import colossalai import torch import copy from typing import List, Callable, Any, Tuple, Dict, Iterable from torch.fx.node import Node, Argument, map_arg, _type_repr, _get_qualified_name from torch.fx.graph import _Namespace, PythonCode, _custom_builtins, _is_from_torch, _format_target, magic_methods, CodeGen, _origin_type_map, inplace_methods, _CustomBuiltin from colossalai.fx.profiler import calculate_fwd_out, calculate_fwd_tmp, parameter_size, activation_size CODEGEN_AVAILABLE = True __all__ = ['ChunkCodeGen'] class NodeIndexTracer(object): def __init__(self, gm) -> None: self.gm = gm self.nodes_list = list(gm.graph.nodes) self.idx_trace_list = [{'idx': [], 'compute': []} for _ in range(len(self.nodes_list))] self.idx_trace_equal = [] self.idx_view_list = [] self.idx_count = -1 def add_index(self): """ Update the count and return it. To record the idx number. Returns: idx_count: int """ self.idx_count += 1 return self.idx_count def inherit_computation(self, node_from, node_to): """ Inherit computed dim from node_from to node_to. If a dim in node_from is marked as computed and exists in node_to, still mark it as computed in node_to. Args: node_from (node): node to be inherited node_to (node): new node to inherit """ _, compute_from = self.find_trace_from_node(node_from) idx_to, compute_to = self.find_trace_from_node(node_to) for i in compute_from: if i in idx_to and i not in compute_to: compute_to.append(i) def mark_idx_equal(self, idx1, idx2): """ Mark 2 index to be equal. Args: idx1 (int): index count. idx2 (int): index count. """ self.idx_trace_equal.append((idx1, idx2)) def mark_computation(self, node, idx, dim): """ Mark some dims of node as computed. Args: node (node) idx (int): node index dim (list or int): dims to be marked as computed """ input_node_idx_trace = self.find_idx_trace_from_node(node) if isinstance(dim, int): dim = [dim] for d in dim: cur_idx = input_node_idx_trace[d] if cur_idx not in self.idx_trace_list[idx]['compute']: self.idx_trace_list[idx]['compute'].append(cur_idx) def find_trace_from_node(self, node): """ Find node idx and compute trace by the node. Args: node (node) Returns: idx (list): idx of the node compute (list): computed idx of the node. """ node_idx = _find_idx_by_name(node.name, self.nodes_list) node_dict = self.idx_trace_list[node_idx] return node_dict['idx'], node_dict['compute'] def find_idx_trace_from_node(self, node): """ Find node idx trace by the node. Args: node (node) Returns: idx (list): idx of the node """ node_idx = _find_idx_by_name(node.name, self.nodes_list) return self.idx_trace_list[node_idx]['idx'] def find_compute_trace_from_node(self, node): """ Find node compute trace by the node. Args: node (node) Returns: compute (list): computed idx of the node. """ node_idx = _find_idx_by_name(node.name, self.nodes_list) return self.idx_trace_list[node_idx]['compute'] def assign_index_as_input(self, node, node_idx): """ Assign node's trace as its input node. Args: node (node) node_idx (int) """ input_node_idx = _find_idx_by_name(node.args[0].name, self.nodes_list) input_node_idx_trace = self.idx_trace_list[input_node_idx]['idx'] new_idx_trace = copy.deepcopy(input_node_idx_trace) self.idx_trace_list[node_idx]['idx'] = new_idx_trace def assign_all_index(self, node, node_idx): """ Add new index for all node's dims. Args: node (node) node_idx (int) """ shape = node.meta['tensor_meta'].shape new_trace = [] for _ in shape: new_trace.append(self.add_index()) self.idx_trace_list[node_idx]['idx'] = new_trace def assign_transpose_index(self, node, node_idx): """ Assign index for transpose op. 1. swap input's dim according to transpose args 2. inherit input's computation Args: node (node) node_idx (int) """ tranpose_dim = node.args[1:] input_node_idx_trace = self.find_idx_trace_from_node(node.args[0]) new_idx_trace = copy.deepcopy(input_node_idx_trace) new_idx_trace[tranpose_dim[0]] = input_node_idx_trace[tranpose_dim[1]] new_idx_trace[tranpose_dim[1]] = input_node_idx_trace[tranpose_dim[0]] self.idx_trace_list[node_idx]['idx'] = new_idx_trace self.inherit_computation(node.args[0], node) def assign_permute_index(self, node, node_idx): """ Assign index for permute op. 1. swap input's dim according to permute args 2. inherit input's computation Args: node (node) node_idx (int) """ permute_dim = node.args[1:] input_node_idx_trace = self.find_idx_trace_from_node(node.args[0]) new_idx_trace = copy.deepcopy(input_node_idx_trace) for idx, d in enumerate(permute_dim): new_idx_trace[idx] = input_node_idx_trace[d] self.idx_trace_list[node_idx]['idx'] = new_idx_trace self.inherit_computation(node.args[0], node) def assign_linear_index(self, node, node_idx): """ Assign index for linear op. 1. copy trace from input node and change last index accroding to weight 2. mark equal for input node last index, weight first dim and bias dim. 3. inherit input's computation, mark computation for last dim. Args: node (node) node_idx (int) """ input_node, weight, bias = node.args input_node_idx_trace = self.find_idx_trace_from_node(input_node) weight_idx_trace = self.find_idx_trace_from_node(weight) new_idx_trace = copy.deepcopy(input_node_idx_trace) new_idx_trace[-1] = weight_idx_trace[1] self.idx_trace_list[node_idx]['idx'] = new_idx_trace self.inherit_computation(input_node, node) self.mark_computation(node, node_idx, [-1]) self.mark_idx_equal(input_node_idx_trace[-1], weight_idx_trace[0]) if bias: bias_idx_trace = self.find_idx_trace_from_node(bias) self.mark_idx_equal(input_node_idx_trace[-1], bias_idx_trace[0]) def assign_matmul_index(self, node, node_idx): """ Assign index for matmul op. 1. copy trace from matmul_left and change last index accroding to matmul_right. (assert they have same length) 2. mark equal for input matmul_left -1 index and matmul_right -2 dim. 3. inherit matmul_left and matmul_right computation, mark computation for last dim. Args: node (node) node_idx (int) """ matmul_left, matmul_right = node.args matmul_left_idx_trace = self.find_idx_trace_from_node(matmul_left) matmul_right_idx_trace = self.find_idx_trace_from_node(matmul_right) assert(len(matmul_left_idx_trace) == len(matmul_right_idx_trace)) new_idx_trace = copy.deepcopy(matmul_left_idx_trace) new_idx_trace[-1] = matmul_right_idx_trace[-1] self.idx_trace_list[node_idx]['idx'] = new_idx_trace self.inherit_computation(matmul_left, node) self.inherit_computation(matmul_right, node) self.mark_computation(node, node_idx, [-1]) self.mark_idx_equal(matmul_left_idx_trace[-1], matmul_right_idx_trace[-2]) def assign_layernorm_index(self, node, idx): """ Assign index for layernorm op. 1. assign index as input node 2. inherit computation and mark last 2 dims as computed. Args: node (node) node_idx (int) """ self.assign_index_as_input(node, idx) self.inherit_computation(node.args[0], node) self.mark_computation(node, idx, [-1, -2]) def assign_elementwise_index(self, node, idx): """ Assign index for element-wise op (eg. relu sigmoid add mul). 1. assign index as input node 2. inherit computation from all input nodes. Args: node (node) node_idx (int) """ self.assign_index_as_input(node, idx) for node_in in node.args: if type(node_in) not in (int, float): self.inherit_computation(node_in, node) def assign_softmax_index(self, node, idx): """ Assign index for softmax op. 1. assign index as input node 2. inherit computation and mark softmax dim as computed. Args: node (node) node_idx (int) """ self.assign_index_as_input(node, idx) self.inherit_computation(node.args[0], node) self.mark_computation(node, idx, [node.kwargs['dim']]) def assign_view_reshape_index(self, node, node_idx): """ Assign index for view and reshape op. 1. get origin shape and target shape by meta info. 2. compute the real value of -1 in target shape. 3. determine changed dim, and assgin index for generated dim. 4. log changed dim and generated dim for restore 5. inherit computation. 6. TODO: look into view list to see whether the view is associated with other, if so assgin equal dim according to previous view. Args: node (node) node_idx (int) """ # get data, turn into number origin_node = node.args[0] origin_shape = origin_node.meta['tensor_meta'].shape target_shape = [] for i in range(1, len(node.args)): if isinstance(node.args[i], int): target_shape.append(node.args[i]) else: target_shape.append(node.args[i].meta['fwd_out'][0]) # compute the value of -1 if -1 in target_shape: origin_product = 1 for i in origin_shape: origin_product *= i target_product = -1 for i in target_shape: target_product *= i shape_idx = target_shape.index(-1) target_shape[shape_idx] = origin_product // target_product # determine changed dim len_diff = len(origin_shape) - len(target_shape) if len_diff == 1: # dim merge dim_equal = [i == j for i, j in zip(origin_shape[:-1], target_shape)] dim_to = [dim_equal.index(False)] dim_from = [dim_equal.index(False), dim_equal.index(False) + 1] elif len_diff == -1: # dim expand dim_equal = [i == j for i, j in zip(origin_shape, target_shape[:-1])] dim_from = [dim_equal.index(False)] dim_to = [dim_equal.index(False), dim_equal.index(False) + 1] else: raise NotImplementedError("shape" + str(origin_shape) + 'and' + str(target_shape) + "view not implemented") # get new index origin_trace = self.find_idx_trace_from_node(origin_node) new_trace = copy.deepcopy(origin_trace) dim_from.reverse() for i in dim_from: new_trace.pop(i) for i in dim_to: new_trace.insert(i, self.add_index()) self.idx_trace_list[node_idx]['idx'] = new_trace # inherit computation self.inherit_computation(origin_node, node) compute_log = self.find_compute_trace_from_node(origin_node) for i in dim_from: if origin_trace[i] in compute_log: for j in dim_to: self.mark_computation(node, node_idx, [j]) break # log view, not used now view_dict = {"idx_from": [origin_trace[i] for i in dim_from], "dim_from": dim_from, "idx_to": [new_trace[i] for i in dim_to], "dim_to": dim_to} self.idx_view_list.append(view_dict) def trace_node_idx(self): for idx, node in enumerate(self.nodes_list): if node.op == 'placeholder': self.assign_all_index(node, idx) elif node.op == 'call_method': if 'transpose' in node.name: self.assign_transpose_index(node, idx) elif 'permute' in node.name: self.assign_permute_index(node, idx) elif 'view' in node.name or 'reshape' in node.name: self.assign_view_reshape_index(node, idx) else: raise NotImplementedError(node.name, "method not implemented yet!") elif node.op == 'call_function': if 'linear' in node.name: self.assign_linear_index(node, idx) elif 'matmul' in node.name: self.assign_matmul_index(node, idx) elif 'softmax' in node.name: self.assign_softmax_index(node, idx) elif any(n in node.name for n in ['mul', 'add', 'sigmoid', 'relu']): self.assign_elementwise_index(node, idx) elif 'getattr' in node.name: continue # get attr like shape elif 'getitem' in node.name: continue # get item in list else: raise NotImplementedError(node.name, "function not implemented yet!") elif node.op == 'call_module': if any(n in node.name for n in ['layernorm', 'norm']): self.assign_layernorm_index(node, idx) else: raise NotImplementedError(node.name, "module not implemented yet!") elif node.op == 'get_attr': self.assign_all_index(node, idx) # get param elif node.op == 'output': continue else: raise NotImplementedError(node.op, "op not implemented yet!") def _get_meta_node_size(x): x = x.meta['tensor_meta'] x = x.numel * torch.tensor([], dtype=x.dtype).element_size() return x def _get_output_node_size(n): fwd_out = {x.uuid: x for x in n.meta["fwd_out"] if isinstance(x, torch.Tensor) and hasattr(x, 'uuid')} return activation_size(fwd_out) def _get_delete_node_size(user, user_to_last_uses): if user.op in ('placeholder', 'output'): return 0 nodes_to_delete = user_to_last_uses.get(user, []) if len(nodes_to_delete): delete_size = sum([_get_output_node_size(i) for i in nodes_to_delete]) return delete_size return 0 def _get_last_usr(nodes): node_to_last_use: Dict[Node, Node] = {} user_to_last_uses: Dict[Node, List[Node]] = {} def register_last_uses(n: Node, user: Node): if n not in node_to_last_use: node_to_last_use[n] = user user_to_last_uses.setdefault(user, []).append(n) for node in reversed(nodes): map_arg(node.args, lambda n: register_last_uses(n, node)) map_arg(node.kwargs, lambda n: register_last_uses(n, node)) return user_to_last_uses def _delete_free_var_from_last_use(user_to_last_uses): for key, value in user_to_last_uses.items(): for n in value: if n.op == 'placeholder': user_to_last_uses[key].remove(n) def _get_contiguous_memory(node, not_contiguous_list, delete=False): mem = 0 not_contiguous_ops = ['transpose', 'permute'] if node.op == 'call_function' and 'matmul' in node.name: for n in node.args: if n in not_contiguous_list: # matmul won't change origin tensor, but create a tmp copy mem += _get_output_node_size(n) elif node.op == 'call_module': for n in node.args: if n in not_contiguous_list: # module will just make origin tensor to contiguous if delete: not_contiguous_list.remove(n) elif node.op == 'call_method' and any(i in node.name for i in not_contiguous_ops): if node not in not_contiguous_list: not_contiguous_list.append(node) elif any(i in node.args for i in not_contiguous_list): if node not in not_contiguous_list: not_contiguous_list.append(node) return mem def _estimate_inference_mem(gm: torch.fx.GraphModule): act_memory = 0.0 act_memory_peak_log = [] act_memory_after_node_log = [] not_contiguous_list = [] user_to_last_uses = _get_last_usr(list(gm.graph.nodes)) _delete_free_var_from_last_use(user_to_last_uses) for node in gm.graph.nodes: # if node is placeholder, just add the size of the node if node.op == 'placeholder': act_memory += _get_meta_node_size(node) / (1024 ** 2) act_memory_peak_log.append(act_memory) act_memory_after_node_log.append(act_memory) # skip output elif node.op == 'output': continue # node is an operation, calculate tmp, output node and delete node memory else: # forward memory act_memory += _get_contiguous_memory(node, not_contiguous_list) / (1024 ** 2) act_memory += _get_output_node_size(node) / (1024 ** 2) # record max act memory act_memory_peak_log.append(act_memory) # delete useless memory act_memory -= _get_delete_node_size(node, user_to_last_uses) / (1024 ** 2) act_memory -= _get_contiguous_memory(node, not_contiguous_list, delete=True) / (1024 ** 2) act_memory_after_node_log.append(act_memory) print("no chunk") _print_mem_log(act_memory_peak_log, list(gm.graph.nodes), "peak") _print_mem_log(act_memory_after_node_log, list(gm.graph.nodes), "after") param_memory = parameter_size(gm) return act_memory + param_memory, param_memory def _get_chunk_ratio(node, chunk_dim, chunk_size): shape = node.meta['tensor_meta'].shape chunk_ratio = float(chunk_size) / shape[chunk_dim] return chunk_ratio def _get_chunk_delete_node_size(user, user_to_last_uses, chunk_ratio, node_list, start_node, end_node): if user.op in ('placeholder', 'output'): return 0 nodes_to_delete = user_to_last_uses.get(user, []) delete_size = 0 for n in nodes_to_delete: node_idx = _find_idx_by_name(n.name, node_list) if start_node <= node_idx < end_node: delete_size += _get_output_node_size(n) * chunk_ratio return delete_size def _print_mem_log(log, nodes, title=None): if title: print(title) for idx, (l, n) in enumerate(zip(log, nodes)): print("%s:%.2f \t" % (n.name, l), end='') if (idx + 1) % 3 == 0: print("") print("\n") def _estimate_chunk_inference_mem(gm: torch.fx.GraphModule, start_nodes, end_nodes, chunk_dims, chunk_sizes): act_memory = 0.0 act_memory_peak_log = [] act_memory_after_node_log = [] not_contiguous_list = [] user_to_last_uses = _get_last_usr(list(gm.graph.nodes)) _delete_free_var_from_last_use(user_to_last_uses) within_chunk = False region_idx = 0 chunk_ratio = 1 # use it to estimate chunk mem node_list = list(gm.graph.nodes) for idx, node in enumerate(node_list): # if node in chunk start nodes, change chunk ratio and add chunk_tensor if idx in start_nodes: within_chunk = True chunk_ratio = _get_chunk_ratio(node, chunk_dims[region_idx], chunk_sizes[region_idx]) act_memory += _get_output_node_size(node_list[end_nodes[region_idx]]) / (1024 ** 2) # if node is placeholder, just add the size of the node if node.op == 'placeholder': act_memory += _get_meta_node_size(node) * chunk_ratio / (1024 ** 2) act_memory_peak_log.append(act_memory) # skip output elif node.op == 'output': continue # node is an operation, calculate tmp, output node and delete node memory else: # forward memory # TODO: permute will create a tmp copy if not contiguous act_memory += _get_contiguous_memory(node, not_contiguous_list) * chunk_ratio / (1024 ** 2) act_memory += _get_output_node_size(node) * chunk_ratio / (1024 ** 2) # record max act memory act_memory_peak_log.append(act_memory) # delete useless memory act_memory -= _get_contiguous_memory(node, not_contiguous_list, delete=True) * chunk_ratio / (1024 ** 2) if within_chunk: act_memory -= _get_chunk_delete_node_size( node, user_to_last_uses, chunk_ratio, node_list, start_nodes[region_idx], end_nodes[region_idx]) / (1024 ** 2) else: act_memory -= _get_delete_node_size(node, user_to_last_uses) / (1024 ** 2) if idx in end_nodes: act_memory -= _get_output_node_size(node) * chunk_ratio / (1024 ** 2) within_chunk = False chunk_ratio = 1 region_idx += 1 act_memory_after_node_log.append(act_memory) print("chunk") _print_mem_log(act_memory_peak_log, node_list, "peak") _print_mem_log(act_memory_after_node_log, node_list, "after") param_memory = parameter_size(gm) return act_memory + param_memory, param_memory def _gen_chunk_slice_dim(chunk_dim, chunk_idx_name, shape): new_shape = "[" for idx, i in enumerate(shape): if idx == chunk_dim: new_shape += "%s:%s + chunk_size" % (chunk_idx_name, chunk_idx_name) else: new_shape += ":" new_shape += ", " new_shape = new_shape[:-2] + "]" return new_shape def _get_first_non_single_dim(shape): for idx, i in enumerate(shape): if i == 1: continue else: return idx raise RuntimeError("can not get first non single dim for shape", shape) def _gen_loop_start(chunk_input_meta, chunk_output, chunk_size=2): if len(chunk_input_meta) == 1: node = chunk_input_meta[0] node_shape = node.meta['tensor_meta'].shape chunk_dim = _get_first_non_single_dim(node_shape) chunk_slice = _gen_chunk_slice_dim(chunk_dim, "gen_chunk_idx", node_shape) out_shape = str(list(chunk_output.meta['tensor_meta'].shape)) context = "chunk_result = torch.empty(%s, dtype=%s.dtype, device=%s.device); chunk_size = %d\nfor gen_chunk_idx in range" % ( out_shape, node.name, node.name, chunk_size) context += "(0, %s.shape[%d], chunk_size):\n" % (node.name, chunk_dim) context += " chunk_tensor = %s%s\n" % (node.name, chunk_slice) else: raise NotImplementedError("input with size %d not implemented" % len(chunk_input_meta)) return context def _gen_loop_end(chunk_outputs, chunk_inputs, node_list): chunk_inputs_name = chunk_inputs[0].name chunk_outputs_name = chunk_outputs.name chunk_outputs_idx = _find_idx_by_name(chunk_outputs_name, node_list) chunk_output_shape = chunk_outputs.meta['tensor_meta'].shape chunk_dim = _get_first_non_single_dim(chunk_output_shape) chunk_slice = _gen_chunk_slice_dim(chunk_dim, "gen_chunk_idx", chunk_output_shape) context = " chunk_result%s = %s\n" % (chunk_slice, chunk_outputs_name) context += chunk_outputs_name + " = chunk_result; chunk_result = None; chunk_size = None" # determine if its the last use for chunk input users_name = list(chunk_inputs[0].users.keys()) if all([_find_idx_by_name(user.name, node_list) <= chunk_outputs_idx for user in users_name]): context += "; %s = None" % chunk_inputs_name context += "\n" return context def _find_input_and_output_nodes(nodes: List[Node]): """ Find the input and output node names which are not found in the given list of nodes. """ input_nodes = [] output_nodes = [] # if a node has an input node which is not in the node list # we treat that input node as the input of the checkpoint function for node in nodes: for input_node in node._input_nodes.keys(): node_repr = repr(input_node) if input_node not in nodes and node_repr not in input_nodes: input_nodes.append(input_node) # if a node has a user node which is not in the node list # we treat that user node as the node receiving the current node output for node in nodes: for output_node in node.users.keys(): node_repr = repr(node) if output_node not in nodes and node_repr not in output_nodes: output_nodes.append(output_node) return input_nodes, output_nodes def _find_idx_by_name(name, nodes_list): for idx, node in enumerate(nodes_list): if node.name == name: return idx raise RuntimeError("name %s not found in node list" % name) def emit_code_with_chunk(body, ckpt_func, nodes, emit_node_func, delete_unused_value_func, meta_nodes, meta_graph): """Emit code with nested activation checkpoint When we detect some of the node.activation_checkpoint is a List, we will use this function to emit the activation checkpoint codes. Args: body: forward code ckpt_func: checkpoint functions code nodes: graph.nodes emit_node_func: function to emit node delete_unused_value_func: function to remove the unused value """ # find the offload regions chunk_regions = [(58, 62)] chunk_starts = [item[0] for item in chunk_regions] chunk_ends = [item[1] for item in chunk_regions] chunk_inputs = [] chunk_outputs = [] within_chunk_region = False node_list = list(nodes) _estimate_chunk_inference_mem(meta_graph, chunk_starts, chunk_ends, [1], [2]) _estimate_inference_mem(meta_graph) node_index_tracer = NodeIndexTracer(meta_graph) node_index_tracer.trace_node_idx() # find the input and output var names for each offload region for idx, (start, end) in enumerate(chunk_regions): offload_node_list = node_list[start:end + 1] inputs, outputs = _find_input_and_output_nodes(offload_node_list) chunk_inputs.append(inputs) chunk_outputs.append(outputs) chunk_inputs_idx = [[_find_idx_by_name(j.name, node_list) for j in i] for i in chunk_inputs] chunk_outputs_idx = [[_find_idx_by_name(j.name, node_list) for j in i] for i in chunk_outputs] chunk_inputs_names = [] for i in chunk_inputs: for j in i: chunk_inputs_names.append(j.name) # this flag is to prevent repeated insert of save tensors # hooks definition in ckpt_func node_idx = 0 region_idx = 0 while node_idx < len(node_list): node = node_list[node_idx] if node_idx in chunk_starts: within_chunk_region = True # add for loop chunk_input_meta = [meta_nodes[i] for i in chunk_inputs_idx[region_idx]] body.append(_gen_loop_start(chunk_input_meta, node_list[chunk_ends[region_idx]])) if within_chunk_region: emit_node_func(node, body) # replace input var with chunk var if node_idx in chunk_starts: body[-1] = body[-1].replace("("+ chunk_inputs[region_idx][0].name +")", '(chunk_tensor)') body[-1] = ' ' + body[-1] delete_unused_value_func(node, body, chunk_inputs_names) else: emit_node_func(node, body) if node_idx not in chunk_inputs: delete_unused_value_func(node, body, chunk_inputs_names) if node_idx in chunk_ends: body.append(_gen_loop_end(node, chunk_inputs[region_idx], node_list)) within_chunk_region = False region_idx += 1 node_idx += 1 if CODEGEN_AVAILABLE: class ChunkCodeGen(CodeGen): def __init__(self, meta_graph): super().__init__() self.meta_graph = meta_graph self.meta_node = list(meta_graph.graph.nodes) def _gen_python_code(self, nodes, root_module: str, namespace: _Namespace) -> PythonCode: free_vars: List[str] = [] body: List[str] = [] globals_: Dict[str, Any] = {} wrapped_fns: Dict[str, None] = {} # Wrap string in list to pass by reference maybe_return_annotation: List[str] = [''] def add_global(name_hint: str, obj: Any): """Add an obj to be tracked as a global. We call this for names that reference objects external to the Graph, like functions or types. Returns: the global name that should be used to reference 'obj' in generated source. """ if _is_from_torch(obj) and obj != torch.device: # to support registering torch.device # HACK: workaround for how torch custom ops are registered. We # can't import them like normal modules so they must retain their # fully qualified name. return _get_qualified_name(obj) # normalize the name hint to get a proper identifier global_name = namespace.create_name(name_hint, obj) if global_name in globals_: assert globals_[global_name] is obj return global_name globals_[global_name] = obj return global_name # set _custom_builtins here so that we needn't import colossalai in forward _custom_builtins["colossalai"] = _CustomBuiltin("import colossalai", colossalai) # Pre-fill the globals table with registered builtins. for name, (_, obj) in _custom_builtins.items(): add_global(name, obj) def type_repr(o: Any): if o == (): # Empty tuple is used for empty tuple type annotation Tuple[()] return '()' typename = _type_repr(o) if hasattr(o, '__origin__'): # This is a generic type, e.g. typing.List[torch.Tensor] origin_type = _origin_type_map.get(o.__origin__, o.__origin__) origin_typename = add_global(_type_repr(origin_type), origin_type) if hasattr(o, '__args__'): # Assign global names for each of the inner type variables. args = [type_repr(arg) for arg in o.__args__] if len(args) == 0: # Bare type, such as `typing.Tuple` with no subscript # This code-path used in Python < 3.9 return origin_typename return f'{origin_typename}[{",".join(args)}]' else: # Bare type, such as `typing.Tuple` with no subscript # This code-path used in Python 3.9+ return origin_typename # Common case: this is a regular module name like 'foo.bar.baz' return add_global(typename, o) def _format_args(args: Tuple[Argument, ...], kwargs: Dict[str, Argument]) -> str: def _get_repr(arg): # Handle NamedTuples (if it has `_fields`) via add_global. if isinstance(arg, tuple) and hasattr(arg, '_fields'): qualified_name = _get_qualified_name(type(arg)) global_name = add_global(qualified_name, type(arg)) return f"{global_name}{repr(tuple(arg))}" return repr(arg) args_s = ', '.join(_get_repr(a) for a in args) kwargs_s = ', '.join(f'{k} = {_get_repr(v)}' for k, v in kwargs.items()) if args_s and kwargs_s: return f'{args_s}, {kwargs_s}' return args_s or kwargs_s # Run through reverse nodes and record the first instance of a use # of a given node. This represents the *last* use of the node in the # execution order of the program, which we will use to free unused # values node_to_last_use: Dict[Node, Node] = {} user_to_last_uses: Dict[Node, List[Node]] = {} def register_last_uses(n: Node, user: Node): if n not in node_to_last_use: node_to_last_use[n] = user user_to_last_uses.setdefault(user, []).append(n) for node in reversed(nodes): map_arg(node.args, lambda n: register_last_uses(n, node)) map_arg(node.kwargs, lambda n: register_last_uses(n, node)) _delete_free_var_from_last_use(user_to_last_uses) # NOTE: we add a variable to distinguish body and ckpt_func def delete_unused_values(user: Node, body, to_keep=[]): """ Delete values after their last use. This ensures that values that are not used in the remainder of the code are freed and the memory usage of the code is optimal. """ if user.op == 'placeholder': return if user.op == 'output': body.append('\n') return nodes_to_delete = user_to_last_uses.get(user, []) nodes_to_delete = [i for i in nodes_to_delete if i.name not in to_keep] if len(nodes_to_delete): to_delete_str = ' = '.join([repr(n) for n in nodes_to_delete] + ['None']) body.append(f'; {to_delete_str}\n') else: body.append('\n') # NOTE: we add a variable to distinguish body and ckpt_func def emit_node(node: Node, body): maybe_type_annotation = '' if node.type is None else f' : {type_repr(node.type)}' if node.op == 'placeholder': assert isinstance(node.target, str) maybe_default_arg = '' if not node.args else f' = {repr(node.args[0])}' free_vars.append(f'{node.target}{maybe_type_annotation}{maybe_default_arg}') raw_name = node.target.replace('*', '') if raw_name != repr(node): body.append(f'{repr(node)} = {raw_name}\n') return elif node.op == 'call_method': assert isinstance(node.target, str) body.append( f'{repr(node)}{maybe_type_annotation} = {_format_target(repr(node.args[0]), node.target)}' f'({_format_args(node.args[1:], node.kwargs)})') return elif node.op == 'call_function': assert callable(node.target) # pretty print operators if node.target.__module__ == '_operator' and node.target.__name__ in magic_methods: assert isinstance(node.args, tuple) body.append(f'{repr(node)}{maybe_type_annotation} = ' f'{magic_methods[node.target.__name__].format(*(repr(a) for a in node.args))}') return # pretty print inplace operators; required for jit.script to work properly # not currently supported in normal FX graphs, but generated by torchdynamo if node.target.__module__ == '_operator' and node.target.__name__ in inplace_methods: body.append(f'{inplace_methods[node.target.__name__].format(*(repr(a) for a in node.args))}; ' f'{repr(node)}{maybe_type_annotation} = {repr(node.args[0])}') return qualified_name = _get_qualified_name(node.target) global_name = add_global(qualified_name, node.target) # special case for getattr: node.args could be 2-argument or 3-argument # 2-argument: attribute access; 3-argument: fall through to attrib function call with default value if global_name == 'getattr' and \ isinstance(node.args, tuple) and \ isinstance(node.args[1], str) and \ node.args[1].isidentifier() and \ len(node.args) == 2: body.append( f'{repr(node)}{maybe_type_annotation} = {_format_target(repr(node.args[0]), node.args[1])}') return body.append( f'{repr(node)}{maybe_type_annotation} = {global_name}({_format_args(node.args, node.kwargs)})') if node.meta.get('is_wrapped', False): wrapped_fns.setdefault(global_name) return elif node.op == 'call_module': assert isinstance(node.target, str) body.append(f'{repr(node)}{maybe_type_annotation} = ' f'{_format_target(root_module, node.target)}({_format_args(node.args, node.kwargs)})') return elif node.op == 'get_attr': assert isinstance(node.target, str) body.append(f'{repr(node)}{maybe_type_annotation} = {_format_target(root_module, node.target)}') return elif node.op == 'output': if node.type is not None: maybe_return_annotation[0] = f" -> {type_repr(node.type)}" body.append(self.generate_output(node.args[0])) return raise NotImplementedError(f'node: {node.op} {node.target}') # Modified for activation checkpointing ckpt_func = [] # if any node has a list of labels for activation_checkpoint, we # will use nested type of activation checkpoint codegen emit_code_with_chunk(body, ckpt_func, nodes, emit_node, delete_unused_values, self.meta_node, self.meta_graph) if len(body) == 0: # If the Graph has no non-placeholder nodes, no lines for the body # have been emitted. To continue to have valid Python code, emit a # single pass statement body.append('pass\n') if len(wrapped_fns) > 0: wrap_name = add_global('wrap', torch.fx.wrap) wrap_stmts = '\n'.join([f'{wrap_name}("{name}")' for name in wrapped_fns]) else: wrap_stmts = '' if self._body_transformer: body = self._body_transformer(body) for name, value in self.additional_globals(): add_global(name, value) # as we need colossalai.utils.checkpoint, we need to import colossalai # in forward function prologue = self.gen_fn_def(free_vars, maybe_return_annotation[0]) prologue = ''.join(ckpt_func) + prologue prologue = prologue code = ''.join(body) code = '\n'.join(' ' + line for line in code.split('\n')) fn_code = f""" {wrap_stmts} {prologue} {code}""" print(fn_code) return PythonCode(fn_code, globals_)