Coding mein galtiyan (Bugs) hona "Gunah" nahi hai, unhein handle na karna gunah hai. AI models train karte waqt aksar errors aate hain — kabhi data missing hota hai, toh kabhi GPU memory khatam ho jati hai. Agar aapne Error Handling nahi seekhi, toh aapka program beech mein hi "Crash" ho jayega. Is guide mein hum Python ki "Safety Net" (Try-Except) ko samjhenge.
1. Try-Except: The Safety Net
- Try: Wo code jo khatarnak hai (jaise file open karna ya web scraping).
- Except: Agar galti hui, toh program band hone ke bajaye yahan aakar koi "Message" dega.
- Example: AI model load karte waqt agar file nahi mili, toh model crash hone ke bajaye error message dekar aage badh sakta hai.
2. Finally aur Else: The Cleanup Crew
- Else: Ye tab chalta hai jab "Try" block mein koi error nahi aaya.
- Finally: Ye hamesha chalta hai, chahe error aaye ya na aaye. Ise hum Cleanup (jaise Database ya File band karna) ke liye use karte hain.
Professional AI developers hamesha
finallyuse karte hain taaki GPU memory "Leak" na ho.
3. Raising Exceptions (Danda dikhana)
Kayi baar hum khud error generate karna chahte hain.
- Raise: Agar user ne "Negative Age" daali, toh hum
raise ValueError("Age cannot be negative")kar sakte hain. - AI mein hum ise Data Validation ke liye use karte hain taaki ganda data model mein na jaye.
4. Logging vs Printing
Beginners print() use karte hain galtiyan dekhne ke liye, professionals Logging use karte hain.
- Logging se aap errors ko ek "File" mein store kar sakte hain taaki aap baad mein dekh sakein ki 10 ghante ki training mein kab galti hui thi.
- 2026 mein AI projects mein hamesha
logginglibrary mandatory hoti hai.
5. Summary Table: Exception Hierarchy for AI
| Exception | When it occurs? | AI Real-World Example |
|---|---|---|
| FileNotFoundError | Missing file | Model .pkl file not found |
| ValueError | Wrong data format | Passing string where float needed |
| ZeroDivisionError | Dividing by zero | When loss becomes zero |
| KeyError | Missing key in dict | Column missing in CSV |
FAQs
1. "Catching All Errors" (except Exception:) kyon bura hai?
Kyonki ye asli galti ko chhipa deta hai. Aapko hamesha "Specific" hona chahiye (e.g. except FileNotFoundError). Ise "Pokemon Exception Handling" (Gotta catch 'em all) kehte hain aur ye ek "Anti-pattern" hai.
2. Custom Exception kaise banayein?
Bas ek class banaiye jo Exception se inherit kare: class DataShapeError(Exception): pass. AI mein ye tensors ki shape mismatch handle karne ke liye bahut kaam aata hai.
3. "Assertion Error" kya hai?
Ise hum model ki requirements check karne ke liye use karte hain: assert data.shape[0] > 0, "Empty Data!". Agar condition fail hui, toh code wahi ruk jayega.
4. 2026 mein AI-assisted Debugging? Ab Python errors ke saath AI suggestions aate hain jo batate hain ki "Shayad aapne column ka naam galat likha hai". Par unhe samajhne ke liye basic logic strong hona zaroori hai.
Error handling aapke code ko "Unbreakable" banati hai. Ise seekh kar aap real-world professional projects banane ke liye taiyar ho jayenge! 🛡️
Tarun ke baare mein: Tarun resilient systems aur automatic error-recovery patterns ke specialist hain. AI-Gyani par har line safety-first hai.