| {"id": "minitorch-001", "question": "Which function computes the derivative of scalar addition?", "answer_node": "minitorch.scalar.Scalar.add_back", "module": 1, "reasoning_type": "autodiff"} | |
| {"id": "minitorch-002", "question": "Where is the backward pass for scalar multiplication implemented?", "answer_node": "minitorch.scalar.Scalar.mul_back", "module": 1, "reasoning_type": "autodiff"} | |
| {"id": "minitorch-003", "question": "Which method computes the gradient for scalar negation?", "answer_node": "minitorch.scalar.Scalar.neg_back", "module": 1, "reasoning_type": "autodiff"} | |
| {"id": "minitorch-004", "question": "How is the chain rule applied during scalar backpropagation?", "answer_node": "minitorch.autodiff.Variable.chain_rule", "module": 1, "reasoning_type": "autodiff"} | |
| {"id": "minitorch-005", "question": "Which class stores the computational graph for scalars?", "answer_node": "minitorch.autodiff.Variable", "module": 1, "reasoning_type": "graph"} | |
| {"id": "minitorch-006", "question": "Where is the ReLU backward function for scalars defined?", "answer_node": "minitorch.scalar.Scalar.relu_back", "module": 1, "reasoning_type": "activation"} | |
| {"id": "minitorch-007", "question": "Which function implements scalar exponentiation backward?", "answer_node": "minitorch.scalar.Scalar.exp_back", "module": 1, "reasoning_type": "autodiff"} | |
| {"id": "minitorch-008", "question": "How is the log derivative computed for scalars?", "answer_node": "minitorch.scalar.Scalar.log_back", "module": 1, "reasoning_type": "autodiff"} | |
| {"id": "minitorch-009", "question": "Which method handles scalar division backward?", "answer_node": "minitorch.scalar.Scalar.div_back", "module": 1, "reasoning_type": "autodiff"} | |
| {"id": "minitorch-010", "question": "Where is the sigmoid backward pass implemented for scalars?", "answer_node": "minitorch.scalar.Scalar.sigmoid_back", "module": 1, "reasoning_type": "activation"} | |
| {"id": "minitorch-011", "question": "Which class represents a trainable parameter in a neural network?", "answer_node": "minitorch.module.Parameter", "module": 2, "reasoning_type": "nn"} | |
| {"id": "minitorch-012", "question": "How is a linear layer's forward pass implemented?", "answer_node": "minitorch.nn.Linear.forward", "module": 2, "reasoning_type": "nn"} | |
| {"id": "minitorch-013", "question": "Where are neural network modules stored in a container?", "answer_node": "minitorch.module.Module._modules", "module": 2, "reasoning_type": "structure"} | |
| {"id": "minitorch-014", "question": "Which function initializes weights using Xavier uniform?", "answer_node": "minitorch.module.xavier_uniform", "module": 2, "reasoning_type": "init"} | |
| {"id": "minitorch-015", "question": "How are model parameters collected recursively?", "answer_node": "minitorch.module.Module.parameters", "module": 2, "reasoning_type": "traversal"} | |
| {"id": "minitorch-016", "question": "Which class implements the base neural network module?", "answer_node": "minitorch.module.Module", "module": 2, "reasoning_type": "nn"} | |
| {"id": "minitorch-017", "question": "Where is the sequential container defined?", "answer_node": "minitorch.module.Sequential", "module": 2, "reasoning_type": "composition"} | |
| {"id": "minitorch-018", "question": "Which method resets a module's gradients to zero?", "answer_node": "minitorch.module.Module.zero_grad", "module": 2, "reasoning_type": "optim"} | |
| {"id": "minitorch-019", "question": "How is the MSE loss computed?", "answer_node": "minitorch.nn.loss.MSELoss.forward", "module": 2, "reasoning_type": "loss"} | |
| {"id": "minitorch-020", "question": "Which optimizer updates parameters using stochastic gradient descent?", "answer_node": "minitorch.optim.SGD.step", "module": 2, "reasoning_type": "optim"} | |
| {"id": "minitorch-021", "question": "Which class represents a multi-dimensional tensor?", "answer_node": "minitorch.tensor.Tensor", "module": 3, "reasoning_type": "tensor"} | |
| {"id": "minitorch-022", "question": "Where is tensor addition implemented?", "answer_node": "minitorch.tensor.Tensor.add", "module": 3, "reasoning_type": "op"} | |
| {"id": "minitorch-023", "question": "How is tensor view creation handled without copying data?", "answer_node": "minitorch.tensor.Tensor.view", "module": 3, "reasoning_type": "memory"} | |
| {"id": "minitorch-024", "question": "Which function computes tensor strides from shape and storage?", "answer_node": "minitorch.tensor.Tensor._compute_strides", "module": 3, "reasoning_type": "memory"} | |
| {"id": "minitorch-025", "question": "Where is the tensor backward pass triggered?", "answer_node": "minitorch.tensor.Tensor.backward", "module": 3, "reasoning_type": "autodiff"} | |
| {"id": "minitorch-026", "question": "How is tensor slicing implemented?", "answer_node": "minitorch.tensor.Tensor.__getitem__", "module": 3, "reasoning_type": "indexing"} | |
| {"id": "minitorch-027", "question": "Which method reshapes a tensor?", "answer_node": "minitorch.tensor.Tensor.reshape", "module": 3, "reasoning_type": "op"} | |
| {"id": "minitorch-028", "question": "Where is the tensor sum reduction implemented?", "answer_node": "minitorch.tensor.Tensor.sum", "module": 3, "reasoning_type": "reduction"} | |
| {"id": "minitorch-029", "question": "How is tensor broadcasting handled during operations?", "answer_node": "minitorch.tensor.broadcast_index", "module": 3, "reasoning_type": "broadcast"} | |
| {"id": "minitorch-030", "question": "Which function creates a tensor from a Python list?", "answer_node": "minitorch.tensor.tensor", "module": 3, "reasoning_type": "creation"} | |
| {"id": "minitorch-031", "question": "Which backend handles parallel tensor operations?", "answer_node": "minitorch.fast_ops", "module": 4, "reasoning_type": "backend"} | |
| {"id": "minitorch-032", "question": "Where is the map operation implemented for tensors?", "answer_node": "minitorch.tensor_ops.map", "module": 4, "reasoning_type": "parallel"} | |
| {"id": "minitorch-033", "question": "How is tensor matrix multiplication computed in parallel?", "answer_node": "minitorch.tensor_ops.tensor_matrix_multiply", "module": 4, "reasoning_type": "parallel"} | |
| {"id": "minitorch-034", "question": "Which function applies a scalar operation element-wise?", "answer_node": "minitorch.tensor_ops.zip", "module": 4, "reasoning_type": "parallel"} | |
| {"id": "minitorch-035", "question": "Where is the reduce operation implemented?", "answer_node": "minitorch.tensor_ops.reduce", "module": 4, "reasoning_type": "parallel"} | |
| {"id": "minitorch-036", "question": "Which module contains optimized tensor operations?", "answer_node": "minitorch.tensor_ops", "module": 4, "reasoning_type": "backend"} | |
| {"id": "minitorch-037", "question": "How are tensor operations dispatched to the correct backend?", "answer_node": "minitorch.tensor.Tensor._op", "module": 4, "reasoning_type": "dispatch"} | |
| {"id": "minitorch-038", "question": "Which function handles tensor indexing in parallel?", "answer_node": "minitorch.tensor_ops.index", "module": 4, "reasoning_type": "parallel"} | |
| {"id": "minitorch-039", "question": "Where is the tensor permute operation implemented?", "answer_node": "minitorch.tensor_ops.permute", "module": 4, "reasoning_type": "op"} | |
| {"id": "minitorch-040", "question": "Which backend is used for CPU tensor operations?", "answer_node": "minitorch.tensor_ops", "module": 4, "reasoning_type": "backend"} | |
| {"id": "minitorch-041", "question": "Which decorator marks a function for CUDA compilation?", "answer_node": "minitorch.cuda.jit", "module": 5, "reasoning_type": "cuda"} | |
| {"id": "minitorch-042", "question": "Where is the CUDA tensor storage implemented?", "answer_node": "minitorch.cuda.CudaStorage", "module": 5, "reasoning_type": "cuda"} | |
| {"id": "minitorch-043", "question": "How is CUDA memory allocated for tensors?", "answer_node": "minitorch.cuda.cuda_malloc", "module": 5, "reasoning_type": "cuda"} | |
| {"id": "minitorch-044", "question": "Which function copies data from CPU to GPU?", "answer_node": "minitorch.cuda.cuda_copy_to", "module": 5, "reasoning_type": "cuda"} | |
| {"id": "minitorch-045", "question": "Where is the CUDA tensor class defined?", "answer_node": "minitorch.cuda.CudaTensor", "module": 5, "reasoning_type": "cuda"} | |
| {"id": "minitorch-046", "question": "How are CUDA kernels launched for tensor operations?", "answer_node": "minitorch.cuda.tensor_ops", "module": 5, "reasoning_type": "cuda"} | |
| {"id": "minitorch-047", "question": "Which module contains CUDA-optimized operations?", "answer_node": "minitorch.cuda.ops", "module": 5, "reasoning_type": "cuda"} | |
| {"id": "minitorch-048", "question": "Where is the CUDA softmax implemented?", "answer_node": "minitorch.cuda.ops.softmax", "module": 5, "reasoning_type": "cuda"} | |
| {"id": "minitorch-049", "question": "How is CUDA tensor backward pass handled?", "answer_node": "minitorch.cuda.CudaTensor.backward", "module": 5, "reasoning_type": "cuda"} | |
| {"id": "minitorch-050", "question": "Which function synchronizes CUDA streams?", "answer_node": "minitorch.cuda.cuda_synchronize", "module": 5, "reasoning_type": "cuda"} | |
| {"id": "minitorch-051", "question": "Where is 2D convolution forward implemented?", "answer_node": "minitorch.tensor.Tensor.conv2d", "module": 6, "reasoning_type": "conv"} | |
| {"id": "minitorch-052", "question": "Which function computes the gradient of 2D convolution?", "answer_node": "minitorch.tensor.Tensor.conv2d_back", "module": 6, "reasoning_type": "conv"} | |
| {"id": "minitorch-053", "question": "How is max pooling implemented?", "answer_node": "minitorch.tensor.Tensor.max_pool2d", "module": 6, "reasoning_type": "pooling"} | |
| {"id": "minitorch-054", "question": "Where is the max pooling backward pass defined?", "answer_node": "minitorch.tensor.Tensor.max_pool2d_back", "module": 6, "reasoning_type": "pooling"} | |
| {"id": "minitorch-055", "question": "Which function implements dropout during training?", "answer_node": "minitorch.tensor.Tensor.dropout", "module": 6, "reasoning_type": "regularization"} | |
| {"id": "minitorch-056", "question": "How is batch normalization forward pass computed?", "answer_node": "minitorch.tensor.Tensor.batch_norm", "module": 6, "reasoning_type": "norm"} | |
| {"id": "minitorch-057", "question": "Where is the layer normalization implemented?", "answer_node": "minitorch.tensor.Tensor.layer_norm", "module": 6, "reasoning_type": "norm"} | |
| {"id": "minitorch-058", "question": "Which method computes the softmax function?", "answer_node": "minitorch.tensor.Tensor.softmax", "module": 6, "reasoning_type": "activation"} | |
| {"id": "minitorch-059", "question": "How is the cross-entropy loss computed?", "answer_node": "minitorch.nn.loss.CrossEntropyLoss.forward", "module": 6, "reasoning_type": "loss"} | |
| {"id": "minitorch-060", "question": "Where is the embedding lookup implemented?", "answer_node": "minitorch.tensor.Tensor.embedding", "module": 6, "reasoning_type": "embedding"} | |
| {"id": "minitorch-061", "question": "Which function creates a scalar from a Python float?", "answer_node": "minitorch.scalar.Scalar", "module": 0, "reasoning_type": "creation"} | |
| {"id": "minitorch-062", "question": "How is scalar addition defined?", "answer_node": "minitorch.scalar.Scalar.__add__", "module": 0, "reasoning_type": "op"} | |
| {"id": "minitorch-063", "question": "Where is the scalar ReLU function implemented?", "answer_node": "minitorch.scalar.Scalar.relu", "module": 0, "reasoning_type": "activation"} | |
| {"id": "minitorch-064", "question": "Which method converts a scalar to a Python float?", "answer_node": "minitorch.scalar.Scalar.data", "module": 0, "reasoning_type": "conversion"} | |
| {"id": "minitorch-065", "question": "How is scalar comparison implemented?", "answer_node": "minitorch.scalar.Scalar.__gt__", "module": 0, "reasoning_type": "op"} | |
| {"id": "minitorch-066", "question": "Where is the scalar log function defined?", "answer_node": "minitorch.scalar.Scalar.log", "module": 0, "reasoning_type": "math"} | |
| {"id": "minitorch-067", "question": "Which function computes scalar exponentiation?", "answer_node": "minitorch.scalar.Scalar.exp", "module": 0, "reasoning_type": "math"} | |
| {"id": "minitorch-068", "question": "How is scalar negation implemented?", "answer_node": "minitorch.scalar.Scalar.__neg__", "module": 0, "reasoning_type": "op"} | |
| {"id": "minitorch-069", "question": "Where is the scalar sigmoid function defined?", "answer_node": "minitorch.scalar.Scalar.sigmoid", "module": 0, "reasoning_type": "activation"} | |
| {"id": "minitorch-070", "question": "Which method checks if a scalar requires gradient?", "answer_node": "minitorch.scalar.Scalar.requires_grad", "module": 0, "reasoning_type": "autodiff"} | |
| {"id": "minitorch-071", "question": "Which class manages the global random number generator?", "answer_node": "minitorch.rand.RandomGenerator", "module": 2, "reasoning_type": "random"} | |
| {"id": "minitorch-072", "question": "How is uniform random tensor generation implemented?", "answer_node": "minitorch.rand.rand", "module": 2, "reasoning_type": "random"} | |
| {"id": "minitorch-073", "question": "Where is the normal distribution sampler defined?", "answer_node": "minitorch.rand.randn", "module": 2, "reasoning_type": "random"} | |
| {"id": "minitorch-074", "question": "Which function sets the random seed?", "answer_node": "minitorch.rand.seed", "module": 2, "reasoning_type": "random"} | |
| {"id": "minitorch-075", "question": "How are model weights initialized randomly?", "answer_node": "minitorch.module.rand", "module": 2, "reasoning_type": "init"} | |
| {"id": "minitorch-076", "question": "Which function computes the L2 norm of a tensor?", "answer_node": "minitorch.tensor.Tensor.norm", "module": 3, "reasoning_type": "reduction"} | |
| {"id": "minitorch-077", "question": "Where is tensor transposition implemented?", "answer_node": "minitorch.tensor.Tensor.T", "module": 3, "reasoning_type": "op"} | |
| {"id": "minitorch-078", "question": "How is tensor concatenation performed?", "answer_node": "minitorch.tensor.Tensor.cat", "module": 3, "reasoning_type": "op"} | |
| {"id": "minitorch-079", "question": "Which method returns the tensor shape?", "answer_node": "minitorch.tensor.Tensor.shape", "module": 3, "reasoning_type": "property"} | |
| {"id": "minitorch-080", "question": "Where is the tensor item() method defined?", "answer_node": "minitorch.tensor.Tensor.item", "module": 3, "reasoning_type": "conversion"} | |
| {"id": "minitorch-081", "question": "Which CUDA function implements element-wise addition?", "answer_node": "minitorch.cuda.ops.add", "module": 5, "reasoning_type": "cuda"} | |
| {"id": "minitorch-082", "question": "How is CUDA tensor creation from CPU data handled?", "answer_node": "minitorch.cuda.CudaTensor.from_numpy", "module": 5, "reasoning_type": "cuda"} | |
| {"id": "minitorch-083", "question": "Where is the CUDA reduce operation implemented?", "answer_node": "minitorch.cuda.ops.reduce", "module": 5, "reasoning_type": "cuda"} | |
| {"id": "minitorch-084", "question": "Which function handles CUDA tensor indexing?", "answer_node": "minitorch.cuda.ops.index", "module": 5, "reasoning_type": "cuda"} | |
| {"id": "minitorch-085", "question": "How is CUDA memory freed?", "answer_node": "minitorch.cuda.cuda_free", "module": 5, "reasoning_type": "cuda"} | |
| {"id": "minitorch-086", "question": "Which function implements adaptive average pooling?", "answer_node": "minitorch.tensor.Tensor.adaptive_avg_pool2d", "module": 6, "reasoning_type": "pooling"} | |
| {"id": "minitorch-087", "question": "Where is the group normalization implemented?", "answer_node": "minitorch.tensor.Tensor.group_norm", "module": 6, "reasoning_type": "norm"} | |
| {"id": "minitorch-088", "question": "How is the GELU activation computed?", "answer_node": "minitorch.tensor.Tensor.gelu", "module": 6, "reasoning_type": "activation"} | |
| {"id": "minitorch-089", "question": "Which method computes the log-softmax?", "answer_node": "minitorch.tensor.Tensor.log_softmax", "module": 6, "reasoning_type": "activation"} | |
| {"id": "minitorch-090", "question": "Where is the NLL loss implemented?", "answer_node": "minitorch.nn.loss.NLLLoss.forward", "module": 6, "reasoning_type": "loss"} | |
| {"id": "minitorch-091", "question": "Which function creates a tensor filled with zeros?", "answer_node": "minitorch.tensor.zeros", "module": 3, "reasoning_type": "creation"} | |
| {"id": "minitorch-092", "question": "How is a tensor of ones created?", "answer_node": "minitorch.tensor.ones", "module": 3, "reasoning_type": "creation"} | |
| {"id": "minitorch-093", "question": "Where is the arange function implemented?", "answer_node": "minitorch.tensor.arange", "module": 3, "reasoning_type": "creation"} | |
| {"id": "minitorch-094", "question": "Which method creates an identity matrix?", "answer_node": "minitorch.tensor.eye", "module": 3, "reasoning_type": "creation"} | |
| {"id": "minitorch-095", "question": "How is a tensor filled with a constant value created?", "answer_node": "minitorch.tensor.full", "module": 3, "reasoning_type": "creation"} | |
| {"id": "minitorch-096", "question": "Which optimizer uses momentum in parameter updates?", "answer_node": "minitorch.optim.SGD", "module": 2, "reasoning_type": "optim"} | |
| {"id": "minitorch-097", "question": "How is the learning rate scheduler implemented?", "answer_node": "minitorch.optim.LRScheduler", "module": 2, "reasoning_type": "optim"} | |
| {"id": "minitorch-098", "question": "Where is the Adam optimizer defined?", "answer_node": "minitorch.optim.Adam", "module": 2, "reasoning_type": "optim"} | |
| {"id": "minitorch-099", "question": "Which function clips gradients by norm?", "answer_node": "minitorch.optim.clip_grad_norm", "module": 2, "reasoning_type": "optim"} | |
| {"id": "minitorch-100", "question": "How is model state saved to disk?", "answer_node": "minitorch.module.save", "module": 2, "reasoning_type": "io"} |