| import copy |
| import operator |
| from copy import deepcopy |
| from typing import cast, Dict, List, Optional, Union |
| |
| import torch |
| import torch.fx._pytree as fx_pytree |
| import torch.utils._pytree as pytree |
| from torch.export import ExportedProgram |
| from torch.export.exported_program import ( |
| ConstantArgument, |
| ModuleCallSignature, |
| SymIntArgument, |
| TensorArgument, |
| ) |
| from torch.fx import GraphModule |
| from .utils import _check_input_constraints_pre_hook |
| |
| |
| # Assign attribute 'from_obj' to the qualified name 'target' on 'to_module |
| # This installs empty Modules where none exist yet if they are subpaths of target |
| def _assign_attr( |
| from_obj: torch.Tensor, |
| to_module: torch.nn.Module, |
| target: str, |
| is_parameter: bool, |
| ): |
| *prefix, field = target.split(".") |
| for item in prefix: |
| t = getattr(to_module, item, None) |
| |
| if t is None: |
| t = torch.nn.Module() |
| setattr(to_module, item, t) |
| to_module = t |
| |
| # If it is a tensor and not a parameter attribute of a module, it should be a named buffer. |
| # So, we register it as a named buffer in the target module. |
| if not isinstance(from_obj, torch.Tensor): |
| raise ValueError("Expected only parameters or buffers, got:", type(from_obj)) |
| |
| if is_parameter: |
| to_module.register_parameter(field, torch.nn.Parameter(from_obj)) |
| else: |
| to_module.register_buffer(field, from_obj) |
| |
| |
| class _UnflattenedModule(torch.fx.GraphModule): |
| def __init__(self, export_module: ExportedProgram): |
| if export_module.graph_signature.backward_signature is not None: |
| raise ValueError("Unflattening on JointExportModule NYI") |
| super().__init__({}, torch.fx.Graph(), "_UnflattenedModule") |
| |
| export_graph = deepcopy(export_module.graph) |
| self.graph_signature = deepcopy(export_module.graph_signature) |
| self.module_call_graph = deepcopy(export_module.module_call_graph) |
| _inplace_buffer_mutations(export_graph, self.graph_signature) |
| _outline_submodules(export_graph, self) |
| |
| self.range_constraints = export_module.range_constraints |
| self.equality_constraints = export_module.equality_constraints |
| |
| state_dict = export_module.state_dict |
| for name in self.graph_signature.parameters: |
| cloned = state_dict[name].clone() |
| _assign_attr( |
| cloned, |
| self, |
| name, |
| is_parameter=True, |
| ) |
| for name in self.graph_signature.buffers: |
| cloned = state_dict[name].clone() |
| _assign_attr( |
| cloned, |
| self, |
| name, |
| is_parameter=False, |
| ) |
| |
| inputs_to_state: Dict[str, str] = { |
| **self.graph_signature.inputs_to_parameters, |
| **self.graph_signature.inputs_to_buffers, |
| } |
| |
| _sink_params(self, inputs_to_state, []) |
| # Check all input nodes has been processed. |
| for module in self.modules(): |
| if not isinstance(module, torch.fx.GraphModule): |
| continue |
| for node in module.graph.nodes: |
| if node.op != "placeholder": |
| continue |
| assert node.name not in inputs_to_state |
| |
| def __call__(self, *args, **kwargs): |
| flat_args, in_spec = pytree.tree_flatten((args, kwargs)) |
| assert self.module_call_graph[0].fqn == "" |
| signature = self.module_call_graph[0].signature |
| if in_spec != signature.in_spec: |
| raise TypeError( |
| f"Input treespec does not match with exported module's. " |
| "Are you sure you are calling this with the right arguments? " |
| f"Input treespec: {in_spec}. ", |
| f"Exported module treespec: {signature.in_spec}", |
| ) |
| |
| # TODO(zhxchen17) Use lineno map to dump the original stacktrace during error handling. |
| tree_out = super().__call__(*flat_args) |
| return pytree.tree_unflatten(tree_out, signature.out_spec) |
| |
| |
| def unflatten(module: ExportedProgram) -> _UnflattenedModule: |
| """Unflatten an ExportedProgram, producing a module with the same module |
| hierarchy as the original eager module. |
| """ |
| module = _UnflattenedModule(module) |
| module.register_forward_pre_hook(_check_input_constraints_pre_hook) |
| return module |
| |
| |
| def _inplace_buffer_mutations(graph: torch.fx.Graph, graph_signature) -> None: |
| """Transform buffer mutations from their functionalized form into a copy_ |
| node in the graph. |
| |
| Functionalization represents buffer mutation by passing the buffer as an input and output. So for example, the eager code: |
| def forward(self, x): |
| self.buffer += x |
| return x * x |
| |
| Will become a graph that looks like: |
| def forward(self, buffer, x): |
| mutated_buffer = aten.add(buffer, x) |
| mul = aten.mul(x, x) |
| return (mutated_buffer, mul) |
| |
| We want to inplace this into something that looks like the original eager code: |
| def forward(self, buffer, x): |
| mutated_buffer = aten.add(buffer, x) |
| buffer.copy_(mutated_buffer) |
| mul = aten.mul(x, x) |
| return (mul,) |
| """ |
| output_node = next(iter(reversed(graph.nodes))) |
| assert output_node.op == "output" and len(output_node.args) == 1 |
| return_args = output_node.args[0] |
| |
| mutation_node_to_buffer = graph_signature.buffers_to_mutate |
| mutations = return_args[: len(mutation_node_to_buffer)] |
| buffers_to_inputs = {v: k for k, v in graph_signature.inputs_to_buffers.items()} |
| input_name_to_node = { |
| node.name: node for node in graph.nodes if node.op == "placeholder" |
| } |
| |
| for mutation in mutations: |
| buffer_name = mutation_node_to_buffer[mutation.name] |
| input_name = buffers_to_inputs[buffer_name] |
| input_node = input_name_to_node[input_name] |
| |
| with graph.inserting_after(mutation): |
| new_node = graph.create_node( |
| "call_function", torch.ops.aten.copy_, (input_node, mutation) |
| ) |
| for k, v in mutation.meta.items(): |
| new_node.meta[k] = v |
| # Replace all uses of the previously functional mutation with our copy_ output. |
| mutation.replace_all_uses_with(new_node, lambda x: x is not new_node) |
| |
| # Remove the mutated buffer from the graph outputs, since we don't need to |
| # thread it through anymore. We don't need to handle the inputs, which will |
| # be handled by _sink_params. |
| user_outputs = tuple( |
| return_args[len(mutation_node_to_buffer) :], |
| ) |
| output_node.args = ((user_outputs),) |
| |
| |
| def is_prefix(candidate, target): |
| """Check whether `candidate` is a prefix of `target`.""" |
| return len(candidate) < len(target) and target[: len(candidate)] == candidate |
| |
| |
| def compute_accessor(parent_fqn: str, child_fqn: str) -> str: |
| if parent_fqn == "": |
| # Handle the root module correctly. |
| return child_fqn |
| |
| parent_split = parent_fqn.split(".") |
| child_split = child_fqn.split(".") |
| |
| assert ( |
| child_split[: len(parent_split)] == parent_split |
| ), f"Child module '{child_fqn}' is not a descendant of parent module '{parent_fqn}'" |
| return ".".join(child_split[len(parent_split) :]) |
| |
| |
| def _verify_graph_equivalence(x: torch.fx.GraphModule, y: torch.fx.GraphModule): |
| def graph_dump(graph: torch.fx.Graph) -> str: |
| ret = [] |
| nodes_idx: Dict[int, int] = {} |
| |
| def arg_dump(arg) -> str: |
| if isinstance(arg, torch.fx.Node): |
| return "%" + str(nodes_idx[id(arg)]) |
| return str(arg) |
| |
| for i, node in enumerate(graph.nodes): |
| args_dump = [str(arg) for arg in pytree.tree_map(arg_dump, node.args)] |
| args_dump += [ |
| f"{key}={value}" |
| for key, value in pytree.tree_map(arg_dump, node.kwargs).items() |
| ] |
| target = node.target if node.op == "call_function" else "" |
| ret.append(f"{i}: {node.op}[{target}]({', '.join(args_dump)})") |
| nodes_idx[id(node)] = i |
| return "\n".join(ret) |
| |
| assert graph_dump(x.graph) == graph_dump(y.graph) |
| |
| |
| def _add_spec(gm: torch.fx.GraphModule, spec) -> str: |
| i = 0 |
| while hasattr(gm, f"_spec_{i}"): |
| i += 1 |
| name = f"_spec_{i}" |
| setattr(gm, name, spec) |
| return name |
| |
| |
| def _generate_flatten(gm: torch.fx.GraphModule, node, spec) -> torch.fx.Node: |
| name = _add_spec(gm, spec) |
| spec_node = gm.graph.get_attr(name) |
| return gm.graph.call_function(fx_pytree.tree_flatten_spec, (node, spec_node)) |
| |
| |
| def _generate_unflatten(gm: torch.fx.GraphModule, nodes, spec) -> torch.fx.Node: |
| name = _add_spec(gm, spec) |
| spec_node = gm.graph.get_attr(name) |
| return gm.graph.call_function(pytree.tree_unflatten, (nodes, spec_node)) |
| |
| |
| class ModuleFrame: |
| def __init__( |
| self, |
| flat_graph, |
| seen_nodes, |
| seen_modules, |
| parent, |
| module_stack, |
| module_id, |
| module_call_graph: Dict[str, ModuleCallSignature], |
| graph_module=None, |
| ): |
| self.flat_graph = flat_graph |
| self.seen_nodes = seen_nodes |
| self.seen_modules = seen_modules |
| self.parent = parent |
| self.module_stack = module_stack |
| self.module_id = module_id |
| |
| self.module_call_graph = module_call_graph |
| self.verbose = False |
| |
| self.fqn = self.module_stack[-1] |
| if graph_module is not None: |
| self.graph_module = graph_module |
| else: |
| # InterpreterModule doesn't work with torch.compile: |
| # 1. in-place compile: nn.Module compile the forward function, and if we overwrite __call__, |
| # in-place compile will not be effective |
| # 2. out-of-place compile: there are a lot of graph guard failures on "self" in the |
| # InterpreterModule |
| # self.graph_module = InterpreterModule( |
| self.graph_module = torch.fx.GraphModule( |
| {}, |
| torch.fx.Graph(), |
| self.fqn, |
| ) |
| self.graph_module.meta["module_call_signature"] = module_call_graph.get( |
| self.fqn |
| ) |
| |
| if self.module_id in self.seen_modules: |
| self.cached_graph_module = self.seen_modules[self.module_id] |
| else: |
| self.cached_graph_module = None |
| self.seen_modules[self.module_id] = self.graph_module |
| |
| self.nodes = list(self.flat_graph.nodes) |
| self.graph = self.graph_module.graph |
| |
| # Mapping of nodes in the flat graph to nodes in this graph. |
| self.node_map: Dict[torch.fx.Node, torch.fx.Node] = {} |
| self.node_to_placeholder = {} |
| |
| self.parent_call_module: Optional[torch.fx.Node] = None |
| if parent is not None: |
| accessor = compute_accessor(parent.fqn, self.fqn) |
| parent.graph_module.add_submodule( |
| accessor, |
| self.graph_module |
| if self.cached_graph_module is None |
| else self.cached_graph_module, |
| ) |
| self.parent_call_module = parent.graph.call_module(accessor) |
| |
| signature = module_call_graph.get(self.fqn) |
| if signature is not None and self.parent is not None: |
| assert len(signature.in_spec.children_specs) == 2 |
| args_spec = signature.in_spec.children_specs[0] |
| kwargs_spec = signature.in_spec.children_specs[1] |
| assert args_spec.context is None |
| assert kwargs_spec.context is not None |
| |
| with self.graph_module.graph.inserting_after(None): |
| arg_nodes = [] |
| for idx in range(len(args_spec.children_specs)): |
| arg_nodes.append( |
| self.graph_module.graph.placeholder(f"_positional_arg_{idx}") |
| ) |
| kwarg_nodes = {} |
| for name in kwargs_spec.context: |
| kwarg_nodes[name] = self.graph_module.graph.placeholder(name) |
| flat_args = _generate_flatten( |
| self.graph_module, |
| (tuple(arg_nodes), kwarg_nodes), |
| signature.in_spec, |
| ) |
| for idx, arg in enumerate(signature.inputs): |
| flat_arg_node = self.graph_module.graph.create_node( |
| op="call_function", |
| target=operator.getitem, |
| args=(flat_args, idx), |
| name=arg.name |
| if not isinstance(arg, ConstantArgument) |
| else f"_constant_{idx}", |
| ) |
| if isinstance(arg, ConstantArgument): |
| continue |
| flat_arg_node.meta = copy.copy(self.seen_nodes[arg.name].meta) |
| self.node_to_placeholder[self.seen_nodes[arg.name]] = flat_arg_node |
| |
| with self.parent.graph.inserting_before(self.parent_call_module): |
| nodes: List[Optional[torch.fx.Node]] = [] |
| for input in signature.inputs: |
| if isinstance(input, ConstantArgument) and input.value is None: |
| nodes.append(None) |
| else: |
| assert isinstance(input, (TensorArgument, SymIntArgument)) |
| nodes.append( |
| self.parent.remap_input(self.seen_nodes[input.name]) |
| ) |
| |
| inputs_node = _generate_unflatten( |
| self.parent.graph_module, |
| nodes, |
| signature.in_spec, |
| ) |
| |
| args_node = self.parent.graph.call_function( |
| operator.getitem, (inputs_node, 0) |
| ) |
| kwargs_node = self.parent.graph.call_function( |
| operator.getitem, (inputs_node, 1) |
| ) |
| arg_nodes = [ |
| self.parent.graph.call_function(operator.getitem, (args_node, i)) |
| for i in range(len(args_spec.children_specs)) |
| ] |
| kwarg_nodes = { |
| k: self.parent.graph.call_function( |
| operator.getitem, (kwargs_node, k) |
| ) |
| for k in kwargs_spec.context |
| } |
| assert self.parent_call_module is not None |
| self.parent_call_module.args = tuple(arg_nodes) |
| self.parent_call_module.kwargs = kwarg_nodes |
| |
| def add_placeholder(self, x): |
| assert x.graph is self.flat_graph |
| # x is not in subgraph, create a new placeholder for subgraph |
| with self.graph.inserting_before(None): |
| placeholder_node = self.graph.placeholder(x.name, type_expr=x.type) |
| # copy all meta fields, even if some fields might be irrelvant for |
| # the placeholder node |
| placeholder_node.meta = copy.copy(x.meta) |
| self.node_to_placeholder[x] = placeholder_node |
| |
| def remap_input(self, x): |
| assert x.graph is self.flat_graph |
| if x in self.node_map: |
| return self.node_map[x] |
| if x not in self.node_to_placeholder: |
| self.add_placeholder(x) |
| if self.parent_call_module is not None: |
| # Important to *prepend* the output to match how we are |
| # inserting placeholder nodes. |
| self.parent_call_module.insert_arg(0, self.parent.remap_input(x)) |
| return self.node_to_placeholder[x] |
| |
| def finalize_outputs(self): |
| orig_outputs = [] |
| |
| signature = self.module_call_graph.get(self.fqn) |
| if signature is not None and self.parent is not None: |
| for output in signature.outputs: |
| if isinstance(output, (TensorArgument, SymIntArgument)): |
| orig_outputs.append(self.seen_nodes[output.name]) |
| else: |
| raise RuntimeError( |
| f"Unsupported data type for output node: {output}" |
| ) |
| |
| tree_out_node = _generate_unflatten( |
| self.graph_module, |
| tuple( |
| self.node_map[self.seen_nodes[output.name]] |
| for output in orig_outputs |
| ), |
| signature.out_spec, |
| ) |
| parent_out: Optional[torch.fx.Node] = _generate_flatten( |
| self.parent.graph_module, self.parent_call_module, signature.out_spec |
| ) |
| graph_outputs: Union[torch.fx.Node, List[torch.fx.Node]] = tree_out_node |
| else: |
| graph_outputs = [] |
| # Iterate through nodes we have copied into self.graph. |
| for orig_node in self.node_map.keys(): |
| for user_node in orig_node.users: |
| if user_node.name not in self.seen_nodes: |
| # external user node, need to expose as an output |
| orig_outputs.append(orig_node) |
| graph_outputs.append(self.node_map[orig_node]) |
| break |
| |
| parent_out = self.parent_call_module |
| if len(graph_outputs) == 1: |
| graph_outputs = graph_outputs[0] |
| |
| assert isinstance(graph_outputs, (list, torch.fx.Node)) |
| |
| self.graph.output(graph_outputs) |
| |
| # lint to ensure correctness |
| self.graph.lint() |
| self.graph_module.recompile() |
| |
| # Rewrite outputs in parent module |
| if parent_out is None: |
| return |
| |
| if len(orig_outputs) == 1 and signature is None: |
| self.parent.node_map[orig_outputs[0]] = parent_out |
| else: |
| for i, orig_output in enumerate(orig_outputs): |
| # Use Proxy to record getitem access. |
| proxy_out = torch.fx.Proxy(parent_out)[i].node # type: ignore[index] |
| self.parent.node_map[orig_output] = proxy_out |
| |
| if self.cached_graph_module is not None: |
| _verify_graph_equivalence(self.cached_graph_module, self.graph_module) |
| |
| def copy_node(self, node): |
| self.print("copying", node.format_node()) |
| self.node_map[node] = self.graph.node_copy(node, self.remap_input) |
| self.seen_nodes[node.name] = node |
| |
| def run_outer(self): |
| i = 0 |
| for node in self.flat_graph.nodes: |
| self.print(i, node.meta.get("nn_module_stack"), node.format_node()) |
| i += 1 |
| |
| # Copy all graph inputs |
| node_idx: int = 0 |
| node = self.nodes[node_idx] |
| while node.op == "placeholder": |
| self.copy_node(node) |
| node_idx += 1 |
| node = self.nodes[node_idx] |
| |
| self.run_from(node_idx) |
| |
| # Copy graph outputs |
| for node in self.flat_graph.nodes: |
| if node.op == "output": |
| self.copy_node(node) |
| |
| def print(self, *args, **kwargs): |
| if self.verbose: |
| print(*args, **kwargs) |
| |
| def run_from(self, node_idx): |
| module_idx = 0 |
| # Walk through the graph, building up a new graph with the right submodules |
| while node_idx < len(self.nodes): |
| node = self.nodes[node_idx] |
| assert node.op != "placeholder" |
| |
| self.print() |
| self.print("STEP", node_idx, node.format_node()) |
| self.print(self.module_stack) |
| if node.op == "output": |
| if len(self.module_stack) == 1: |
| # We want the output node of the original graph to be handled |
| # specially by the outermost stack frame (in run_outer). So |
| # skip finalization here. |
| return node_idx |
| |
| # We've reached the end of the graph. Wrap up all the existing stack frames. |
| self.finalize_outputs() |
| return node_idx |
| |
| node_module_stack = ( |
| [path for path, ty in node.meta["nn_module_stack"].values()] |
| if "nn_module_stack" in node.meta |
| else self.module_stack |
| ) |
| if node_module_stack[: len(self.module_stack)] != self.module_stack: |
| # This means that the current module is done executing and the |
| # current node is the beginning of a new module. |
| # |
| # In this case, we should finalize this module and return without |
| # incrementing the node counter. |
| self.finalize_outputs() |
| self.print("outlining", self.fqn) |
| self.print(self.graph) |
| return node_idx |
| |
| assert node_module_stack is not None |
| |
| if is_prefix(self.module_stack, node_module_stack): |
| # This means that the current node represents the execution of a new |
| # module. |
| next_module = node_module_stack[len(self.module_stack)] |
| self.print("Creating new stack frame for", next_module) |
| # Run a nested version of module outliner from the current node |
| # counter. Once it is complete, continue from that point. |
| node_idx = ModuleFrame( |
| self.flat_graph, |
| self.seen_nodes, |
| self.seen_modules, |
| self, |
| self.module_stack + [next_module], |
| list(node.meta["nn_module_stack"].keys())[len(self.module_stack)], |
| self.module_call_graph, |
| ).run_from(node_idx) |
| module_idx += 1 |
| continue |
| |
| # The only remaining possibility is that we are in the right stack |
| # frame. Copy the node into this frame's graph and increment the node counter. |
| assert node_module_stack == self.module_stack |
| self.copy_node(node) |
| node_idx += 1 |
| |
| |
| def _outline_submodules(orig_graph: torch.fx.Graph, root_module: torch.fx.GraphModule): |
| seen_nodes: Dict[str, torch.fx.Node] = {} |
| seen_modules: Dict[int, torch.nn.Module] = {} |
| ModuleFrame( |
| orig_graph, |
| seen_nodes, |
| seen_modules, |
| None, |
| [""], |
| "", |
| { |
| entry.fqn: entry.signature |
| for entry in root_module.module_call_graph |
| if entry.signature |
| }, |
| graph_module=root_module, |
| ).run_outer() |
| |
| |
| def _sink_params( |
| module: GraphModule, |
| inputs_to_state: Dict[str, str], |
| scope: List[str], |
| ): |
| """Sink params and buffers from graph inputs into get_attr nodes. |
| |
| Exported modules are purely functional, so they pass their parameters and |
| buffers in as inputs to the graph. |
| |
| To replicate eager's semantics, we need to get them from the module state |
| via get_attr instead. |
| |
| module: GraphModule, potentially containining nested submodules. |
| inputs_to_state: mapping graph input names to the corresponding key in the state_dict. |
| scope: tracks where we are in the module hierarchy, so that we can emit the |
| right `getattr(self, "foo.bar")` calls, etc. |
| """ |
| for name, submodule in module._modules.items(): |
| _sink_params(cast(GraphModule, submodule), inputs_to_state, scope + [name]) |
| |
| if not isinstance(module, GraphModule): |
| return |
| |
| graph = module.graph |
| inputs = filter(lambda n: n.op == "placeholder", graph.nodes) |
| |
| # Also remove from call_module nodes |
| call_module_nodes = filter(lambda n: n.op == "call_module", graph.nodes) |
| for node in call_module_nodes: |
| node.args = tuple(filter(lambda n: n.name not in inputs_to_state, node.args)) |
| |
| for node in inputs: |
| if node.name not in inputs_to_state: |
| continue |
| |
| if len(node.users) > 0: |
| state_name = inputs_to_state[node.name].split(".") |
| # If there's a mismatch beteewn scope name and state name, then there must be multuple scopes |
| # pointing to the same state name, meaning some modules are shared. In such case, we can simply |
| # skip updating the current node because another later iteration will take care of this input |
| # node when the unique match between scope and state name occurs. |
| # To make sure this always happen, we should enforce the invariant that no placeholder node |
| # in the unflattened graph appears in inputs_to_state dict, which means all the extra input |
| # nodes have been handled. |
| if state_name[: len(scope)] != scope: |
| continue |
| attr_path = state_name[len(scope) :] |
| state_attr = _recursive_getattr(module, attr_path) |
| assert isinstance(state_attr, torch.Tensor) |
| |
| with graph.inserting_after(node): |
| new_node = graph.create_node("get_attr", ".".join(attr_path)) |
| |
| node.replace_all_uses_with(new_node, propagate_meta=True) |
| graph.erase_node(node) |
| module.recompile() |
| |
| |
| def _recursive_getattr(obj, attr_path): |
| for attr in attr_path: |
| obj = getattr(obj, attr) |
| |
| return obj |
