How to File a Good Bug#

Found an issue? We’d love to hear about it! Filing a good bug report helps us fix problems faster and makes visualkeras better for everyone.

Quick Checklist#

Before filing a bug, check off these items:

✅ You’re using the latest version: pip install --upgrade visualkeras
✅ The issue reproduces consistently
✅ You’ve searched existing issues
✅ The issue is specifically with visualkeras (not TensorFlow/Keras itself)

Where to Report#

File bugs on GitHub: paulgavrikov/visualkeras/issues

Essential Information#

Every good bug report includes these 5 things:

What Version of visualkeras?

import visualkeras
print(visualkeras.__version__)

Or check via pip:

pip show visualkeras

What Python Version?
```
python --version
```
Environment Details

Include: - Operating System (Windows/macOS/Linux) - TensorFlow/Keras version: pip show tensorflow - Keras standalone or tf.keras? - Virtual environment (conda/venv/Poetry)?

Minimal Reproducible Example

Provide the shortest possible code that shows the problem:

import tensorflow as tf
from tensorflow import keras
import visualkeras

# Create minimal model that shows the issue
model = keras.Sequential([
    keras.layers.Input(shape=(28, 28, 1)),
    keras.layers.Conv2D(32, (3, 3)),
    keras.layers.Dense(10),
])

# This causes the bug:
image = visualkeras.show(model, mode='layered')
# Error: ... (paste full error here)

Do: - Use the smallest model that reproduces it - Use standard layers (Conv2D, Dense, etc.) - Include the exact error message/traceback - Copy-paste code directly (no screenshots of code)

Don’t: - Share entire model files - Use obscure custom layers if not necessary - Describe the problem in words instead of showing code

What Happened vs. What You Expected

Actual behavior: ` When I call visualkeras.show(model, mode='functional'), I get a TypeError: 'NoneType' object is not subscriptable at line 245 of functional.py `

Expected behavior: ` The function should return a PIL Image showing the model visualization. `

Screenshots (optional but helpful): - Screenshot of error message - Screenshot of expected vs. actual output

Example Bug Report#

Title: show()` with mode='functional' crashes on models with skip connections

Description:

When visualizing a model with skip connections using mode='functional', visualkeras crashes with a TypeError.

Version Info: - visualkeras: 2.4.1 - Python: 3.11.4 - TensorFlow: 2.17.0 (tf.keras) - OS: Ubuntu 22.04

Minimal Reproducible Example:

import tensorflow as tf
from tensorflow import keras
import visualkeras

# Simple model with skip connection
inputs = keras.Input(shape=(224, 224, 3))
x = keras.layers.Conv2D(64, (3, 3))(inputs)
x = keras.layers.Conv2D(64, (3, 3))(x)

# Skip connection
residual = inputs
x = keras.layers.Add()([x, residual])  # ← Fails here

outputs = keras.layers.Dense(10)(x)
model = keras.Model(inputs=inputs, outputs=outputs)

# This crashes:
img = visualkeras.show(model, mode='functional')

Error Message:

Traceback (most recent call last):
  File "reproduce_bug.py", line 20, in <module>
    img = visualkeras.show(model, mode='functional')
  File ".../visualkeras/show.py", line 115, in show
    return functional_view(model, preset=preset, options=options, **overrides)
  File ".../visualkeras/functional.py", line 245, in functional_view
    dims = shape_dict[layer.name]
TypeError: 'NoneType' object is not subscriptable

Expected Behavior:

Should return a PIL Image with the functional visualization showing the skip connection.

Actual Behavior:

Crashes with TypeError. The issue doesn’t occur with mode='layered' or mode='graph'.

—

Common Issues and Solutions#

Before Filing, Try These First#

“ModuleNotFoundError: No module named ‘aggdraw’”

# Install missing dependencies
pip install aggdraw pillow numpy

“visualkeras.show() opens no image window”

You need an image viewer configured:

# Ubuntu/Debian
sudo apt-get install imagemagick display

# macOS
brew install imagemagick

# Or just save to file instead
image = visualkeras.show(model)
image.save('model.png')

Model visualization looks wrong or empty

Check these:

# Make sure model has Input layer
model = keras.Sequential([
    keras.layers.Input(shape=(28, 28, 1)),  # ← Required!
    keras.layers.Dense(10),
])

# Build model if needed
model.build((None, 28, 28, 1))

# Check model summary
model.summary()

“This works on the latest GitHub version but not PyPI”

Install from main branch:

pip install git+https://github.com/paulgavrikov/visualkeras

Color/styling parameters not working

Make sure you’re passing correct format:

# ✅ CORRECT - dict with fill and outline
visualkeras.show(
    model,
    color_map={
        keras.layers.Conv2D: {'fill': '#1976d2', 'outline': '#0d47a1'}
    }
)

# ❌ WRONG - plain strings
visualkeras.show(
    model,
    color_map={keras.layers.Conv2D: 'blue'}  # This won't work!
)