Question 1

How to fix matplotlib "TypeError: Axes.plot() got an unexpected keyword argument 's'" when plotting positive and negative examples?

Accepted Answer

`plot()` does not support the scatter-size parameter `s`. Replace `ax.plot()` with `ax.scatter()` and specify colors with `c`:

```python
ax1.scatter(pos_examples[:, 0], pos_examples[:, 1],
            c='blue', marker='o', label='Positive', s=100, alpha=0.6)
ax1.scatter(neg_examples[:, 0], neg_examples[:, 1],
            c='red', marker='o', label='Negative', s=100, alpha=0.6)
```

Do not use the `'bo'`/`'ro'` format strings with `scatter()`; use `c='blue'` and `c='red'` instead.

Question 2

Why is the Tamil translation showing Hindi words, and how do I fix it?

Accepted Answer

This is caused by residual machine-translated Hindi (Devanagari) text committed under the Tamil locale. Audit the Tamil files with `rg '[\x{0900}-\x{097F}]' translations/ta/`, replace every Hindi span with the correct Tamil equivalent based on the English source, and restore `translations/ta/README.md` if it contains Hindi sections. The fix is covered by the PRs "fix(translations): replace residual Hindi words in Tamil lessons" and "fix: restore Tamil README translation" (`Fixes #734`). As a temporary workaround, use the English version or manually remove the Devanagari text from the affected Tamil Markdown files.

Question 3

How do I fix bare except clauses that catch SystemExit and KeyboardInterrupt in Python image validation?

Accepted Answer

Replace every bare `except:` with `except Exception:` so only normal application exceptions are caught while system-level exceptions propagate. Update these four files:

- `lessons/4-ComputerVision/07-ConvNets/tfcv.py` (line 51)
- `lessons/4-ComputerVision/07-ConvNets/pytorchcv.py` (line 137)
- `lessons/4-ComputerVision/08-TransferLearning/tfcv.py` (line 51)
- `lessons/4-ComputerVision/08-TransferLearning/pytorchcv.py` (line 137)

Example fix:

```python
def check_image(fn):
    try:
        im = Image.open(fn)
        im.verify()
        return True
    except Exception:
        return False
```

This follows PEP 8 and prevents `KeyboardInterrupt`, `SystemExit`, and other critical exceptions from being silently masked.

Question 4

How do you fix the incorrect delta p formula in the OwnFramework backpropagation lesson?

Accepted Answer

The formula should use the chain rule with partial derivatives, not a finite change in the loss. If `p` is a parameter, `z` the layer output, and `L` the loss, use:

```text
∂L/∂p = ∂L/∂z · ∂z/∂p
```

Here, `∂L/∂z` is the upstream gradient from the next layer and `∂z/∂p` is the local gradient. Do not multiply by `ΔL`, the actual change in loss, or treat a finite-loss change as a gradient.

For a linear layer `z = Wx + b`, the gradients are:

```text
∂L/∂W = ∂L/∂z · xᵀ
∂L/∂b = ∂L/∂z
∂L/∂x = Wᵀ · ∂L/∂z
```

Keep the optimizer update separate:

```text
p_new = p - learning_rate · ∂L/∂p
```

Pull the latest version of `OwnFramework.ipynb` or replace the erroneous `δp` equation with the chain-rule expression above. The corrected notebook also fixes the `obtanes` typo and uses consistent `p` and `z` partial-derivative notation.

Question 5

How to fix the Tamil translation showing Hindi text in the README?

Accepted Answer

The issue was caused by the Tamil (`ta`) README containing Hindi/Devanagari content. It was fixed by restoring the Tamil-language README at `translations/ta/README.md`. After updating, verify the file contains Tamil script and no Devanagari characters; the fix reported 102 Tamil-script lines and 0 Devanagari lines. Pull the latest changes or apply PR `fix(ta): restore Tamil translation in README`. Report remaining translation issues at https://github.com/Azure/co-op-translator.

AI-For-Beginners

Overview

README Preview

FAQ (5)

同类型项目

freeCodeCamp

generative-ai-for-beginners

cs-self-learning

ai-agents-for-beginners