is_channels_last_contiguous_2d

is_channels_last_contiguous_3d

is_channels_last_strides_2d

is_channels_last_strides_3d

is_contiguous

is_non_overlapping_and_dense_indicator

method_to_operator

sympy_is_channels_last_contiguous_2d

sympy_is_channels_last_contiguous_3d

sympy_is_channels_last_strides_2d

sympy_is_channels_last_strides_3d

sympy_is_channels_last_strides_generic

sympy_is_contiguous

sympy_is_contiguous_generic

to_node

ShapeEnv

DimDynamic

Controls how to perform symbol allocation for a dimension.

StrictMinMaxConstraint

For clients: the size at this dimension must be within 'vr' (which specifies a lower and upper bound, inclusive-inclusive) AND it must be non-negative and should not be 0 or 1 (but see NB below).

RelaxedUnspecConstraint

For clients: no explicit constraint; constraint is whatever is implicitly inferred by guards from tracing.

EqualityConstraint

Represent and decide various kinds of equality constraints between input sources.

SymbolicContext

Data structure specifying how we should create symbols increate_symbolic_sizes_strides_storage_offset; e.g., should they be static or dynamic.

StatelessSymbolicContext

Create symbols increate_symbolic_sizes_strides_storage_offset via a symbolic_context determination as given byDimDynamic andDimConstraint.

StatefulSymbolicContext

Create symbols increate_symbolic_sizes_strides_storage_offset via a symbolic_context determination as given by a cache of Source:Symbol.

SubclassSymbolicContext

The correct symbolic context for a given inner tensor of a traceable tensor subclass may differ from that of the outer symbolic context.

DimConstraints

Custom solver for a system of constraints on symbolic dimensions.

ShapeEnvSettings

Encapsulates all shape env settings that could potentially affect FakeTensor dispatch.

ConvertIntKey

CallMethodKey

PropagateUnbackedSymInts

DivideByKey

InnerTensorKey

Specialization

This class is used in multi-graph compilation contexts where we generate multiple specialized graphs and dispatch to the appropriate one at runtime.

hint_int

Retrieve the hint for an int (based on the underlying real values as observed at runtime).

is_concrete_int

Utility to check if underlying object in SymInt is concrete value.

is_concrete_bool

Utility to check if underlying object in SymBool is concrete value.

is_concrete_float

Utility to check if underlying object in SymInt is concrete value.

has_free_symbols

Faster version of bool(free_symbols(val))

has_free_unbacked_symbols

Faster version of bool(free_unbacked_symbols(val))

guard_or_true

Try to guard a, if data dependent error encountered just return true.

guard_or_false

Try to guard a, if data dependent error encountered just return false.

guard_size_oblivious

Perform a guard on a symbolic boolean expression in a size oblivious way.

sym_and

and, but for symbolic expressions, without bool casting.

sym_eq

Like ==, but when run on list/tuple, it will recursively test equality and use sym_and to join the results together, without guarding.

sym_or

or, but for symbolic expressions, without bool casting.

constrain_range

Applies a constraint that the passed in SymInt must lie between min-max inclusive-inclusive, WITHOUT introducing a guard on the SymInt (meaning that it can be used on unbacked SymInts).

constrain_unify

Given two SymInts, constrain them so that they must be equal.

canonicalize_bool_expr

Canonicalize a boolean expression by transforming it into a lt / le inequality and moving all the non-constant terms to the rhs.

statically_known_true

Returns True if x can be simplified to a constant and is true.

statically_known_false

Returns True if x can be simplified to a constant and is False.

has_static_value

User-code friendly utility to check if a value is static or dynamic.

lru_cache

check_consistent

Test that two "meta" values (typically either Tensor or SymInt) have the same values, e.g., after retracing.

compute_unbacked_bindings

After having run fake tensor propagation and producing example_value result, traverse example_value looking for freshly bound unbacked symbols and record their paths for later.

rebind_unbacked

Suppose we are retracing a pre-existing FX graph that previously had fake tensor propagation (and therefore unbacked SymInts).

resolve_unbacked_bindings

When we do fake tensor prop, we oftentimes will allocate new unbacked symints.

is_accessor_node

Helper function to determine if a node is trying to access a symbolic integer such as size, stride, offset or item.

cast_symbool_to_symint_guardless

Converts a SymBool or bool to a SymInt or int without introducing guards.

create_contiguous

error

eval_guards

eval_is_non_overlapping_and_dense

find_symbol_binding_fx_nodes

Find all nodes in an FX graph that bind sympy Symbols.

free_symbols

Recursively collect all free symbols from a value.

free_unbacked_symbols

Like free_symbols, but filtered to only report unbacked symbols

fx_placeholder_targets

fx_placeholder_vals

guard_bool

guard_float

guard_int

guard_scalar

Guard a scalar value, which can be a symbolic or concrete boolean, integer, or float.

has_hint

has_symbolic_sizes_strides

is_nested_int

is_symbol_binding_fx_node

Check if a given FX node is a symbol binding node.

is_symbolic

expect_true

log_lru_cache_stats

make_fx

Given a function f, return a new function which when executed with valid arguments to f, returns an FX GraphModule representing the set of operations that were executed during the course of execution.

handle_sym_dispatch

Call into the currently active proxy tracing mode to do a SymInt/SymFloat/SymBool dispatch trace on a function that operates on these arguments.

get_proxy_mode

Current the currently active proxy tracing mode, or None if we are not currently tracing.

maybe_enable_thunkify

Within this context manager, if you are doing make_fx tracing, we will thunkify all SymNode compute and avoid tracing it into the graph unless it is actually needed.

maybe_disable_thunkify

Within a context, disable thunkification.

thunkify

Delays computation of f until it's called again Also caches the result

track_tensor

track_tensor_tree

decompose

disable_autocast_cache

disable_proxy_modes_tracing

extract_val

fake_signature

FX gets confused by varargs, de-confuse it

fetch_object_proxy

fetch_sym_proxy

has_proxy_slot

is_sym_node

maybe_handle_decomp

proxy_call

set_meta

set_original_aten_op

set_proxy_slot

snapshot_fake

assoc

Return a new dict with new key value pair

assoc_in

Return a new dict with new, potentially nested, key value pair

dissoc

Return a new dict with the given key(s) removed.

first

The first element in a sequence

keyfilter

Filter items in dictionary by key

keymap

Apply function to keys of dictionary

merge

Merge a collection of dictionaries

merge_with

Merge dictionaries and apply function to combined values

update_in

Update value in a (potentially) nested dictionary

valfilter

Filter items in dictionary by value

valmap

Apply function to values of dictionary

itemfilter

Filter items in dictionary by item

itemmap

Apply function to items of dictionary

transform_algebraic_expression

Transforms an algebraic expression to z3 format :param expr: An expression is either a dimension variable or an algebraic-expression

transform_all_constraints

Given a trace, generates constraints and transforms them to z3 format

transform_all_constraints_trace_time

Takes a node and a graph and generates two sets of constraints.

transform_dimension

Takes a dimension variable or a number and transforms it to a tuple according to our scheme :param dimension: The dimension to be transformed :param counter: variable tracking

transform_to_z3

transform_var

Transforms tensor variables to a format understood by z3 :param tensor: Tensor variable or a tensor type potentially with variable dimensions