fp16_utils.py 22.9 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
#   Copyright (c) 2019 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

from __future__ import print_function

from ... import core
18
from ... import framework
19
from ... import layers
20 21
from ... import global_scope
from ...log_helper import get_logger
22 23 24
from ...wrapped_decorator import signature_safe_contextmanager
from .fp16_lists import AutoMixedPrecisionLists
import collections
25 26
import logging
import numpy as np
27

28
__all__ = ["fp16_guard", "cast_model_to_fp16", "cast_parameters_to_fp16"]
29

30 31
_logger = get_logger(
    __name__, logging.INFO, fmt='%(asctime)s-%(levelname)s: %(message)s')
32

33 34 35 36 37 38 39
_valid_types = [
    core.VarDesc.VarType.LOD_TENSOR, core.VarDesc.VarType.SELECTED_ROWS,
    core.VarDesc.VarType.LOD_TENSOR_ARRAY
]

_fp16_guard_pattern = "__use_fp16__"

40

J
Jie Fang 已提交
41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57
def _rename_arg(op, old_name, new_name):
    """
    If an op has old_name input and output, rename these input 
    args new_name.

    Args:
        op (Operator): Current operator.
        old_name (str): The old name of input args.
        new_name (str): The new name of input args.
    """
    op_desc = op.desc
    if isinstance(op_desc, tuple):
        op_desc = op_desc[0]
    op_desc._rename_input(old_name, new_name)
    op_desc._rename_output(old_name, new_name)


58 59 60 61 62 63 64 65 66 67 68 69
def _rename_op_input(program, op_var_rename_map, origin_ops, keep_fp32_ops):
    for block in program.blocks:
        ops = block.ops
        block_id = block.idx
        for op in ops:
            if op not in origin_ops or op in keep_fp32_ops:
                continue
            for name in op.input_arg_names:
                if name in op_var_rename_map[block_id]:
                    op._rename_input(name, op_var_rename_map[block_id][name])


J
Jie Fang 已提交
70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91
def _dtype_to_str(dtype):
    """
    Convert specific variable type to its corresponding string.

    Args:
        dtype (VarType): Variable type.
    """
    if dtype == core.VarDesc.VarType.FP16:
        return 'fp16'
    else:
        return 'fp32'


def _insert_cast_op(block, op, idx, src_dtype, dest_dtype):
    """
    Insert cast op and rename args of input and output.

    Args:
        block (Program): The block in which the operator is.
        op (Operator): The operator to insert cast op.
        idx (int): The index of current operator.
        src_dtype (VarType): The input variable dtype of cast op.
Z
Zhen Wang 已提交
92
        dest_dtype (VarType): The output variable dtype of cast op.
J
Jie Fang 已提交
93 94 95 96 97

    Returns:
        num_cast_op (int): The number of cast ops that have been inserted.
    """
    num_cast_ops = 0
98

J
Jie Fang 已提交
99
    for in_name in op.input_names:
Z
Zhang Ting 已提交
100
        if src_dtype == core.VarDesc.VarType.FP32 and op.type in [
F
furnace 已提交
101
                'batch_norm', 'fused_bn_add_activation', 'layer_norm'
Z
Zhang Ting 已提交
102 103
        ]:
            if in_name not in {'X', 'Z'}:
104
                continue
J
Jie Fang 已提交
105 106
        for in_var_name in op.input(in_name):
            in_var = block.var(in_var_name)
107
            if in_var.type not in _valid_types or in_var.dtype == dest_dtype:
J
Jie Fang 已提交
108 109
                continue
            if in_var.dtype == src_dtype:
110 111 112 113 114 115 116
                cast_name = in_var.name + '.cast_' + _dtype_to_str(dest_dtype)
                out_var = block.vars.get(cast_name)
                if out_var is None or out_var.dtype != dest_dtype:
                    out_var = block.create_var(
                        name=cast_name,
                        dtype=dest_dtype,
                        persistable=False,
Z
Zhen Wang 已提交
117
                        stop_gradient=in_var.stop_gradient)
118 119 120 121 122 123 124 125 126 127 128

                    block._insert_op(
                        idx,
                        type="cast",
                        inputs={"X": in_var},
                        outputs={"Out": out_var},
                        attrs={
                            "in_dtype": in_var.dtype,
                            "out_dtype": out_var.dtype
                        })
                    num_cast_ops += 1
J
Jie Fang 已提交
129 130 131 132
                _rename_arg(op, in_var.name, out_var.name)
            else:
                if op.has_attr('in_dtype'):
                    op._set_attr('in_dtype', dest_dtype)
Z
Zhen Wang 已提交
133
    if src_dtype == core.VarDesc.VarType.FP32 and dest_dtype == core.VarDesc.VarType.FP16:
J
Jie Fang 已提交
134
        for out_name in op.output_names:
F
furnace 已提交
135 136 137
            if op.type in [
                    'batch_norm', 'fused_bn_add_activation', 'layer_norm'
            ] and out_name != 'Y':
138
                continue
J
Jie Fang 已提交
139 140
            for out_var_name in op.output(out_name):
                out_var = block.var(out_var_name)
141
                if out_var.type not in _valid_types:
J
Jie Fang 已提交
142
                    continue
143 144
                if out_var.dtype == core.VarDesc.VarType.FP32:
                    out_var.desc.set_dtype(core.VarDesc.VarType.FP16)
J
Jie Fang 已提交
145
                    if op.has_attr('out_dtype'):
146
                        op._set_attr('out_dtype', core.VarDesc.VarType.FP16)
J
Jie Fang 已提交
147 148 149
    return num_cast_ops


150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181
def _insert_cast_post_op(block, op, idx, src_dtype, dest_dtype, target_name,
                         op_var_rename_map):
    num_cast_ops = 0

    target_var = block.var(target_name)
    if target_var.type not in _valid_types or target_var.dtype == dest_dtype:
        return num_cast_ops

    assert target_var.dtype == src_dtype, \
           "The real dtype({}) is not equal to the src dtype({})".format(_dtype_to_str(target_var.dtype), _dtype_to_str(src_dtype))

    cast_name = target_var.name + '.cast_' + _dtype_to_str(dest_dtype)
    cast_var = block.vars.get(cast_name)
    if cast_var is None or cast_var.dtype != dest_dtype:
        cast_var = block.create_var(
            name=cast_name,
            dtype=dest_dtype,
            persistable=False,
            stop_gradient=target_var.stop_gradient)
        block._insert_op(
            idx,
            type="cast",
            inputs={"X": target_var},
            outputs={"Out": cast_var},
            attrs={"in_dtype": target_var.dtype,
                   "out_dtype": cast_var.dtype})
        num_cast_ops += 1
        op_var_rename_map[block.idx][target_var.name] = cast_var.name

    return num_cast_ops


182 183 184 185 186 187 188 189 190 191
def find_true_prev_op(ops, cur_op, var_name):
    """
    Find the true prev op that outputs var_name variable.

    Args:
        ops (list): A list of ops.
        cur_op (Operator): Current operator which has var_name variable.
        var_name (string): Variable name.
    """
    prev_op = []
J
Jie Fang 已提交
192
    for op in ops:
193 194
        if op == cur_op:
            break
J
Jie Fang 已提交
195 196 197
        for out_name in op.output_names:
            for out_var_name in op.output(out_name):
                if out_var_name == var_name:
198 199 200 201 202 203 204 205
                    prev_op.append(op)
    if prev_op:
        if not len(prev_op) == 1:
            raise ValueError("There must be only one previous op "
                             "that outputs {0} variable".format(var_name))
        else:
            return prev_op[0]
    return None
J
Jie Fang 已提交
206 207


M
mapingshuo 已提交
208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227
def find_true_post_op(ops, cur_op, var_name):
    """
    if there are post ops, return them, if there is no post op,
    return None instead.
    Args:
        ops (list): A list of ops.
        cur_op (Operator): Current operator which has var_name variable.
        var_name (string): Variable name.
    """
    post_op = []
    for idx, op in enumerate(ops):
        if op == cur_op:
            break

    for i in range(idx + 1, len(ops)):
        op = ops[i]
        for in_name in op.input_names:
            for in_var_name in op.input(in_name):
                if in_var_name == var_name:
                    post_op.append(op)
228 229

    return post_op
M
mapingshuo 已提交
230 231 232 233 234 235 236 237 238 239 240


def find_op_index(block_desc, cur_op_desc):
    """
    """
    for idx in range(block_desc.op_size()):
        if cur_op_desc == block_desc.op(idx):
            return idx
    return -1


241 242 243 244 245 246 247 248 249 250 251 252
def _is_in_black_varnames(op, amp_lists):
    for in_name in op.input_arg_names:
        if in_name in amp_lists.black_varnames:
            return True

    for out_name in op.output_arg_names:
        if out_name in amp_lists.black_varnames:
            return True

    return False


253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290
def _need_keep_fp32(op, unsupported_op_list, use_fp16_guard):
    if op.type in unsupported_op_list:
        # the highest priority condition: If ops don't have fp16 computing kernels,
        # they must be executed in fp32 calculation pattern.
        return True

    # process ops about learning rate
    in_out_arg_names = []
    in_out_arg_names.extend(list(op.input_arg_names))
    in_out_arg_names.extend(list(op.output_arg_names))
    for name in in_out_arg_names:
        if "learning_rate" in name:
            return True

    if use_fp16_guard:
        if op.has_attr("op_namescope") and \
            (_fp16_guard_pattern in op.attr("op_namescope")):
            # op in fp16 guard
            return False
        else:
            # op not in fp16 guard
            return True
    else:
        return False


@signature_safe_contextmanager
def fp16_guard():
    """
    As for the pure fp16 training, if users set `use_fp16_guard` to True,
    only those ops created in the context manager `fp16_guard` will be
    transformed as float16 type.
    """
    with framework.name_scope(prefix=_fp16_guard_pattern):
        yield


def cast_model_to_fp16(program, amp_lists=None, use_fp16_guard=True):
291 292 293 294 295 296
    """
    Traverse all ops in the whole model and set their inputs and outputs
    to the fp16 data type. This function will do some special process for
    the batch normalization, which keeps the computational process of
    batchnorms in FP32.
    Args:
297 298 299 300
        program (Program): The used program.
        amp_lists (AutoMixedPrecisionLists): An AutoMixedPrecisionLists object.
        use_fp16_guard(bool): Determine whether to use `fp16_guard` when
                              constructing the program. Default True.
301 302
    """

303 304 305 306 307 308 309 310 311 312
    if amp_lists is None:
        amp_lists = AutoMixedPrecisionLists()
    global_block = program.global_block()
    keep_fp32_ops = set()
    to_fp16_var_names = set()
    origin_ops = []
    for block in program.blocks:
        origin_ops.extend(block.ops)

    for block in program.blocks:
313 314 315 316
        ops = block.ops
        for op in ops:
            if op.type == 'create_py_reader' or op.type == 'read':
                continue
317 318 319
            if _need_keep_fp32(op, amp_lists.unsupported_list, use_fp16_guard):
                keep_fp32_ops.add(op)
                continue  # processed below
320 321 322 323 324 325 326 327 328 329 330
            for in_name in op.input_names:
                if op.type in {
                        'batch_norm', 'fused_bn_add_activation', 'layer_norm'
                } and in_name not in {'X', 'Z'}:
                    continue
                for in_var_name in op.input(in_name):
                    in_var = None
                    try:
                        in_var = block.var(in_var_name)
                    except ValueError as e:
                        _logger.debug(
331
                            "-- {}, try to get it in the global block --".
332 333 334 335
                            format(e))
                        in_var = global_block.var(in_var_name)
                        if in_var is not None:
                            _logger.debug(
336
                                "-- var {} is got in the global block --".
337 338
                                format(in_var_name))

339
                    if in_var is None or in_var.type not in _valid_types:
340 341 342 343
                        continue

                    if in_var.dtype == core.VarDesc.VarType.FP32:
                        in_var.desc.set_dtype(core.VarDesc.VarType.FP16)
344
                        to_fp16_var_names.add(in_var_name)
345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360

                    _logger.debug(
                        "-- op type: {}, in var name: {}, in var dtype: {} --".
                        format(op.type, in_var_name, in_var.dtype))

            for out_name in op.output_names:
                if op.type in {
                        'batch_norm', 'fused_bn_add_activation', 'layer_norm'
                } and out_name != 'Y':
                    continue
                for out_var_name in op.output(out_name):
                    out_var = None
                    try:
                        out_var = block.var(out_var_name)
                    except ValueError as e:
                        _logger.debug(
361
                            "-- {}, try to get it in the global block --".
362 363 364 365
                            format(e))
                        out_var = global_block.var(out_var_name)
                        if out_var is not None:
                            _logger.debug(
366
                                "-- var {} is got in the global block --".
367 368
                                format(out_var_name))

369
                    if out_var is None or out_var.type not in _valid_types:
370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387
                        continue

                    if out_var.dtype == core.VarDesc.VarType.FP32:
                        out_var.desc.set_dtype(core.VarDesc.VarType.FP16)

                    _logger.debug(
                        "-- op type: {}, out var name: {}, out var dtype: {} --".
                        format(op.type, out_var_name, out_var.dtype))
            if op.has_attr('in_dtype') and op.attr(
                    'in_dtype') == core.VarDesc.VarType.FP32:
                op._set_attr('in_dtype', core.VarDesc.VarType.FP16)
            if op.has_attr('out_dtype') and op.attr(
                    'out_dtype') == core.VarDesc.VarType.FP32:
                op._set_attr('out_dtype', core.VarDesc.VarType.FP16)
            if op.has_attr('dtype') and op.attr(
                    'dtype') == core.VarDesc.VarType.FP32:
                op._set_attr('dtype', core.VarDesc.VarType.FP16)

388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422
    # process ops in keep_fp32_ops
    op_var_rename_map = [
        collections.OrderedDict() for _ in range(len(program.blocks))
    ]
    for block in program.blocks:
        ops = block.ops
        idx = 0
        while idx < len(ops):
            op = ops[idx]
            num_cast_ops = 0
            if op in keep_fp32_ops:
                pre_cast_num = _insert_cast_op(block, op, idx,
                                               core.VarDesc.VarType.FP16,
                                               core.VarDesc.VarType.FP32)
                num_cast_ops += pre_cast_num
                for out_var_name in op.output_arg_names:
                    out_var = block.vars.get(out_var_name)
                    if out_var is None or out_var.type not in _valid_types:
                        continue
                    if out_var.dtype == core.VarDesc.VarType.FP16:
                        out_var.desc.set_dtype(core.VarDesc.VarType.FP32)
                        post_ops = find_true_post_op(ops, op, out_var_name)
                        for post_op in post_ops:
                            if post_op in keep_fp32_ops:
                                continue
                            post_cast_num = _insert_cast_post_op(
                                block, op, idx + pre_cast_num + 1,
                                core.VarDesc.VarType.FP32,
                                core.VarDesc.VarType.FP16, out_var_name,
                                op_var_rename_map)
                            num_cast_ops += post_cast_num
            idx += num_cast_ops + 1

    _rename_op_input(program, op_var_rename_map, origin_ops, keep_fp32_ops)
    return to_fp16_var_names
423

424 425

def cast_parameters_to_fp16(place, program, scope=None, to_fp16_var_names=None):
426
    """
427
    Traverse all parameters in the whole model and set them to the FP16 data type.
428 429
    Whereas, this function will keep parameters of batchnorms in FP32.
    Args:
430 431 432 433 434 435 436
        place(fluid.CPUPlace|fluid.CUDAPlace): `place` is used to restore the FP16 weight tensors.
        program (Program): The used program.
        scope(fluid.Scope, optional): `scope` is used to get the FP32 weight tensor values.
                                      Default is None.
        to_fp16_var_names(set|list, optional): The data types of vars in `to_fp16_var_names`
                                               will be set to FP16. Usually, it is the returned
                                               value of `cast_model_to_fp16` API.
437
    """
438 439 440 441 442 443
    all_parameters = []
    for block in program.blocks:
        all_parameters.extend(block.all_parameters())

    fp16_var_names = to_fp16_var_names if to_fp16_var_names else set()
    var_scope = scope if scope else global_scope()
444
    for param in all_parameters:
445 446
        if param.name in fp16_var_names:
            _logger.debug("---- cast {} to fp16 dtype ----".format(param.name))
447 448 449 450 451
            param_t = var_scope.find_var(param.name).get_tensor()
            data = np.array(param_t)
            param_t.set(np.float16(data), place)


J
Jie Fang 已提交
452
def rewrite_program(main_prog, amp_lists):
J
Jie Fang 已提交
453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475
    """
    Traverse all ops in current block and insert cast op according to 
    which set current op belongs to.

    1. When an op belongs to the black list, add it to black set
    2. When an op belongs to the white list, add it to white set
    3. When an op belongs to the gray list. If one 
       of its inputs is the output of black set op or black list op, 
       add it to black set. If all of its previous ops are not black 
       op and one of its inputs is the output of white set op or 
       white list op, add it to white set.
    4. When an op isn't in the lists, add it to black op set.
    5. Add necessary cast ops to make sure that black set op will be 
       computed in fp32 mode, while white set op will be computed in 
       fp16 mode.

    Args:
        main_prog (Program): The main program for training.
    """
    block = main_prog.global_block()
    ops = block.ops
    white_op_set = set()
    black_op_set = set()
476
    for op in ops:
477 478 479 480 481 482 483 484

        # NOTE(zhiqiu): 'create_py_reader' and 'read' is used in non-iterable DataLoder, 
        # we don't need to handle reader op and the input of 'create_py_reader' is not 
        # in block, which may result in errors.
        # See GeneratorLoader._init_non_iterable() for details.
        if op.type == 'create_py_reader' or op.type == 'read':
            continue

485 486 487 488 489
        if amp_lists.black_varnames is not None and _is_in_black_varnames(
                op, amp_lists):
            black_op_set.add(op)
            continue

J
Jie Fang 已提交
490
        if op.type in amp_lists.black_list:
J
Jie Fang 已提交
491
            black_op_set.add(op)
J
Jie Fang 已提交
492
        elif op.type in amp_lists.white_list:
J
Jie Fang 已提交
493
            white_op_set.add(op)
J
Jie Fang 已提交
494
        elif op.type in amp_lists.gray_list:
J
Jie Fang 已提交
495 496 497 498 499 500 501 502 503 504
            is_black_op = False
            is_white_op = False
            for in_name in op.input_names:
                # if this op has inputs
                if in_name:
                    for in_var_name in op.input(in_name):
                        in_var = block.var(in_var_name)
                        # this in_var isn't the output of other op
                        if in_var.op is None:
                            continue
505 506 507 508
                        elif in_var.op is op:
                            prev_op = find_true_prev_op(ops, op, in_var_name)
                            if prev_op is None:
                                continue
J
Jie Fang 已提交
509 510 511 512
                        else:
                            prev_op = in_var.op
                        # if it's one of inputs
                        if prev_op in black_op_set or \
J
Jie Fang 已提交
513
                                prev_op.type in amp_lists.black_list:
J
Jie Fang 已提交
514
                            is_black_op = True
515
                        elif prev_op in white_op_set or \
J
Jie Fang 已提交
516
                                prev_op.type in amp_lists.white_list:
J
Jie Fang 已提交
517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546
                            is_white_op = True
            if is_black_op:
                black_op_set.add(op)
            elif is_white_op:
                white_op_set.add(op)
            else:
                pass
        else:
            # For numerical safe, we apply fp32 computation on ops that
            # are not determined which list they should stay.
            black_op_set.add(op)

    idx = 0
    while idx < len(ops):
        op = ops[idx]
        num_cast_ops = 0
        if op in black_op_set:
            num_cast_ops = _insert_cast_op(block, op, idx,
                                           core.VarDesc.VarType.FP16,
                                           core.VarDesc.VarType.FP32)
        elif op in white_op_set:
            num_cast_ops = _insert_cast_op(block, op, idx,
                                           core.VarDesc.VarType.FP32,
                                           core.VarDesc.VarType.FP16)
        else:
            pass

        idx += num_cast_ops + 1


547 548 549
def update_role_var_grad(main_prog, params_grads):
    """
    Update op_role_var attr for some ops to make sure the gradients
Z
Zhen Wang 已提交
550
    transferred across GPUs is FP16.
551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581
    1. Check whether the op that outputs gradient is cast or not.
    2. If op is cast and gradient is FP32, remove the op_role_var
       and find the prev op which outputs FP16 gradient
    3. Update the op_role_var of the prev op.

    Args:
        main_prog (Program): The main program for training.
        params_grads (list): A list of params and grads.
    """
    block = main_prog.global_block()
    BACKWARD = core.op_proto_and_checker_maker.OpRole.Backward
    OPTIMIZE = core.op_proto_and_checker_maker.OpRole.Optimize
    for p, g in params_grads:
        op = g.op
        if g.dtype == core.VarDesc.VarType.FP32 and op.type == 'cast':
            role = op.attr('op_role')
            if role & int(BACKWARD) and op.has_attr('op_role_var'):
                op.desc.remove_attr("op_role_var")
            else:
                raise ValueError("The cast op {0} must be in BACKWARD role "
                                 "and have op_role_var attr.".format(op))

            fp16_grad_name = op.input(op.input_names[0])[0]
            op_for_fp16_grad = find_true_prev_op(block.ops, op, fp16_grad_name)
            op_role_var_attr_name = \
                core.op_proto_and_checker_maker.kOpRoleVarAttrName()
            attr_val = [p.name, fp16_grad_name]
            if op_for_fp16_grad.has_attr(op_role_var_attr_name):
                attr_val.extend(op_for_fp16_grad.attr(op_role_var_attr_name))
            op_for_fp16_grad._set_attr(op_role_var_attr_name, attr_val)

Z
Zhen Wang 已提交
582 583
            # Maximize the all_reduce overlap, and perform the cast
            # operation after gradients transfer.
584
            op._set_attr('op_role', OPTIMIZE)
M
mapingshuo 已提交
585 586 587 588
            # optimize op should stay behind forward and backward ops
            if op == block.ops[-1]:
                continue
            post_ops = find_true_post_op(block.ops, op, g.name)
589
            if post_ops:
M
mapingshuo 已提交
590 591 592 593 594 595 596 597 598 599 600
                raise ValueError("The cast op {0}'s output should not be"
                                 "used by a non-optimize op, however, it"
                                 "is used by {1}".format(op, post_ops[0]))
            new_op_desc = block.desc.append_op()
            new_op_desc.copy_from(op.desc)

            op_idx = find_op_index(block.desc, op.desc)
            if op_idx == -1:
                raise ValueError("The op {0} is not in program".format(op))
            block.desc._remove_op(op_idx, op_idx + 1)
        block._sync_with_cpp()