File size: 18,333 Bytes
453a6ca | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 | # Copyright (c) 2023 Alibaba PAI and Nvidia Megatron-LM Team.
#
# 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.
import torch
from megatron.core import mpu
try:
from megatron import get_args
except:
from megatron.training import get_args
from megatron_patch.tokenizer import get_tokenizer
def get_ltor_masks_and_position_ids(data,
eod_token,
reset_position_ids,
reset_attention_mask,
eod_mask_loss,
create_attention_mask: bool=True):
"""Build masks and position id for left to right model."""
# Extract batch size and sequence length.
micro_batch_size, seq_length = data.size()
# Attention mask (lower triangular).
if reset_attention_mask:
att_mask_batch = micro_batch_size
else:
att_mask_batch = 1
if create_attention_mask:
attention_mask = torch.tril(torch.ones(
(att_mask_batch, seq_length, seq_length), device=data.device)).view(
att_mask_batch, 1, seq_length, seq_length)
else:
attention_mask = None
# Loss mask.
loss_mask = torch.ones(data.size(), dtype=torch.float, device=data.device)
if eod_mask_loss:
loss_mask[data == eod_token] = 0.0
# Position ids.
position_ids = torch.arange(seq_length, dtype=torch.long,
device=data.device)
position_ids = position_ids.unsqueeze(0).expand_as(data)
# We need to clone as the ids will be modifed based on batch index.
if reset_position_ids:
position_ids = position_ids.clone()
if reset_position_ids or reset_attention_mask:
# Loop through the batches:
for b in range(micro_batch_size):
# Find indecies where EOD token is.
eod_index = position_ids[b, data[b] == eod_token]
# Detach indecies from positions if going to modify positions.
if reset_position_ids:
eod_index = eod_index.clone()
# Loop through EOD indecies:
prev_index = 0
for j in range(eod_index.size()[0]):
i = eod_index[j]
# Mask attention loss.
if reset_attention_mask and attention_mask is not None:
attention_mask[b, 0, (i + 1):, :(i + 1)] = 0
# Reset positions.
if reset_position_ids:
position_ids[b, (i + 1):] -= (i + 1 - prev_index)
prev_index = i + 1
if attention_mask is not None:
# Convert attention mask to binary:
attention_mask = (attention_mask < 0.5)
return attention_mask, loss_mask, position_ids
def get_ltor_position_ids_packed_seq(data):
"""
Given a input_seqs from custom mmap dataset, generate a
position_ids by searching negative tokens.
"""
# Extract batch size and sequence length.
micro_batch_size, seq_length = data.size()
# Position ids.
position_ids = torch.arange(seq_length, dtype=torch.long, device=data.device)
position_ids = position_ids.unsqueeze(0).expand_as(data)
# We need to clone as the ids will be modifed based on batch index.
position_ids = position_ids.clone()
# Loop through the batches:
for b in range(micro_batch_size):
# Find indecies where EOD token is.
eod_index = position_ids[b, data[b] < 0]
# Detach indecies from positions if going to modify positions.
eod_index = eod_index.clone()
# Loop through EOD indecies:
prev_index = 0
for j in range(eod_index.size()[0]):
i = eod_index[j]
position_ids[b, (i + 1):] -= (i + 1 - prev_index)
prev_index = i + 1
return position_ids
def get_batch_on_this_tp_rank_original(data_iterator, per_seq_average=False):
args = get_args()
tokenizer = get_tokenizer()
def _broadcast(item):
if item is None:
return
torch.distributed.broadcast(item, mpu.get_tensor_model_parallel_src_rank(),
group=mpu.get_tensor_model_parallel_group())
if mpu.get_tensor_model_parallel_rank() == 0:
if isinstance(data_iterator, dict):
data = data_iterator
else:
data = next(data_iterator)
tokens_ = data['input_ids'].long()
labels_ = data['labels'].long()
tokens = tokens_[:, :-1].contiguous()
labels = labels_[:, 1:].contiguous()
# core/tensor_parallel/cross_entropy.py, target_mask = (target < vocab_start_index) | (target >= vocab_end_index)
# labels[labels == tokenizer.eos_token_id] = -100
# NOTE: if eos == pad, we map <eos> to - 1 - eos_id, map these tokens back
tokens[tokens < 0] = - 1 - tokens[tokens < 0]
eos_indices = (labels < 0).nonzero()
labels[labels == tokenizer.pad_token_id] = -100
labels[eos_indices[:, 0], eos_indices[:, 1]] = - 1 - labels[eos_indices[:, 0], eos_indices[:, 1]]
attention_mask, loss_mask, position_ids = get_ltor_masks_and_position_ids(
labels,
-100,
args.reset_position_ids,
args.reset_attention_mask,
args.eod_mask_loss)
num_seqs = None
if per_seq_average:
# NOTE: raw dataset does not support sequence packing
num_seqs = torch.ones(position_ids.shape[0], device=torch.cuda.current_device(), dtype=torch.int64)
loss_mask = loss_mask / loss_mask.sum(dim=-1, keepdims=True) # [mbs]
batch = {
'tokens': tokens.cuda(non_blocking=True),
'labels': labels.cuda(non_blocking=True),
'loss_mask': loss_mask.cuda(non_blocking=True),
'attention_mask': attention_mask.cuda(non_blocking=True),
'position_ids': position_ids.cuda(non_blocking=True),
'num_seqs': num_seqs.cuda(non_blocking=True) if num_seqs is not None else None
}
if args.pipeline_model_parallel_size == 1:
_broadcast(batch['tokens'])
_broadcast(batch['labels'])
_broadcast(batch['loss_mask'])
_broadcast(batch['attention_mask'])
_broadcast(batch['position_ids'])
_broadcast(batch['num_seqs'])
elif mpu.is_pipeline_first_stage():
_broadcast(batch['tokens'])
_broadcast(batch['attention_mask'])
_broadcast(batch['position_ids'])
elif mpu.is_pipeline_last_stage():
# Multi-Token Prediction (MTP) layers need tokens and position_ids to calculate embedding.
# Currently the Multi-Token Prediction (MTP) layers is fixed on the last stage, so we need
# to broadcast tokens and position_ids to all of the tensor parallel ranks on the last stage.
if getattr(args, 'mtp_num_layers', None) is not None:
_broadcast(batch['tokens'])
_broadcast(batch['position_ids'])
_broadcast(batch['labels'])
_broadcast(batch['loss_mask'])
_broadcast(batch['attention_mask'])
_broadcast(batch['num_seqs'])
else:
tokens = torch.empty((args.micro_batch_size, args.seq_length), dtype=torch.int64,
device=torch.cuda.current_device())
labels = torch.empty((args.micro_batch_size, args.seq_length), dtype=torch.int64,
device=torch.cuda.current_device())
loss_mask = torch.empty((args.micro_batch_size, args.seq_length), dtype=torch.float32,
device=torch.cuda.current_device())
mbs = args.micro_batch_size if args.reset_attention_mask else 1
attention_mask = torch.empty((mbs, 1, args.seq_length, args.seq_length), dtype=torch.bool,
device=torch.cuda.current_device())
position_ids = torch.empty((args.micro_batch_size, args.seq_length), dtype=torch.int64,
device=torch.cuda.current_device())
num_seqs = None
if per_seq_average:
num_seqs = torch.empty((args.micro_batch_size,), dtype=torch.int64,
device=torch.cuda.current_device())
if args.pipeline_model_parallel_size == 1:
_broadcast(tokens)
_broadcast(labels)
_broadcast(loss_mask)
_broadcast(attention_mask)
_broadcast(position_ids)
_broadcast(num_seqs)
elif mpu.is_pipeline_first_stage():
labels = None
loss_mask = None
num_seqs = None
_broadcast(tokens)
_broadcast(attention_mask)
_broadcast(position_ids)
elif mpu.is_pipeline_last_stage():
# Multi-Token Prediction (MTP) layers need tokens and position_ids to calculate embedding.
# Currently the Multi-Token Prediction (MTP) layers is fixed on the last stage, so we need
# to broadcast tokens and position_ids to all of the tensor parallel ranks on the last stage.
if getattr(args, 'mtp_num_layers', None) is not None:
_broadcast(tokens)
_broadcast(position_ids)
else:
tokens = None
position_ids = None
_broadcast(labels)
_broadcast(loss_mask)
_broadcast(attention_mask)
_broadcast(num_seqs)
batch = {
'tokens': tokens,
'labels': labels,
'loss_mask': loss_mask,
'attention_mask': attention_mask,
'position_ids': position_ids,
'num_seqs': num_seqs
}
return batch
def get_position_id_on_this_tp_rank_idxmap_sft_packing(data_iterator):
args = get_args()
tokenizer = get_tokenizer()
def _broadcast(item):
if item is None:
return
torch.distributed.broadcast(item, mpu.get_tensor_model_parallel_src_rank(),
group=mpu.get_tensor_model_parallel_group())
if mpu.get_tensor_model_parallel_rank() == 0:
if isinstance(data_iterator, dict):
data = data_iterator
else:
data = next(data_iterator)
actual_seqlen = args.seq_length
data['tokens'] = data['tokens'].long()
tokens = data['tokens'][..., :actual_seqlen]
position_ids = get_ltor_position_ids_packed_seq(tokens).cuda(non_blocking=True)
else:
# dtype: long
position_ids = torch.empty((args.micro_batch_size, args.seq_length), dtype=torch.int64,
device=torch.cuda.current_device())
_broadcast(position_ids)
return position_ids
def get_batch_on_this_tp_rank_idxmap_sft(data_iterator, per_seq_average=False):
args = get_args()
tokenizer = get_tokenizer()
def _broadcast(item):
if item is None:
return
torch.distributed.broadcast(item, mpu.get_tensor_model_parallel_src_rank(),
group=mpu.get_tensor_model_parallel_group())
if mpu.get_tensor_model_parallel_rank() == 0:
if isinstance(data_iterator, dict):
data = data_iterator
else:
data = next(data_iterator)
# sanity check
assert data['tokens'].shape[-1] == 2 * args.seq_length
actual_seqlen = args.seq_length
data['tokens'] = data['tokens'].long()
tokens = data['tokens'][..., :actual_seqlen]
labels = data['tokens'][..., actual_seqlen:]
loss_mask = (labels != -100).float()
if args.reset_position_ids:
attention_mask = None
position_ids = get_ltor_position_ids_packed_seq(tokens)
has_pad = tokens[:, -1] >= 0
tokens[tokens < 0] = - tokens[tokens < 0] - 1
else:
tokens[tokens < 0] = - tokens[tokens < 0] - 1
attention_mask, _, position_ids = get_ltor_masks_and_position_ids(
tokens,
tokenizer.eod,
args.reset_position_ids,
args.reset_attention_mask,
False,
args.create_attention_mask_in_dataloader
)
num_seqs = None
if per_seq_average:
num_seqs = torch.ones(position_ids.shape[0], device=torch.cuda.current_device(), dtype=torch.int64)
if args.reset_position_ids:
for b in range(position_ids.shape[0]):
p = position_ids[b]
start_indices = (p == 0).nonzero(as_tuple=True)[0]
seqlens = start_indices[1:] - start_indices[:-1]
seqlens = seqlens.cpu().numpy().tolist() + [p.shape[0] - start_indices[-1].item()]
subseqs = torch.split(loss_mask[b], seqlens)
num_seqs[b] = len(seqlens) - int(has_pad[b])
for subseq_idx, (start_idx, seqlen, subseq) in enumerate(zip(start_indices, seqlens, subseqs)):
if subseq_idx == num_seqs[b]:
# NOTE: do not process pad sequence
continue
assert subseq.sum() > 0
loss_mask[b, start_idx: start_idx + seqlen] /= subseq.sum()
else:
loss_mask = loss_mask / loss_mask.sum(dim=-1, keepdims=True) # [mbs]
# dtype: long, long, float, bool, long
batch = {
'tokens': tokens.cuda(non_blocking=True),
'labels': labels.cuda(non_blocking=True),
'loss_mask': loss_mask.cuda(non_blocking=True),
'attention_mask': attention_mask.cuda(non_blocking=True) if attention_mask is not None else None,
'position_ids': position_ids.cuda(non_blocking=True),
'num_seqs': num_seqs.cuda(non_blocking=True) if num_seqs is not None else None
}
if args.pipeline_model_parallel_size == 1:
_broadcast(batch['tokens'])
_broadcast(batch['labels'])
_broadcast(batch['loss_mask'])
_broadcast(batch['attention_mask'])
_broadcast(batch['num_seqs'])
elif mpu.is_pipeline_first_stage():
_broadcast(batch['tokens'])
_broadcast(batch['attention_mask'])
elif mpu.is_pipeline_last_stage():
# Multi-Token Prediction (MTP) layers need tokens and position_ids to calculate embedding.
# Currently the Multi-Token Prediction (MTP) layers is fixed on the last stage, so we need
# to broadcast tokens and position_ids to all of the tensor parallel ranks on the last stage.
if getattr(args, 'mtp_num_layers', None) is not None:
_broadcast(batch['tokens'])
_broadcast(batch['labels'])
_broadcast(batch['loss_mask'])
_broadcast(batch['attention_mask'])
_broadcast(batch['num_seqs'])
_broadcast(batch['position_ids'])
else:
# dtype: long, long, float, bool, long
tokens = torch.empty((args.micro_batch_size, args.seq_length), dtype=torch.int64,
device=torch.cuda.current_device())
labels = torch.empty((args.micro_batch_size, args.seq_length), dtype=torch.int64,
device=torch.cuda.current_device())
loss_mask = torch.empty((args.micro_batch_size, args.seq_length), dtype=torch.float32,
device=torch.cuda.current_device())
attention_mask = None
if args.create_attention_mask_in_dataloader:
mbs = args.micro_batch_size if args.reset_attention_mask else 1
attention_mask = torch.empty((mbs, 1, args.seq_length, args.seq_length), dtype=torch.bool,
device=torch.cuda.current_device())
position_ids = torch.empty((args.micro_batch_size, args.seq_length), dtype=torch.int64,
device=torch.cuda.current_device())
num_seqs = None
if per_seq_average:
num_seqs = torch.empty((args.micro_batch_size,), dtype=torch.int64,
device=torch.cuda.current_device())
if args.pipeline_model_parallel_size == 1:
_broadcast(tokens)
_broadcast(labels)
_broadcast(loss_mask)
_broadcast(attention_mask)
_broadcast(num_seqs)
elif mpu.is_pipeline_first_stage():
labels = None
loss_mask = None
num_seqs = None
_broadcast(tokens)
_broadcast(attention_mask)
elif mpu.is_pipeline_last_stage():
# Multi-Token Prediction (MTP) layers need tokens and position_ids to calculate embedding.
# Currently the Multi-Token Prediction (MTP) layers is fixed on the last stage, so we need
# to broadcast tokens and position_ids to all of the tensor parallel ranks on the last stage.
if getattr(args, 'mtp_num_layers', None) is not None:
_broadcast(tokens)
else:
tokens = None
_broadcast(labels)
_broadcast(loss_mask)
_broadcast(attention_mask)
_broadcast(num_seqs)
_broadcast(position_ids)
batch = {
'tokens': tokens,
'labels': labels,
'loss_mask': loss_mask,
'attention_mask': attention_mask,
'position_ids': position_ids,
'num_seqs': num_seqs
}
return batch |