Free practice problems
Free Deep Learning Interview Questions
Backprop, optimizers, normalization, CNNs, RNNs, training dynamics. 22 of the 123 Deep Learning problems in the LeetML bank are free (7 easy, 11 medium, 4 hard). Open any of them to read the statement, then solve it in your browser.
22 free Deep Learning problems
- Self-Attention, LoRA, Optimizers and ViT Patch SizeConceptualMediumSolve
- Regularizers, Adam vs AdamW and the Decoder BlockConceptualHardSolve
- Deriving Loss Gradients by Hand: Softmax Cross-EntropyCalculationEasySolve
- Fully Connected Networks vs CNNsConceptualEasySolve
- CNNs Versus Transformers Under Data and Latency LimitsConceptualMediumSolve
- Two Regression Heads With a Negative, Heavy-Tailed TargetConceptualMediumSolve
- VAE Objective: ELBO Gap, Gaussian KL and Posterior CollapseConceptualMediumSolve
- Conv2D Arithmetic: Parameters, Output Size and ComputeCalculationEasySolve
- What torch.distributed.barrier Does and How It FailsConceptualMediumSolve
- Memory of a Sharded Array on a Device MeshCalculationEasySolve
101 more Deep Learning problems are in the full bank inside the app.