TensorFlow Quest

Tracks

1. 01🌍Why TensorFlow and Where It Fits

   0/7 lessons

   The Origin Story, the Ecosystem, and the Decision Tree

   Where TensorFlow came from, what it competes with, and the production-shaped niche it still owns in 2026.

   Lesson list (7)

   • 01
     From DistBelief to TensorFlow 2.21~12 min · history, tf2, eager
   • 02
     TF, PyTorch, JAX — Where Each One Lives in 2026~12 min · frameworks, comparison, ecosystem
   • 03
     What TensorFlow Is Actually Good At~11 min · deployment, production, tflite, tf-serving
   • 04
     The Eager / Graph / @tf.function Mental Model~14 min · eager, tf-function, graph, tracing
   • 05
     The Ecosystem Map — What Else Lives in TF~10 min · ecosystem, keras-cv, keras-hub, tfx
   • 06
     When to Choose TF vs Alternatives — A Decision Tree~10 min · decision-making, tradeoffs
   • 07
     Installing TF 2.21 — CPU, NVIDIA GPU, Apple Silicon, Docker~12 min · install, gpu, cuda, metal

   ✓Quiz · 5 questions→

2. 02🧊Tensors — TensorFlow's Fundamental Data Structure

   0/7 lessons

   Shapes, dtypes, devices, and the bugs you'll meet most often

   Every input, every weight, every gradient is a tensor. This track makes shape and dtype bugs go away by making them obvious.

   Lesson list (7)

   • 01
     What Is a Tensor?~11 min · tensor, rank, shape, dtype
   • 02
     Creating Tensors: constant, zeros, ones, random~12 min · create, init, random, constant
   • 03
     Tensor Properties: Shape, Dtype, Rank, Device~10 min · shape, dtype, device, introspection
   • 04
     Eager Execution and NumPy Interop~10 min · eager, numpy, interop
   • 05
     Immutable Tensors vs Mutable Variables~11 min · variable, constant, trainable, gradient
   • 06
     Indexing, Slicing, Reshaping~12 min · indexing, slicing, reshape, expand-dims
   • 07
     Common Shape Errors and How to Debug Them~13 min · debugging, shape-error, broadcasting

   ✓Quiz · 5 questions→

3. 03∇Operations, Math & Automatic Differentiation

   0/6 lessons

   Element-wise math, matmul, broadcasting, GradientTape — the engine of training

   How TF computes gradients automatically and how the entire training loop is built on top of those four steps.

   Lesson list (6)

   • 01
     Element-wise Math and Reductions~11 min · ops, reduce, argmax
   • 02
     Matrix Multiplication and Broadcasting~12 min · matmul, broadcasting, linear-algebra
   • 03
     GradientTape Basics — Your First Gradient~13 min · autodiff, gradient-tape, backprop
   • 04
     Persistent Tapes and Higher-Order Gradients~11 min · gradient-tape, persistent, second-derivative
   • 05
     Linear Regression from Scratch — The Universal Loop~14 min · training-loop, gradient-descent, regression
   • 06
     Gradient Clipping and stop_gradient~11 min · clipping, stop-gradient, exploding-gradient, rl

   ✓Quiz · 5 questions→

4. 04🏗️Building Models with tf.keras

   0/6 lessons

   Sequential, Functional, Subclassing — all three in your toolkit

   Three Keras model APIs, when each is right, and how to write your own custom layer when none of them is.

   Lesson list (6)

   • 01
     Why Keras Is TF's High-Level API~9 min · keras, history, keras3
   • 02
     Sequential API — The Linear Stack~11 min · sequential, dense, dropout
   • 03
     Functional API — DAGs and Skip Connections~14 min · functional, skip-connection, resnet, multi-io
   • 04
     Model Subclassing — Maximum Flexibility~13 min · subclassing, research, training-flag
   • 05
     Essential Layers and Activation Functions~11 min · layers, activations, relu, gelu
   • 06
     Custom Layers — When Built-Ins Don't Cover It~13 min · custom-layer, build, add-weight

   ✓Quiz · 5 questions→

5. 05🔄The Training Loop

   0/6 lessons

   From compile() to fit() to GradientTape — and why callbacks are non-negotiable

   Optimizers, losses, metrics, callbacks, and how to write a custom training loop when fit() isn't flexible enough.

   Lesson list (6)

   • 01
     model.compile() — Wiring Optimizer, Loss, Metrics~9 min · compile, optimizer, loss
   • 02
     Optimizers — Adam, SGD, AdamW, Schedules~13 min · optimizer, adam, sgd, schedule
   • 03
     Loss Functions and Metrics — The Right Choice~11 min · loss, metric, classification, regression
   • 04
     model.fit() — The Training Loop You Don't Write~12 min · fit, history, validation
   • 05
     Callbacks — EarlyStopping, ModelCheckpoint, ReduceLROnPlateau~11 min · callbacks, early-stopping, checkpoint
   • 06
     Custom Training Loops with GradientTape~13 min · custom-loop, gradient-tape, metrics

   ✓Quiz · 5 questions→

6. 06📊Data Pipelines with tf.data

   0/6 lessons

   Lazy, parallel, GPU-friendly — and the four antipatterns to never write

   How to feed your model fast enough that the GPU never sits idle. Order, AUTOTUNE, caching, prefetching.

   Lesson list (6)

   • 01
     Why tf.data Exists~8 min · tf-data, pipeline, performance
   • 02
     Dataset Basics — Creating and Iterating~11 min · dataset, from-tensor-slices, iteration
   • 03
     Chaining Transformations — map, batch, shuffle, filter~13 min · map, batch, shuffle, filter
   • 04
     AUTOTUNE, Caching, Prefetching~11 min · autotune, cache, prefetch, performance
   • 05
     TFRecord and Interleave — Large-Scale Pipelines~12 min · tfrecord, interleave, large-scale
   • 06
     Practical Image Loading + Antipatterns~13 min · image-loading, antipatterns, debugging

   ✓Quiz · 4 questions→

7. 07👁️CNNs in TensorFlow

   0/6 lessons

   Convolutions, pooling, augmentation, transfer learning, Grad-CAM

   How to build, fine-tune, and interpret convolutional networks for vision tasks.

   Lesson list (6)

   • 01
     Conv2D Explained — Filters, Strides, Padding~13 min · conv2d, filter, stride, padding
   • 02
     Pooling Layers and Modern Alternatives~10 min · pooling, max-pool, global-avg-pool
   • 03
     Data Augmentation Layers~11 min · augmentation, regularization
   • 04
     Transfer Learning — Feature Extraction vs Fine-Tuning~14 min · transfer-learning, fine-tuning, pretrained
   • 05
     Common CNN Architectures~10 min · resnet, efficientnet, mobilenet, convnext
   • 06
     Grad-CAM — Seeing What CNNs See~11 min · grad-cam, interpretability, visualization

   ✓Quiz · 4 questions→

8. 08📝Sequences, RNNs & Attention

   0/6 lessons

   From SimpleRNN to LSTM to Multi-Head Attention — and why we got here

   Order-dependent data: text, audio, time series. Why each new architecture solved the previous one's pain.

   Lesson list (6)

   • 01
     Why Sequences Need Special Treatment~10 min · sequences, rnn, ordering
   • 02
     SimpleRNN, LSTM, and GRU~14 min · rnn, lstm, gru
   • 03
     TextVectorization and Embeddings~12 min · text, tokenization, embedding
   • 04
     Attention and MultiHeadAttention~14 min · attention, transformer, self-attention
   • 05
     Why Transformers Replaced RNNs~11 min · transformer, rnn, scaling
   • 06
     Sequence-to-Sequence with Cross Attention~13 min · seq2seq, encoder-decoder, translation

   ✓Quiz · 5 questions→

9. 09💾Saving, Loading & Model Formats

   0/5 lessons

   Three formats, when to use each, and how to verify before deployment

   From quick checkpoints to production-ready SavedModels — and the verification step that prevents accidents.

   Lesson list (5)

   • 01
     model.save() and load_model() — Three Formats~10 min · save, load, keras-format
   • 02
     SavedModel Format In Depth~11 min · savedmodel, deployment, signatures
   • 03
     Checkpoints — Resume Interrupted Training~12 min · checkpoint, training, resume
   • 04
     ONNX Export and TF Hub~11 min · onnx, tf-hub, interop
   • 05
     Practical: Save, Load, and Verify~12 min · verification, deployment-prep

   ✓Quiz · 4 questions→

10. 10📱TFLite & Edge Deployment

    0/5 lessons

    From a SavedModel to a 3.5 MB int8 binary that runs on a phone

    How to shrink models 4-10× for mobile and microcontroller deployment without losing meaningful accuracy.

    Lesson list (5)

    • 01
      What Is TFLite — and Why It's Becoming LiteRT~9 min · tflite, litert, edge, mobile
    • 02
      TFLiteConverter and Post-Training Quantization~13 min · converter, quantization, int8
    • 03
      Quantization-Aware Training and Pruning~12 min · qat, pruning, model-optimization
    • 04
      TFLite Interpreter and Hardware Delegates~11 min · interpreter, delegates, edge-tpu, core-ml
    • 05
      Practical: Full TFLite Conversion Pipeline~13 min · practical, mobilenet, deployment

    ✓Quiz · 4 questions→

11. 11🚀TF Serving & Production

    0/5 lessons

    From SavedModel to a REST endpoint that scales — and why TFX exists

    How to put a model behind a production server, manage versions, and prevent the silent killer that is train/serve skew.

    Lesson list (5)

    • 01
      What Is TF Serving?~9 min · serving, production, rest, grpc
    • 02
      Docker Deployment~11 min · docker, deployment, config
    • 03
      REST and gRPC APIs~12 min · rest, grpc, api, client
    • 04
      TFX Pipelines Overview~11 min · tfx, mlops, pipelines
    • 05
      Vertex AI and Cloud Deployment~11 min · vertex-ai, cloud, managed

    ✓Quiz · 4 questions→

12. 12📈TensorBoard & Debugging

    0/4 lessons

    Watch your model train, find bottlenecks, debug systematically

    The visualization suite that turns opaque training into a visual story, plus the systematic debugging routine experienced ML engineers actually use.

    Lesson list (4)

    • 01
      TensorBoard Overview — Eight Dashboards~10 min · tensorboard, visualization, monitoring
    • 02
      tf.summary API — Custom Logging~11 min · tf-summary, scalars, images, histograms
    • 03
      HParams Dashboard and the Profiler~12 min · hparams, profiler, performance
    • 04
      Debugging Strategies for Neural Networks~13 min · debugging, nan, overfit, underfit

    ✓Quiz · 4 questions→

13. 13⚡TPUs & Distributed Training

    0/4 lessons

    From 1 GPU to a TPU pod — distribution strategies, mixed precision, scaling recipes

    How Google trains Gemini-class models — and how to make a single Keras script scale to a fleet of accelerators.

    Lesson list (4)

    • 01
      What Are TPUs?~10 min · tpu, hardware, xla
    • 02
      tf.distribute Strategies~13 min · mirrored, tpu-strategy, multi-worker
    • 03
      Mixed Precision Training~11 min · mixed-precision, fp16, bfloat16
    • 04
      MultiWorker, TF_CONFIG, and Performance Recipes~12 min · multi-worker, tf-config, scaling-recipes

    ✓Quiz · 4 questions→

14. 14🌐The TF Ecosystem & What's Next

    0/5 lessons

    TFP, TF-DF, TFRS, TF-Agents, TFF, KerasHub — and where the field is heading

    The specialized libraries that fix what core TF doesn't, and the convergence between TF/JAX/PyTorch via Keras 3 and StableHLO.

    Lesson list (5)

    • 01
      TF Probability and TF Decision Forests~11 min · tfp, tf-df, tabular, uncertainty
    • 02
      TF Recommenders, TF Agents, TF Federated~12 min · tfrs, tf-agents, tff, specialized
    • 03
      KerasCV, KerasHub, and the Keras 3 Convergence~13 min · keras-cv, keras-hub, keras3, multi-backend
    • 04
      TF and JAX — Google's Two Frameworks~11 min · jax, convergence, stablehlo
    • 05
      Where to Go Next — Career and Learning Paths~9 min · career, learning, next-steps

    ✓Quiz · 5 questions→