rime.util.score_array

Functions

auto_cast_lazy_score(other)
auto_device()
auto_tensor(x[, device])
batch_op_iter(S, op[, device])
find_iloc(old_index, new_index[, allow_missing])
get_batch_size(shape)	round to similar batch sizes
matrix_reindex(csr, old_index, new_index, axis)	pandas.reindex functionality on sparse or dense matrices as well as 1d arrays
score_op(S, op[, device, reduce_fn])	aggregation operations (e.g., max, min, sum) across entire matrix
sps_to_torch(x, device)	convert scipy.sparse to torch.sparse

Classes

ElementWiseExpression(op, children)	Tree representation of an element-wise expression; auto-broadcast
LazyDenseMatrix(c)	cast scalar and arrays to 2-d arrays
LazyExpressionBase(op, children)
LazyScoreBase(shape)	low-rank, sparse, and other expressions with deferred evaluations after subsampling
LazyScoreModel(user_index, item_index[, ...])
LazySparseMatrix(c)
MatMulExpression(op, children)
RandScore(row_seeds, col_seeds[, distn])	add random noise to break ties
VAEExpression(op, children)	op=_VAEOp, [weight (n, 2d), randn (n, d)], auto-broadcast on d-axis