MareArts Computer Vision Study.

yaml to Dict and update element like class and save dict to yaml again, python example code

May 28, 2023

refer to code and example yaml

before yaml to run code

a: a-value
b: b-value
c:
  d: Nested
  e: Values

code

pip install yaml, attract

import yaml
from attrdict import AttrDict

# Load YAML file
with open('file.yaml', 'r') as f:
    data = yaml.safe_load(f)

# Convert to AttrDict for dot notation access
data = AttrDict(data)

# Access elements
print(data.a)
print(data.b)

# Modify elements
data.a = 'new value 2'
data.b = 'another new value 2'

# Convert back to dictionary
data = dict(data)

# Save back to YAML
with open('file.yaml', 'w') as f:
    yaml.safe_dump(data, f)

after yaml running code

a: new value 2
b: another new value 2
c:
  d: Nested
  e: Values

Thank you.

www.marearts.com

🙇🏻‍♂️

Create custom tokenizer simple code.

May 23, 2023

In the sample code, vocabulary is "0,1,2,3,4" and max length is 20.

from typing import List, Union

class CustomTokenizer:
    def __init__(self, vocab: Union[str, List[str]], pad_token="<PAD>", cls_token="<BOS>", sep_token="<SEP>", max_len=20):
        if isinstance(vocab, str):
            with open(vocab, 'r') as f:
                self.vocab = {word.strip(): i for i, word in enumerate(f.readlines())}
        elif isinstance(vocab, list):
            self.vocab = {word: i for i, word in enumerate(vocab)}
        else:
            raise ValueError("vocab must be either a filepath (str) or a list of words")
        
        print('vocab: ', self.vocab)
        self.pad_token = pad_token
        self.cls_token = cls_token
        self.sep_token = sep_token
        self.max_len = max_len
        self.inv_vocab = {v: k for k, v in self.vocab.items()}

    def tokenize(self, text: str):
        tokens = [c for c in text if c in self.vocab]
        tokens = tokens[:self.max_len]
        padding_length = self.max_len - len(tokens)
        return [self.cls_token] + tokens + [self.sep_token] + [self.pad_token] * padding_length

    def convert_tokens_to_ids(self, tokens):
        return [self.vocab.get(token, self.vocab.get(self.pad_token)) for token in tokens]

    def convert_ids_to_tokens(self, ids):
        return [self.inv_vocab.get(id, self.pad_token) for id in ids]



vocab = ["<PAD>", "<BOS>", "<SEP>", "0", "1", "2", "3", "4"]
with open('vocab.txt', 'w') as f:
    for token in vocab:
        f.write(token + '\n')

# Initialize your custom tokenizer
tokenizer = CustomTokenizer(vocab='vocab.txt')

# Now you can use this tokenizer to tokenize your data, study.marearts.com
tokenized_text = tokenizer.tokenize('22342')
print("tokenized_text: ", tokenized_text)

# Convert tokens to ids
token_ids = tokenizer.convert_tokens_to_ids(tokenized_text)
print("token_ids: ", token_ids)

# Convert ids back to tokens, marearts.com
tokens = tokenizer.convert_ids_to_tokens(token_ids)
print("tokens: ", tokens)

Thank you.

🙇🏻‍♂️

timm swin transformer v2 model review

May 20, 2023

refer to code:

import timm
import torch
from PIL import Image
from timm.data import resolve_data_config
from timm.data.transforms_factory import create_transform

# Load from Hub 🔥
model = timm.create_model(
    'hf-hub:nateraw/resnet50-oxford-iiit-pet',
    pretrained=True
)

# Set model to eval mode for inference
model.eval()

#resolve
print(resolve_data_config(model.pretrained_cfg, model=model))

# Get the labels from the model config
print(model.pretrained_cfg)

# labels
labels = model.pretrained_cfg['label_names']
print(labels)

top_k = min(len(labels), 5)

# Create Transform
transform = create_transform(**resolve_data_config(model.pretrained_cfg, model=model))

# Use your own image file here...
image = Image.open('boxer.jpg').convert('RGB')

# Process PIL image with transforms and add a batch dimension
x = transform(image).unsqueeze(0)

print(x.shape)
# Pass inputs to model forward function to get outputs
out = model(x)
print(out.shape)

# Apply softmax to get predicted probabilities for each class
probabilities = torch.nn.functional.softmax(out[0], dim=0)

print(probabilities)

# Grab the values and indices of top 5 predicted classes
values, indices = torch.topk(probabilities, top_k)

# Prepare a nice dict of top k predictions
predictions = [
    {"label": labels[i], "score": v.item()}
    for i, v in zip(indices, values)
]
print(predictions)

from torchsummary import summary
input_size = (3, 224, 224)
# Use torchsummary to print the model summary
summary(model, input_size)


# Remove the last layer
model = torch.nn.Sequential(*list(model.children())[:-1])
input_size = (3, 224, 224)
# Use torchsummary to print the model summary
summary(model, input_size)

First Summary:

----------------------------------------------------------------

Layer (type) Output Shape Param #

================================================================

Conv2d-1 [-1, 64, 112, 112] 9,408

BatchNorm2d-2 [-1, 64, 112, 112] 128

ReLU-3 [-1, 64, 112, 112] 0

MaxPool2d-4 [-1, 64, 56, 56] 0

Conv2d-5 [-1, 64, 56, 56] 4,096

BatchNorm2d-6 [-1, 64, 56, 56] 128

ReLU-7 [-1, 64, 56, 56] 0

Conv2d-8 [-1, 64, 56, 56] 36,864

BatchNorm2d-9 [-1, 64, 56, 56] 128

Identity-10 [-1, 64, 56, 56] 0

ReLU-11 [-1, 64, 56, 56] 0

Identity-12 [-1, 64, 56, 56] 0

Conv2d-13 [-1, 256, 56, 56] 16,384

BatchNorm2d-14 [-1, 256, 56, 56] 512

Conv2d-15 [-1, 256, 56, 56] 16,384

BatchNorm2d-16 [-1, 256, 56, 56] 512

ReLU-17 [-1, 256, 56, 56] 0

Bottleneck-18 [-1, 256, 56, 56] 0

Conv2d-19 [-1, 64, 56, 56] 16,384

BatchNorm2d-20 [-1, 64, 56, 56] 128

ReLU-21 [-1, 64, 56, 56] 0

Conv2d-22 [-1, 64, 56, 56] 36,864

BatchNorm2d-23 [-1, 64, 56, 56] 128

Identity-24 [-1, 64, 56, 56] 0

ReLU-25 [-1, 64, 56, 56] 0

Identity-26 [-1, 64, 56, 56] 0

Conv2d-27 [-1, 256, 56, 56] 16,384

BatchNorm2d-28 [-1, 256, 56, 56] 512

ReLU-29 [-1, 256, 56, 56] 0

Bottleneck-30 [-1, 256, 56, 56] 0

Conv2d-31 [-1, 64, 56, 56] 16,384

BatchNorm2d-32 [-1, 64, 56, 56] 128

ReLU-33 [-1, 64, 56, 56] 0

Conv2d-34 [-1, 64, 56, 56] 36,864

BatchNorm2d-35 [-1, 64, 56, 56] 128

Identity-36 [-1, 64, 56, 56] 0

ReLU-37 [-1, 64, 56, 56] 0

Identity-38 [-1, 64, 56, 56] 0

Conv2d-39 [-1, 256, 56, 56] 16,384

BatchNorm2d-40 [-1, 256, 56, 56] 512

ReLU-41 [-1, 256, 56, 56] 0

Bottleneck-42 [-1, 256, 56, 56] 0

Conv2d-43 [-1, 128, 56, 56] 32,768

BatchNorm2d-44 [-1, 128, 56, 56] 256

ReLU-45 [-1, 128, 56, 56] 0

Conv2d-46 [-1, 128, 28, 28] 147,456

BatchNorm2d-47 [-1, 128, 28, 28] 256

Identity-48 [-1, 128, 28, 28] 0

ReLU-49 [-1, 128, 28, 28] 0

Identity-50 [-1, 128, 28, 28] 0

Conv2d-51 [-1, 512, 28, 28] 65,536

BatchNorm2d-52 [-1, 512, 28, 28] 1,024

Conv2d-53 [-1, 512, 28, 28] 131,072

BatchNorm2d-54 [-1, 512, 28, 28] 1,024

ReLU-55 [-1, 512, 28, 28] 0

Bottleneck-56 [-1, 512, 28, 28] 0

Conv2d-57 [-1, 128, 28, 28] 65,536

BatchNorm2d-58 [-1, 128, 28, 28] 256

ReLU-59 [-1, 128, 28, 28] 0

Conv2d-60 [-1, 128, 28, 28] 147,456

BatchNorm2d-61 [-1, 128, 28, 28] 256

Identity-62 [-1, 128, 28, 28] 0

ReLU-63 [-1, 128, 28, 28] 0

Identity-64 [-1, 128, 28, 28] 0

Conv2d-65 [-1, 512, 28, 28] 65,536

BatchNorm2d-66 [-1, 512, 28, 28] 1,024

ReLU-67 [-1, 512, 28, 28] 0

Bottleneck-68 [-1, 512, 28, 28] 0

Conv2d-69 [-1, 128, 28, 28] 65,536

BatchNorm2d-70 [-1, 128, 28, 28] 256

ReLU-71 [-1, 128, 28, 28] 0

Conv2d-72 [-1, 128, 28, 28] 147,456

BatchNorm2d-73 [-1, 128, 28, 28] 256

Identity-74 [-1, 128, 28, 28] 0

ReLU-75 [-1, 128, 28, 28] 0

Identity-76 [-1, 128, 28, 28] 0

Conv2d-77 [-1, 512, 28, 28] 65,536

BatchNorm2d-78 [-1, 512, 28, 28] 1,024

ReLU-79 [-1, 512, 28, 28] 0

Bottleneck-80 [-1, 512, 28, 28] 0

Conv2d-81 [-1, 128, 28, 28] 65,536

BatchNorm2d-82 [-1, 128, 28, 28] 256

ReLU-83 [-1, 128, 28, 28] 0

Conv2d-84 [-1, 128, 28, 28] 147,456

BatchNorm2d-85 [-1, 128, 28, 28] 256

Identity-86 [-1, 128, 28, 28] 0

ReLU-87 [-1, 128, 28, 28] 0

Identity-88 [-1, 128, 28, 28] 0

Conv2d-89 [-1, 512, 28, 28] 65,536

BatchNorm2d-90 [-1, 512, 28, 28] 1,024

ReLU-91 [-1, 512, 28, 28] 0

Bottleneck-92 [-1, 512, 28, 28] 0

Conv2d-93 [-1, 256, 28, 28] 131,072

BatchNorm2d-94 [-1, 256, 28, 28] 512

ReLU-95 [-1, 256, 28, 28] 0

Conv2d-96 [-1, 256, 14, 14] 589,824

BatchNorm2d-97 [-1, 256, 14, 14] 512

Identity-98 [-1, 256, 14, 14] 0

ReLU-99 [-1, 256, 14, 14] 0

Identity-100 [-1, 256, 14, 14] 0

Conv2d-101 [-1, 1024, 14, 14] 262,144

BatchNorm2d-102 [-1, 1024, 14, 14] 2,048

Conv2d-103 [-1, 1024, 14, 14] 524,288

BatchNorm2d-104 [-1, 1024, 14, 14] 2,048

ReLU-105 [-1, 1024, 14, 14] 0

Bottleneck-106 [-1, 1024, 14, 14] 0

Conv2d-107 [-1, 256, 14, 14] 262,144

BatchNorm2d-108 [-1, 256, 14, 14] 512

ReLU-109 [-1, 256, 14, 14] 0

Conv2d-110 [-1, 256, 14, 14] 589,824

BatchNorm2d-111 [-1, 256, 14, 14] 512

Identity-112 [-1, 256, 14, 14] 0

ReLU-113 [-1, 256, 14, 14] 0

Identity-114 [-1, 256, 14, 14] 0

Conv2d-115 [-1, 1024, 14, 14] 262,144

BatchNorm2d-116 [-1, 1024, 14, 14] 2,048

ReLU-117 [-1, 1024, 14, 14] 0

Bottleneck-118 [-1, 1024, 14, 14] 0

Conv2d-119 [-1, 256, 14, 14] 262,144

BatchNorm2d-120 [-1, 256, 14, 14] 512

ReLU-121 [-1, 256, 14, 14] 0

Conv2d-122 [-1, 256, 14, 14] 589,824

BatchNorm2d-123 [-1, 256, 14, 14] 512

Identity-124 [-1, 256, 14, 14] 0

ReLU-125 [-1, 256, 14, 14] 0

Identity-126 [-1, 256, 14, 14] 0

Conv2d-127 [-1, 1024, 14, 14] 262,144

BatchNorm2d-128 [-1, 1024, 14, 14] 2,048

ReLU-129 [-1, 1024, 14, 14] 0

Bottleneck-130 [-1, 1024, 14, 14] 0

Conv2d-131 [-1, 256, 14, 14] 262,144

BatchNorm2d-132 [-1, 256, 14, 14] 512

ReLU-133 [-1, 256, 14, 14] 0

Conv2d-134 [-1, 256, 14, 14] 589,824

BatchNorm2d-135 [-1, 256, 14, 14] 512

Identity-136 [-1, 256, 14, 14] 0

ReLU-137 [-1, 256, 14, 14] 0

Identity-138 [-1, 256, 14, 14] 0

Conv2d-139 [-1, 1024, 14, 14] 262,144

BatchNorm2d-140 [-1, 1024, 14, 14] 2,048

ReLU-141 [-1, 1024, 14, 14] 0

Bottleneck-142 [-1, 1024, 14, 14] 0

Conv2d-143 [-1, 256, 14, 14] 262,144

BatchNorm2d-144 [-1, 256, 14, 14] 512

ReLU-145 [-1, 256, 14, 14] 0

Conv2d-146 [-1, 256, 14, 14] 589,824

BatchNorm2d-147 [-1, 256, 14, 14] 512

Identity-148 [-1, 256, 14, 14] 0

ReLU-149 [-1, 256, 14, 14] 0

Identity-150 [-1, 256, 14, 14] 0

Conv2d-151 [-1, 1024, 14, 14] 262,144

BatchNorm2d-152 [-1, 1024, 14, 14] 2,048

ReLU-153 [-1, 1024, 14, 14] 0

Bottleneck-154 [-1, 1024, 14, 14] 0

Conv2d-155 [-1, 256, 14, 14] 262,144

BatchNorm2d-156 [-1, 256, 14, 14] 512

ReLU-157 [-1, 256, 14, 14] 0

Conv2d-158 [-1, 256, 14, 14] 589,824

BatchNorm2d-159 [-1, 256, 14, 14] 512

Identity-160 [-1, 256, 14, 14] 0

ReLU-161 [-1, 256, 14, 14] 0

Identity-162 [-1, 256, 14, 14] 0

Conv2d-163 [-1, 1024, 14, 14] 262,144

BatchNorm2d-164 [-1, 1024, 14, 14] 2,048

ReLU-165 [-1, 1024, 14, 14] 0

Bottleneck-166 [-1, 1024, 14, 14] 0

Conv2d-167 [-1, 512, 14, 14] 524,288

BatchNorm2d-168 [-1, 512, 14, 14] 1,024

ReLU-169 [-1, 512, 14, 14] 0

Conv2d-170 [-1, 512, 7, 7] 2,359,296

BatchNorm2d-171 [-1, 512, 7, 7] 1,024

Identity-172 [-1, 512, 7, 7] 0

ReLU-173 [-1, 512, 7, 7] 0

Identity-174 [-1, 512, 7, 7] 0

Conv2d-175 [-1, 2048, 7, 7] 1,048,576

BatchNorm2d-176 [-1, 2048, 7, 7] 4,096

Conv2d-177 [-1, 2048, 7, 7] 2,097,152

BatchNorm2d-178 [-1, 2048, 7, 7] 4,096

ReLU-179 [-1, 2048, 7, 7] 0

Bottleneck-180 [-1, 2048, 7, 7] 0

Conv2d-181 [-1, 512, 7, 7] 1,048,576

BatchNorm2d-182 [-1, 512, 7, 7] 1,024

ReLU-183 [-1, 512, 7, 7] 0

Conv2d-184 [-1, 512, 7, 7] 2,359,296

BatchNorm2d-185 [-1, 512, 7, 7] 1,024

Identity-186 [-1, 512, 7, 7] 0

ReLU-187 [-1, 512, 7, 7] 0

Identity-188 [-1, 512, 7, 7] 0

Conv2d-189 [-1, 2048, 7, 7] 1,048,576

BatchNorm2d-190 [-1, 2048, 7, 7] 4,096

ReLU-191 [-1, 2048, 7, 7] 0

Bottleneck-192 [-1, 2048, 7, 7] 0

Conv2d-193 [-1, 512, 7, 7] 1,048,576

BatchNorm2d-194 [-1, 512, 7, 7] 1,024

ReLU-195 [-1, 512, 7, 7] 0

Conv2d-196 [-1, 512, 7, 7] 2,359,296

BatchNorm2d-197 [-1, 512, 7, 7] 1,024

Identity-198 [-1, 512, 7, 7] 0

ReLU-199 [-1, 512, 7, 7] 0

Identity-200 [-1, 512, 7, 7] 0

Conv2d-201 [-1, 2048, 7, 7] 1,048,576

BatchNorm2d-202 [-1, 2048, 7, 7] 4,096

ReLU-203 [-1, 2048, 7, 7] 0

Bottleneck-204 [-1, 2048, 7, 7] 0

AdaptiveAvgPool2d-205 [-1, 2048, 1, 1] 0

Flatten-206 [-1, 2048] 0

SelectAdaptivePool2d-207 [-1, 2048] 0

Linear-208 [-1, 37] 75,813

================================================================

Total params: 23,583,845

Trainable params: 23,583,845

Non-trainable params: 0

----------------------------------------------------------------

Input size (MB): 0.57

Forward/backward pass size (MB): 307.64

Params size (MB): 89.97

Estimated Total Size (MB): 398.18

Second Summary:

----------------------------------------------------------------

Layer (type) Output Shape Param #

================================================================

Conv2d-1 [-1, 64, 112, 112] 9,408

BatchNorm2d-2 [-1, 64, 112, 112] 128

ReLU-3 [-1, 64, 112, 112] 0

MaxPool2d-4 [-1, 64, 56, 56] 0

Conv2d-5 [-1, 64, 56, 56] 4,096

BatchNorm2d-6 [-1, 64, 56, 56] 128

ReLU-7 [-1, 64, 56, 56] 0

Conv2d-8 [-1, 64, 56, 56] 36,864

BatchNorm2d-9 [-1, 64, 56, 56] 128

Identity-10 [-1, 64, 56, 56] 0

ReLU-11 [-1, 64, 56, 56] 0

Identity-12 [-1, 64, 56, 56] 0

Conv2d-13 [-1, 256, 56, 56] 16,384

BatchNorm2d-14 [-1, 256, 56, 56] 512

Conv2d-15 [-1, 256, 56, 56] 16,384

BatchNorm2d-16 [-1, 256, 56, 56] 512

ReLU-17 [-1, 256, 56, 56] 0

Bottleneck-18 [-1, 256, 56, 56] 0

Conv2d-19 [-1, 64, 56, 56] 16,384

BatchNorm2d-20 [-1, 64, 56, 56] 128

ReLU-21 [-1, 64, 56, 56] 0

Conv2d-22 [-1, 64, 56, 56] 36,864

BatchNorm2d-23 [-1, 64, 56, 56] 128

Identity-24 [-1, 64, 56, 56] 0

ReLU-25 [-1, 64, 56, 56] 0

Identity-26 [-1, 64, 56, 56] 0

Conv2d-27 [-1, 256, 56, 56] 16,384

BatchNorm2d-28 [-1, 256, 56, 56] 512

ReLU-29 [-1, 256, 56, 56] 0

Bottleneck-30 [-1, 256, 56, 56] 0

Conv2d-31 [-1, 64, 56, 56] 16,384

BatchNorm2d-32 [-1, 64, 56, 56] 128

ReLU-33 [-1, 64, 56, 56] 0

Conv2d-34 [-1, 64, 56, 56] 36,864

BatchNorm2d-35 [-1, 64, 56, 56] 128

Identity-36 [-1, 64, 56, 56] 0

ReLU-37 [-1, 64, 56, 56] 0

Identity-38 [-1, 64, 56, 56] 0

Conv2d-39 [-1, 256, 56, 56] 16,384

BatchNorm2d-40 [-1, 256, 56, 56] 512

ReLU-41 [-1, 256, 56, 56] 0

Bottleneck-42 [-1, 256, 56, 56] 0

Conv2d-43 [-1, 128, 56, 56] 32,768

BatchNorm2d-44 [-1, 128, 56, 56] 256

ReLU-45 [-1, 128, 56, 56] 0

Conv2d-46 [-1, 128, 28, 28] 147,456

BatchNorm2d-47 [-1, 128, 28, 28] 256

Identity-48 [-1, 128, 28, 28] 0

ReLU-49 [-1, 128, 28, 28] 0

Identity-50 [-1, 128, 28, 28] 0

Conv2d-51 [-1, 512, 28, 28] 65,536

BatchNorm2d-52 [-1, 512, 28, 28] 1,024

Conv2d-53 [-1, 512, 28, 28] 131,072

BatchNorm2d-54 [-1, 512, 28, 28] 1,024

ReLU-55 [-1, 512, 28, 28] 0

Bottleneck-56 [-1, 512, 28, 28] 0

Conv2d-57 [-1, 128, 28, 28] 65,536

BatchNorm2d-58 [-1, 128, 28, 28] 256

ReLU-59 [-1, 128, 28, 28] 0

Conv2d-60 [-1, 128, 28, 28] 147,456

BatchNorm2d-61 [-1, 128, 28, 28] 256

Identity-62 [-1, 128, 28, 28] 0

ReLU-63 [-1, 128, 28, 28] 0

Identity-64 [-1, 128, 28, 28] 0

Conv2d-65 [-1, 512, 28, 28] 65,536

BatchNorm2d-66 [-1, 512, 28, 28] 1,024

ReLU-67 [-1, 512, 28, 28] 0

Bottleneck-68 [-1, 512, 28, 28] 0

Conv2d-69 [-1, 128, 28, 28] 65,536

BatchNorm2d-70 [-1, 128, 28, 28] 256

ReLU-71 [-1, 128, 28, 28] 0

Conv2d-72 [-1, 128, 28, 28] 147,456

BatchNorm2d-73 [-1, 128, 28, 28] 256

Identity-74 [-1, 128, 28, 28] 0

ReLU-75 [-1, 128, 28, 28] 0

Identity-76 [-1, 128, 28, 28] 0

Conv2d-77 [-1, 512, 28, 28] 65,536

BatchNorm2d-78 [-1, 512, 28, 28] 1,024

ReLU-79 [-1, 512, 28, 28] 0

Bottleneck-80 [-1, 512, 28, 28] 0

Conv2d-81 [-1, 128, 28, 28] 65,536

BatchNorm2d-82 [-1, 128, 28, 28] 256

ReLU-83 [-1, 128, 28, 28] 0

Conv2d-84 [-1, 128, 28, 28] 147,456

BatchNorm2d-85 [-1, 128, 28, 28] 256

Identity-86 [-1, 128, 28, 28] 0

ReLU-87 [-1, 128, 28, 28] 0

Identity-88 [-1, 128, 28, 28] 0

Conv2d-89 [-1, 512, 28, 28] 65,536

BatchNorm2d-90 [-1, 512, 28, 28] 1,024

ReLU-91 [-1, 512, 28, 28] 0

Bottleneck-92 [-1, 512, 28, 28] 0

Conv2d-93 [-1, 256, 28, 28] 131,072

BatchNorm2d-94 [-1, 256, 28, 28] 512

ReLU-95 [-1, 256, 28, 28] 0

Conv2d-96 [-1, 256, 14, 14] 589,824

BatchNorm2d-97 [-1, 256, 14, 14] 512

Identity-98 [-1, 256, 14, 14] 0

ReLU-99 [-1, 256, 14, 14] 0

Identity-100 [-1, 256, 14, 14] 0

Conv2d-101 [-1, 1024, 14, 14] 262,144

BatchNorm2d-102 [-1, 1024, 14, 14] 2,048

Conv2d-103 [-1, 1024, 14, 14] 524,288

BatchNorm2d-104 [-1, 1024, 14, 14] 2,048

ReLU-105 [-1, 1024, 14, 14] 0

Bottleneck-106 [-1, 1024, 14, 14] 0

Conv2d-107 [-1, 256, 14, 14] 262,144

BatchNorm2d-108 [-1, 256, 14, 14] 512

ReLU-109 [-1, 256, 14, 14] 0

Conv2d-110 [-1, 256, 14, 14] 589,824

BatchNorm2d-111 [-1, 256, 14, 14] 512

Identity-112 [-1, 256, 14, 14] 0

ReLU-113 [-1, 256, 14, 14] 0

Identity-114 [-1, 256, 14, 14] 0

Conv2d-115 [-1, 1024, 14, 14] 262,144

BatchNorm2d-116 [-1, 1024, 14, 14] 2,048

ReLU-117 [-1, 1024, 14, 14] 0

Bottleneck-118 [-1, 1024, 14, 14] 0

Conv2d-119 [-1, 256, 14, 14] 262,144

BatchNorm2d-120 [-1, 256, 14, 14] 512

ReLU-121 [-1, 256, 14, 14] 0

Conv2d-122 [-1, 256, 14, 14] 589,824

BatchNorm2d-123 [-1, 256, 14, 14] 512

Identity-124 [-1, 256, 14, 14] 0

ReLU-125 [-1, 256, 14, 14] 0

Identity-126 [-1, 256, 14, 14] 0

Conv2d-127 [-1, 1024, 14, 14] 262,144

BatchNorm2d-128 [-1, 1024, 14, 14] 2,048

ReLU-129 [-1, 1024, 14, 14] 0

Bottleneck-130 [-1, 1024, 14, 14] 0

Conv2d-131 [-1, 256, 14, 14] 262,144

BatchNorm2d-132 [-1, 256, 14, 14] 512

ReLU-133 [-1, 256, 14, 14] 0

Conv2d-134 [-1, 256, 14, 14] 589,824

BatchNorm2d-135 [-1, 256, 14, 14] 512

Identity-136 [-1, 256, 14, 14] 0

ReLU-137 [-1, 256, 14, 14] 0

Identity-138 [-1, 256, 14, 14] 0

Conv2d-139 [-1, 1024, 14, 14] 262,144

BatchNorm2d-140 [-1, 1024, 14, 14] 2,048

ReLU-141 [-1, 1024, 14, 14] 0

Bottleneck-142 [-1, 1024, 14, 14] 0

Conv2d-143 [-1, 256, 14, 14] 262,144

BatchNorm2d-144 [-1, 256, 14, 14] 512

ReLU-145 [-1, 256, 14, 14] 0

Conv2d-146 [-1, 256, 14, 14] 589,824

BatchNorm2d-147 [-1, 256, 14, 14] 512

Identity-148 [-1, 256, 14, 14] 0

ReLU-149 [-1, 256, 14, 14] 0

Identity-150 [-1, 256, 14, 14] 0

Conv2d-151 [-1, 1024, 14, 14] 262,144

BatchNorm2d-152 [-1, 1024, 14, 14] 2,048

ReLU-153 [-1, 1024, 14, 14] 0

Bottleneck-154 [-1, 1024, 14, 14] 0

Conv2d-155 [-1, 256, 14, 14] 262,144

BatchNorm2d-156 [-1, 256, 14, 14] 512

ReLU-157 [-1, 256, 14, 14] 0

Conv2d-158 [-1, 256, 14, 14] 589,824

BatchNorm2d-159 [-1, 256, 14, 14] 512

Identity-160 [-1, 256, 14, 14] 0

ReLU-161 [-1, 256, 14, 14] 0

Identity-162 [-1, 256, 14, 14] 0

Conv2d-163 [-1, 1024, 14, 14] 262,144

BatchNorm2d-164 [-1, 1024, 14, 14] 2,048

ReLU-165 [-1, 1024, 14, 14] 0

Bottleneck-166 [-1, 1024, 14, 14] 0

Conv2d-167 [-1, 512, 14, 14] 524,288

BatchNorm2d-168 [-1, 512, 14, 14] 1,024

ReLU-169 [-1, 512, 14, 14] 0

Conv2d-170 [-1, 512, 7, 7] 2,359,296

BatchNorm2d-171 [-1, 512, 7, 7] 1,024

Identity-172 [-1, 512, 7, 7] 0

ReLU-173 [-1, 512, 7, 7] 0

Identity-174 [-1, 512, 7, 7] 0

Conv2d-175 [-1, 2048, 7, 7] 1,048,576

BatchNorm2d-176 [-1, 2048, 7, 7] 4,096

Conv2d-177 [-1, 2048, 7, 7] 2,097,152

BatchNorm2d-178 [-1, 2048, 7, 7] 4,096

ReLU-179 [-1, 2048, 7, 7] 0

Bottleneck-180 [-1, 2048, 7, 7] 0

Conv2d-181 [-1, 512, 7, 7] 1,048,576

BatchNorm2d-182 [-1, 512, 7, 7] 1,024

ReLU-183 [-1, 512, 7, 7] 0

Conv2d-184 [-1, 512, 7, 7] 2,359,296

BatchNorm2d-185 [-1, 512, 7, 7] 1,024

Identity-186 [-1, 512, 7, 7] 0

ReLU-187 [-1, 512, 7, 7] 0

Identity-188 [-1, 512, 7, 7] 0

Conv2d-189 [-1, 2048, 7, 7] 1,048,576

BatchNorm2d-190 [-1, 2048, 7, 7] 4,096

ReLU-191 [-1, 2048, 7, 7] 0

Bottleneck-192 [-1, 2048, 7, 7] 0

Conv2d-193 [-1, 512, 7, 7] 1,048,576

BatchNorm2d-194 [-1, 512, 7, 7] 1,024

ReLU-195 [-1, 512, 7, 7] 0

Conv2d-196 [-1, 512, 7, 7] 2,359,296

BatchNorm2d-197 [-1, 512, 7, 7] 1,024

Identity-198 [-1, 512, 7, 7] 0

ReLU-199 [-1, 512, 7, 7] 0

Identity-200 [-1, 512, 7, 7] 0

Conv2d-201 [-1, 2048, 7, 7] 1,048,576

BatchNorm2d-202 [-1, 2048, 7, 7] 4,096

ReLU-203 [-1, 2048, 7, 7] 0

Bottleneck-204 [-1, 2048, 7, 7] 0

AdaptiveAvgPool2d-205 [-1, 2048, 1, 1] 0

Flatten-206 [-1, 2048] 0

SelectAdaptivePool2d-207 [-1, 2048] 0

================================================================

Total params: 23,508,032

Trainable params: 23,508,032

Non-trainable params: 0

----------------------------------------------------------------

Input size (MB): 0.57

Forward/backward pass size (MB): 307.64

Params size (MB): 89.68

Estimated Total Size (MB): 397.89

----------------------------------------------------------------

www.marearts.com

🙇🏻‍♂️

Output exceeds the size limit. Open the full output data in a text editor

May 20, 2023

Jupyter notebook

unlimited size of line:

Code #1

from IPython.core.display import display, HTML
display(HTML("<style>.container { width:100% !important; }</style>"))

Code #2

from IPython.core.interactiveshell import InteractiveShell

InteractiveShell.ast_node_interactivity = "all"

Thank you.

🙇🏻‍♂️

swin transformer v2 - model forward and export onnx

May 18, 2023

1. load pre-trained model

2. export onnx

3. load onnx

refer to code:

import warnings
from torch.jit import TracerWarning
warnings.filterwarnings("ignore", category=TracerWarning)

#------------------
#swin-transformer v2 pretrained model
#------------------

from transformers import AutoImageProcessor, Swinv2Model
import torch
from datasets import load_dataset

dataset = load_dataset("huggingface/cats-image")
image = dataset["test"]["image"][0]

image_processor = AutoImageProcessor.from_pretrained("microsoft/swinv2-tiny-patch4-window8-256")
model = Swinv2Model.from_pretrained("microsoft/swinv2-tiny-patch4-window8-256")

inputs = image_processor(image, return_tensors="pt")

with torch.no_grad():
    outputs = model(**inputs)

last_hidden_states = outputs.last_hidden_state

# print( list(last_hidden_states.shape) )
# Convert last_hidden_states to numpy
last_hidden_states_numpy = last_hidden_states.detach().numpy()
print(f"Shape of last_hidden_states: {last_hidden_states_numpy.shape}")
print(last_hidden_states)



#----------------
#onnx export
#------------------
import torch
from torch.autograd import Variable

# ensure the model is in evaluation mode
model.eval()

# create a dummy variable with the same size as your input
# for this example, let's assume the input is of size [1, 3, 256, 256]
dummy_input = Variable(torch.randn(1, 3, 256, 256))

# specify the file path
file_path = "./swinv2_tiny.onnx"

# export the model
torch.onnx.export(model, dummy_input, file_path)

#------------------
#onnx inference
#------------------
import onnxruntime as ort

# load the ONNX model
ort_session = ort.InferenceSession(file_path)

# convert the PyTorch tensor to numpy array for onnxruntime
print(inputs.keys())
inputs_numpy = inputs["pixel_values"].numpy()
# inputs_numpy = inputs["input_ids"].numpy()

# create a dictionary from model input name to the actual input data
ort_inputs = {ort_session.get_inputs()[0].name: inputs_numpy}

# forward
ort_outs = ort_session.run(None, ort_inputs)
print(f"Shape of ort_outs: {ort_outs[0].shape}")
print(ort_outs)
# print(type(ort_outs))
# print( list(ort_outs.shape) )

Thank you.

www.marearts.com

🙇🏻‍♂️

Generate EAN13 barcode number as random

May 04, 2023

refer to code:

def generate_random_ean13():
    number = [random.randint(0, 9) for _ in range(12)]
    number_str = ''.join(str(digit) for digit in number)
    ean = EAN13(number_str)
    return ean.get_fullcode()

Thank you.

study.marearts.com

Yolo V7 vs V8

May 02, 2023

V7 vs V8 comparison

https://youtu.be/k1dOZFcLOek

https://youtu.be/tpOGDclq7KY

https://youtu.be/u5qxN2ACEP4

https://youtu.be/85SH08jN4dY

This is a comparison video between yolo v7 and v8.

Here is information for each version

Yolo V7
- Github : https://github.com/WongKinYiu/yolov7
- Model : yolov7x.pt
Yolo V8
- Github : https://github.com/ultralytics/ultralytics
- Model : yolov8x.pt

Testing Computer :

Intel(R) Core(TM) i7-9800X CPU @ 3.80GHz
RTX 4090

Something might be useful code

yolo v8, video writer for detection result

import cv2
import time
from ultralytics import YOLO

def process_video(model, video_path, output_path):
    cap = cv2.VideoCapture(video_path)
    width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
    height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
    fps = int(cap.get(cv2.CAP_PROP_FPS))

    # Create a VideoWriter object to save the annotated video
    fourcc = cv2.VideoWriter_fourcc(*'mp4v')
    out = cv2.VideoWriter(output_path, fourcc, fps, (width, height))

    while cap.isOpened():
        success, frame = cap.read()

        if success:
            start_time = time.time()
            results = model(frame)
            end_time = time.time()
            processing_time = end_time - start_time
            fps = 1/processing_time
            # Visualize the results on the frame
            annotated_frame = results[0].plot()
            
            # Display the processing time on the annotated frame
            cv2.putText(annotated_frame, f"Processing time: {processing_time:.4f} seconds / {fps:.4f} fps",
                        (10, 30), cv2.FONT_HERSHEY_SIMPLEX, 1, (255, 255, 255), 2)

            # Write the annotated frame to the output video
            out.write(annotated_frame)

            # cv2.imshow("YOLOv8 Inference", annotated_frame)
            # if cv2.waitKey(1) & 0xFF == ord("q"):
            #     break
        else:
            break

    cap.release()
    out.release()

def main():
    # Load the YOLO model
    model = YOLO('yolov8x.pt')

    # List of video files
    video_paths = [
        "../video/videoplayback-1.mp4",
        "../video/videoplayback-2.mp4",
        "../video/videoplayback-3.mp4",
        "../video/videoplayback-4.mp4",
    ]

    # Loop through video files and process them
    for i, video_path in enumerate(video_paths):
        output_path = f"../video/yolo_88_output_{i+1}.mp4"
        process_video(model, video_path, output_path)

    cv2.destroyAllWindows()

if __name__ == '__main__':
    main()

make 2 video to side by side

Combine Two Videos Side by Side with OpenCV python

Thank you! 😺