Mohammad Farid Hendianto

Mohammad Farid Hendianto

ID
@ndik
Gaming
7.4K
Video Count
1.2M
Video View
1.7K
Subscriber
#8,399
Indonesia Rank
#275,336
Global Rank
Mohammad Farid Hendianto YouTube channel subscribers:1,740- Seelive statisticsand growth insights below.

Mohammad Farid Hendianto YouTube Statistics & Analytics

Subscribers
1.7K
Total Views
1.2M
Videos
7.4K
Activity
Unknown

Mohammad Farid Hendianto Content Analysis

Content Type Distribution

Long videosLong
71%
147 videos
ShortsShorts
29%
59 videos

📽️ This channel specializes in long-form videos. Deep dives and comprehensive content perform well here.

Content Categories

Primary CategoryNews & Politics
44%
News & Politics
90(44%)
Gaming
75(36%)
Education
34(17%)
Howto & Style
7(3%)

🎯 Primary focus: News & Politics with 90 videos (44% of categorized content).

Latest Video

Long video
Membangun AI Pengoreksi Grammar  - Presentasi Responsi Deep Learning || Universitas Ahmad Dahlan
10:00
New

Membangun AI Pengoreksi Grammar - Presentasi Responsi Deep Learning || Universitas Ahmad Dahlan

2
Views
0
Likes
2 days ago
Published

Bagaimana cara melatih model AI Transformer dengan dataset yang sangat kecil tapi hasilnya bisa menyerupai State-of-the-Art (SOTA)? 🚀 Di video presentasi Responsi Praktikum Deep Learning ini, saya mendemokan perjalanan membangun model AI pengoreksi tata bahasa Inggris (Grammar Error Correction) dari nol. Kita akan membahas 11 eksperimen arsitektur yang berhasil meningkatkan akurasi model hingga 3 kali lipat (dari 23% menjadi 72.88%) menggunakan dataset JFLEG yang sangat terbatas (~1.200 kalimat). 💡 COBA MODEL & BACA KODE LENGKAPNYA DI SINI: 🔗 Kaggle Notebook (Source Code): https://www.kaggle.com/code/ireddragonicy/transformer-evolution-for-grammar-error-correction 🔗 Hugging Face (Pre-trained Models): https://huggingface.co/IRedDragonICY/jfleg-gec-transformer-ensemble/tree/main Di video ini, kita akan membedah: 1️⃣ Masalah Halusinasi Model: Kenapa Transformer standar gagal total di data kecil. 2️⃣ Modifikasi Arsitektur (V2): Efek ajaib dari Pre-Layer Normalization, GELU, dan Weight Tying. 3️⃣ Mekanisme Pointer-Generator: Cara mengajari AI untuk "copy-paste" kata dari input untuk mengatasi kata langka (Out-of-Vocabulary). 4️⃣ Dynamic Mined Noise: Rahasia menyulap 1.200 kalimat menjadi puluhan ribu data latihan baru yang berubah-ubah setiap epoch. 5️⃣ Weighted Ensemble: Menggabungkan probabilitas beberapa model untuk hasil akhir dengan tata bahasa yang flawless. ⏱️ TIMESTAMPS (Bab Video): 0:00 - Intro & Tantangan Dataset JFLEG (Low-Resource) 1:15 - Demo Model Baseline (Masalah Halusinasi Parah) 3:30 - Modifikasi Arsitektur V2: Pre-LN, GELU & Weight Tying 5:00 - Mengatasi OOV dengan Mekanisme Pointer-Generator 6:45 - Rahasia Augmentasi Data: Dynamic Mined Noise 8:10 - Hasil Akhir Weighted Ensemble (Akurasi Meroket 72.88%!) 9:15 - Kesimpulan, Tips Riset AI, & Umpan Balik Praktikum 🛠️ Tech Stack & Libraries: Python, TensorFlow, Keras, Hugging Face Datasets, NLTK (GLEU Score). Komputasi didukung oleh Modal.com (NVIDIA H100 GPU). 👨‍💻 Dibuat oleh: Mohammad Farid Hendianto (2200018401) Proyek Akhir / Responsi Mata Kuliah Praktikum Deep Learning Jika kalian merasa video ini bermanfaat untuk belajar Natural Language Processing (NLP) dan Deep Learning, jangan lupa untuk LIKE, COMMENT, dan SUBSCRIBE! Jika ada pertanyaan seputar arsitektur kodenya, silakan tulis di kolom komentar. #DeepLearning #TransformerModel #NaturalLanguageProcessing #MachineLearningIndonesia #Kaggle #HuggingFace #TensorFlow #ArtificialIntelligence #NLP #ProgrammingIndonesia

deep learning nlp natural language processing

See Top YouTube Channels in Indonesia

Jump to the full Indonesia ranking and compare this channel with the country leaders.

Ranking: Indonesia
Open ranking

Mohammad Farid Hendianto Channel Snapshot

Score: 2.5/10

A high-level snapshot of content cadence, library size, and consistency derived from this channel's recent uploads.

Overall Score
2.5
Consistency
95%
Cadence
2-3/wk
Library
50

Growth Potential

3.8/10

Library of 50 videos with ~218 avg views per upload. Combined size + reach signal suggests early-stage development.

Audience Engagement

3.8/10

Avg engagement rate of 2.29% (likes + comments / views) across 33 videos. Below the ~3% industry baseline; community-building plays could lift this.

Niche Specialization

0/10

22% of recent videos cluster in Video game culture. Generalist mix — niche consolidation often unlocks growth at this stage.

Suggested Actions

Recommendations grouped by typical impact for channels at this stage

  1. 1
    Increase upload frequency to 2-3 videos per week
    High ImpactCadence
  2. 2
    Focus on SEO optimization for better discoverability
    High ImpactSEO
  3. 3
    Analyze top-performing content for pattern replication
    MediumStrategy
  4. 4
    Increase community engagement through comments and polls
    MediumEngagement

Frequently Asked Questions About Mohammad Farid Hendianto

Data Source & Accuracy

Source: YouTube Data API v3
Accuracy: Real-time statistics from official YouTube API
Data is updated hourly and sourced directly from official APIs to ensure accuracy and reliability.

Data from YouTube Data API v3 • Updated hourly • Last updated: 09:38 PM