds
rice_preds.shape,dis_preds.shape
(torch.Size([2081, 10]), torch.Size([2081, 10]))
need to check Fastai create head function