File size: 41,979 Bytes
d91766b | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 | import os
# os.environ["ALL_PROXY"] = "socks5h://127.0.0.1:13659"
os.environ['CURL_CA_BUNDLE'] = ''
os.environ['REQUESTS_CA_BUNDLE'] = ''
os.environ["HF_ENDPOINT"] = "https://hf-mirror.com"
import accelerate
import torch
import random
import numpy as np
import torch.nn.functional as F
from datasets import Dataset
from lm_eval.__main__ import cli_evaluate
from lm_eval.api.model import LM
from lm_eval.api.registry import register_model
from tqdm import tqdm
from transformers import AutoTokenizer, AutoConfig, AutoModel
from model.modeling_llada import LLaDAModelLM
import json
import time
from accelerate import (
Accelerator,
InitProcessGroupKwargs,
)
from datetime import timedelta
def set_seed(seed):
torch.manual_seed(seed)
random.seed(seed)
np.random.seed(seed)
torch.backends.cudnn.deterministic = True
torch.backends.cudnn.benchmark = False
def add_gumbel_noise(logits, temperature):
if temperature == 0:
return logits
logits = logits.to(torch.float64)
noise = torch.rand_like(logits, dtype=torch.float64)
gumbel_noise = (-torch.log(noise)) ** temperature
return logits.exp() / gumbel_noise
def get_num_transfer_tokens(mask_index, steps):
mask_num = mask_index.sum(dim=1, keepdim=True)
base = mask_num // steps
remainder = mask_num % steps
num_transfer_tokens = torch.zeros(mask_num.size(0), steps, device=mask_index.device, dtype=torch.int64) + base
for i in range(mask_num.size(0)):
num_transfer_tokens[i, :remainder[i]] += 1
return num_transfer_tokens
def get_transfer_index_entropy(
logits,
temperature,
remasking,
mask_index,
x,
entropy_threshold=None,
num_transfer_tokens=None,
):
logits_with_noise = add_gumbel_noise(logits, temperature=temperature)
x0 = torch.argmax(logits_with_noise, dim=-1)
p = F.softmax(logits.to(torch.float64), dim=-1)
if remasking == "low_confidence":
entropy = -torch.sum(p * torch.log(p + 1e-12), dim=-1)
elif remasking == "random":
entropy = torch.rand((x0.shape[0], x0.shape[1]), device=x0.device)
else:
raise NotImplementedError(remasking)
x0 = torch.where(mask_index, x0, x)
entropy_for_selection = torch.where(mask_index, entropy, torch.inf)
transfer_index = torch.zeros_like(x0, dtype=torch.bool, device=x0.device)
if entropy_threshold is not None:
transfer_index = entropy_for_selection < entropy_threshold
for j in range(entropy_for_selection.shape[0]):
if mask_index[j].sum() > 0 and transfer_index[j].sum() == 0:
min_index = torch.argmin(entropy_for_selection[j])
transfer_index[j, min_index] = True
else:
for j in range(entropy_for_selection.shape[0]):
_, select_index = torch.topk(entropy_for_selection[j], k=num_transfer_tokens[j], largest=False)
transfer_index[j, select_index] = True
return x0, transfer_index
def _check_early_stop(x, prompt_length, eos_token_id, mask_id):
"""
Check for EOS token in the generation region.
If found, set all tokens after the first EOS to EOS (not mask) and
return (True, first_eos_absolute_position).
Otherwise return (False, None).
"""
if eos_token_id is None:
return False, None
gen_region = x[:, prompt_length:]
eos_mask = (gen_region == eos_token_id) & (gen_region != mask_id)
if not eos_mask.any():
return False, None
pos = torch.arange(gen_region.shape[1], device=x.device).unsqueeze(0)
first_eos_rel = torch.where(eos_mask, pos, gen_region.shape[1]).amin(dim=1)
first_eos_abs = prompt_length + first_eos_rel[0].item()
# Set everything after first EOS to EOS (not mask)
x[:, first_eos_abs + 1:] = eos_token_id
return True, first_eos_abs
@torch.no_grad()
def generate(
model,
prompt,
steps=128,
gen_length=128,
block_length=128,
temperature=0.0,
remasking="low_confidence",
mask_id=126336,
threshold=None,
eos_token_id=None,
):
x = torch.full((1, prompt.shape[1] + gen_length), mask_id, dtype=torch.long).to(model.device)
x[:, :prompt.shape[1]] = prompt.clone()
prompt_length = prompt.shape[1]
assert gen_length % block_length == 0
num_blocks = gen_length // block_length
assert steps % num_blocks == 0
steps = steps // num_blocks
nfe = 0
for num_block in range(num_blocks):
# Early stop: skip remaining blocks if EOS already found
if eos_token_id is not None:
has_eos, _ = _check_early_stop(x, prompt_length, eos_token_id, mask_id)
if has_eos:
break
block_mask_index = (
x[:, prompt.shape[1] + num_block * block_length : prompt.shape[1] + (num_block + 1) * block_length] == mask_id
)
num_transfer_tokens = get_num_transfer_tokens(block_mask_index, steps)
i = 0
while True:
nfe += 1
mask_index = x == mask_id
logits = model(x).logits
mask_index[:, prompt.shape[1] + (num_block + 1) * block_length :] = 0
x0, transfer_index = get_transfer_index_entropy(
logits,
temperature,
remasking,
mask_index,
x,
entropy_threshold=threshold,
num_transfer_tokens=num_transfer_tokens[:, i] if threshold is None else None,
)
x[transfer_index] = x0[transfer_index]
i += 1
# Early stop: check after each decode step
if eos_token_id is not None:
has_eos, _ = _check_early_stop(x, prompt_length, eos_token_id, mask_id)
if has_eos:
break
if (
x[:, prompt.shape[1] + num_block * block_length : prompt.shape[1] + (num_block + 1) * block_length] == mask_id
).sum() == 0:
break
return x, nfe
@torch.no_grad()
def generate_with_prefix_cache(
model,
prompt,
steps=128,
gen_length=128,
block_length=128,
temperature=0.0,
remasking="low_confidence",
mask_id=126336,
threshold=None,
eos_token_id=None,
):
x = torch.full((1, prompt.shape[1] + gen_length), mask_id, dtype=torch.long).to(model.device)
x[:, :prompt.shape[1]] = prompt.clone()
prompt_length = prompt.shape[1]
assert gen_length % block_length == 0
num_blocks = gen_length // block_length
assert steps % num_blocks == 0
steps = steps // num_blocks
nfe = 0
for num_block in range(num_blocks):
# Early stop: skip remaining blocks if EOS already found
if eos_token_id is not None:
has_eos, _ = _check_early_stop(x, prompt_length, eos_token_id, mask_id)
if has_eos:
break
current_block_start = prompt.shape[1] + num_block * block_length
current_block_end = current_block_start + block_length
block_mask_index = x[:, current_block_start:current_block_end] == mask_id
num_transfer_tokens = get_num_transfer_tokens(block_mask_index, steps)
output = model(x, use_cache=True)
past_key_values = output.past_key_values
mask_index = x == mask_id
mask_index[:, current_block_end:] = 0
x0, transfer_index = get_transfer_index_entropy(
output.logits,
temperature,
remasking,
mask_index,
x,
entropy_threshold=threshold,
num_transfer_tokens=num_transfer_tokens[:, 0] if threshold is None else None,
)
x[transfer_index] = x0[transfer_index]
new_past_key_values = []
for i in range(len(past_key_values)):
new_past_key_values.append(())
for j in range(len(past_key_values[i])):
new_past_key_values[i] += (past_key_values[i][j][:, :, :current_block_start],)
past_key_values = new_past_key_values
nfe += 1
i = 1
while True:
nfe += 1
mask_index = x[:, current_block_start:] == mask_id
mask_index[:, block_length:] = 0
logits = model(x[:, current_block_start:], past_key_values=past_key_values, use_cache=True).logits
x0, transfer_index = get_transfer_index_entropy(
logits,
temperature,
remasking,
mask_index,
x[:, current_block_start:],
entropy_threshold=threshold,
num_transfer_tokens=num_transfer_tokens[:, i] if threshold is None else None,
)
x[:, current_block_start:][transfer_index] = x0[transfer_index]
# Early stop: check after each decode step
if eos_token_id is not None:
has_eos, _ = _check_early_stop(x, prompt_length, eos_token_id, mask_id)
if has_eos:
break
if (x[:, current_block_start:current_block_end] == mask_id).sum() == 0:
break
i += 1
return x, nfe
@torch.no_grad()
def generate_with_dual_cache(
model,
prompt,
steps=128,
gen_length=128,
block_length=128,
temperature=0.0,
remasking="low_confidence",
mask_id=126336,
threshold=None,
eos_token_id=None,
):
x = torch.full((1, prompt.shape[1] + gen_length), mask_id, dtype=torch.long).to(model.device)
x[:, :prompt.shape[1]] = prompt.clone()
prompt_length = prompt.shape[1]
assert gen_length % block_length == 0
num_blocks = gen_length // block_length
assert steps % num_blocks == 0
steps = steps // num_blocks
nfe = 0
for num_block in range(num_blocks):
# Early stop: skip remaining blocks if EOS already found
if eos_token_id is not None:
has_eos, _ = _check_early_stop(x, prompt_length, eos_token_id, mask_id)
if has_eos:
break
current_block_start = prompt.shape[1] + num_block * block_length
current_block_end = current_block_start + block_length
block_mask_index = x[:, current_block_start:current_block_end] == mask_id
num_transfer_tokens = get_num_transfer_tokens(block_mask_index, steps)
output = model(x, use_cache=True)
past_key_values = output.past_key_values
mask_index = x == mask_id
mask_index[:, current_block_end:] = 0
x0, transfer_index = get_transfer_index_entropy(
output.logits,
temperature,
remasking,
mask_index,
x,
entropy_threshold=threshold,
num_transfer_tokens=num_transfer_tokens[:, 0] if threshold is None else None,
)
x[transfer_index] = x0[transfer_index]
nfe += 1
i = 1
replace_position = torch.zeros_like(x, dtype=torch.bool)
replace_position[:, current_block_start:current_block_end] = 1
while True:
nfe += 1
mask_index = x[:, current_block_start:current_block_end] == mask_id
logits = model(
x[:, current_block_start:current_block_end],
past_key_values=past_key_values,
use_cache=True,
replace_position=replace_position,
).logits
x0, transfer_index = get_transfer_index_entropy(
logits,
temperature,
remasking,
mask_index,
x[:, current_block_start:current_block_end],
entropy_threshold=threshold,
num_transfer_tokens=num_transfer_tokens[:, i] if threshold is None else None,
)
x[:, current_block_start:current_block_end][transfer_index] = x0[transfer_index]
# Early stop: check after each decode step
if eos_token_id is not None:
has_eos, _ = _check_early_stop(x, prompt_length, eos_token_id, mask_id)
if has_eos:
break
if (x[:, current_block_start:current_block_end] == mask_id).sum() == 0:
break
i += 1
return x, nfe
@torch.no_grad()
def generate_multi_block(
model,
prompt,
steps=128,
max_new_tokens=512,
block_size=32,
temperature=0.0,
remasking="low_confidence",
mask_id=126336,
threshold=0.5,
block_add_threshold=0.5,
decoded_token_threshold=0.5,
eos_token_id=None,
):
"""
Pipelined parallel decoding without cache.
Args:
block_add_threshold: Add new block when last block progress >= this threshold.
Set to 1.0 for fully sequential processing (like generate()).
decoded_token_threshold: Block becomes fully activated when previous block progress >= this threshold.
Set to 1.0 for fully sequential processing (like generate()).
threshold: Entropy threshold for decoding (lower entropy = higher confidence).
Tokens with entropy > threshold are skipped.
Same semantics as in generate() method. Typical value: 0.5
When block_add_threshold=1.0 and decoded_token_threshold=1.0, this method behaves
identically to generate() with sequential block processing.
"""
x = torch.full((1, prompt.shape[1] + max_new_tokens), mask_id, dtype=torch.long).to(
model.device
)
x[:, : prompt.shape[1]] = prompt.clone()
prompt_length = prompt.shape[1]
# Track block states: {block_id: {start, end, mask_count, total_masks, is_complete}}
# Initialize with prompt block
block_states = {
0: {
"start": 0,
"end": prompt.shape[1],
"mask_count": 0,
"total_masks": prompt.shape[1],
"is_complete": True,
}
}
# Create first generation block
num_blocks = max_new_tokens // block_size
next_block_id = 1
if next_block_id <= num_blocks:
block_start = prompt.shape[1] + (next_block_id - 1) * block_size
block_end = min(block_start + block_size, prompt.shape[1] + max_new_tokens)
# First block should be immediately activated since prompt (block 0) is already complete
should_activate = 1.0 >= decoded_token_threshold # prompt progress is always 1.0
block_states[next_block_id] = {
"start": block_start,
"end": block_end,
"mask_count": block_end - block_start,
"total_masks": block_end - block_start,
"is_complete": should_activate,
}
next_block_id += 1
nfe = 0
while True:
# Check if all blocks are exhausted AND no more blocks to create
mask_index = x == mask_id
total_masks = mask_index[:, prompt_length:].sum()
if total_masks == 0 and next_block_id > num_blocks:
break
nfe += 1
# Early stop: check for EOS token
if eos_token_id is not None:
has_eos, first_eos_abs = _check_early_stop(x, prompt_length, eos_token_id, mask_id)
if has_eos:
# Create all remaining blocks after EOS and mark them as complete
while next_block_id <= num_blocks:
block_start = prompt_length + (next_block_id - 1) * block_size
block_end = min(block_start + block_size, prompt_length + max_new_tokens)
if block_start > first_eos_abs:
block_states[next_block_id] = {
"start": block_start, "end": block_end, "mask_count": 0,
"total_masks": block_end - block_start, "is_complete": True,
}
next_block_id += 1
else:
break
# Recalculate: if no masks remain, we're done
if (x == mask_id)[:, prompt_length:].sum() == 0:
break
# Update block completion states
def update_block_activation_states():
"""Update which blocks should be fully activated based on previous block progress."""
for bid in sorted(block_states.keys()):
if bid > 0 and not block_states[bid]["is_complete"]:
prev_progress = (
1
- block_states[bid - 1]["mask_count"]
/ block_states[bid - 1]["total_masks"]
)
if prev_progress >= decoded_token_threshold:
block_states[bid]["is_complete"] = True
update_block_activation_states()
# Add new block dynamically based on last block's progress
if next_block_id <= num_blocks:
last_bid = max(block_states.keys())
if last_bid > 0: # Not just prompt
last_progress = (
1
- block_states[last_bid]["mask_count"]
/ block_states[last_bid]["total_masks"]
)
# Create next block when:
# 1. Last block progress >= block_add_threshold (for parallel processing), OR
# 2. Last block is complete (mask_count == 0) for sequential processing
should_add_block = (last_progress >= block_add_threshold) or (block_states[last_bid]["mask_count"] == 0)
if should_add_block:
# Add next block
block_start = prompt.shape[1] + (next_block_id - 1) * block_size
block_end = min(block_start + block_size, prompt.shape[1] + max_new_tokens)
if block_end > block_start:
# Check how many positions in this block are actually masked
actual_mask_count = (x[:, block_start:block_end] == mask_id).sum().item()
# Determine if this block should be immediately activated
# Check if previous block is complete enough
prev_bid = next_block_id - 1
prev_progress = (
1 - block_states[prev_bid]["mask_count"] / block_states[prev_bid]["total_masks"]
)
should_activate = prev_progress >= decoded_token_threshold
block_states[next_block_id] = {
"start": block_start,
"end": block_end,
"mask_count": actual_mask_count,
"total_masks": block_end - block_start,
"is_complete": should_activate,
}
next_block_id += 1
# Forward pass: only process up to the last complete or semi-activated block
# Find the rightmost block that should be processed
rightmost_active_bid = 0
for bid in sorted(block_states.keys()):
if block_states[bid]["is_complete"] or block_states[bid]["mask_count"] > 0:
rightmost_active_bid = bid
if rightmost_active_bid == 0:
break
active_end = block_states[rightmost_active_bid]["end"]
# Always do forward pass on entire sequence (like generate() does)
logits = model(x).logits
# Mask out future blocks (positions after active_end) to prevent them from being decoded
mask_index_for_decode = mask_index.clone()
mask_index_for_decode[:, active_end:] = 0
# Decode using full-sequence approach (like generate())
x0, transfer_index = get_transfer_index_entropy(
logits,
temperature,
remasking,
mask_index_for_decode,
x,
entropy_threshold=threshold if threshold is not None else 999.0,
num_transfer_tokens=None,
)
# For fully activated blocks, ensure at least one token is decoded (guaranteed progress)
# Find the first fully activated block with masks
first_fully_activated_bid = None
for bid in sorted(block_states.keys()):
if bid > 0 and block_states[bid]["is_complete"] and block_states[bid]["mask_count"] > 0:
first_fully_activated_bid = bid
break
if first_fully_activated_bid is not None:
# Check if any token was decoded in this block
start, end = block_states[first_fully_activated_bid]["start"], block_states[first_fully_activated_bid]["end"]
block_transfer = transfer_index[:, start:end]
if not block_transfer.any():
# Force decode the lowest entropy token in this fully activated block
p = F.softmax(logits[:, start:end].to(torch.float64), dim=-1)
entropy = -torch.sum(p * torch.log(p + 1e-12), dim=-1)
block_mask = mask_index_for_decode[:, start:end]
entropy = torch.where(block_mask, entropy, torch.inf)
best_idx = entropy[0].argmin()
transfer_index[0, start + best_idx] = True
x0_resample = torch.argmax(logits[0, start + best_idx], dim=-1)
x0[0, start + best_idx] = x0_resample
# Apply the decoded tokens
x[transfer_index] = x0[transfer_index]
# Update block states based on which positions were decoded
for bid in sorted(block_states.keys()):
if bid > 0 and block_states[bid]["mask_count"] > 0:
start, end = block_states[bid]["start"], block_states[bid]["end"]
block_decoded = transfer_index[:, start:end].sum().item()
if block_decoded > 0:
block_states[bid]["mask_count"] -= block_decoded
if nfe > 10000:
break
return x, nfe
@register_model("llada_dist")
class LLaDAEvalHarness(LM):
def __init__(
self,
model_path="/home/u2025104115/LLada_sft/SAFT-LLaDA/checkpoint/save_checkpoints/SAFT-LLaDA-5k_data",
mask_id=126336,
max_length=1024,
batch_size=32,
mc_num=128,
is_check_greedy=True,
steps=1024,
gen_length=1024,
block_length=1024,
remasking="low_confidence",
device="cuda",
use_cache=False,
threshold=None,
save_dir=None,
stats_dir=None,
show_speed=False,
dual_cache=False,
multi_block=False,
block_add_threshold=0.5,
decoded_token_threshold=0.5,
early_stop=False,
task="null",
**kwargs,
):
super().__init__()
accelerator_kwargs = InitProcessGroupKwargs(timeout=timedelta(weeks=52))
accelerator = Accelerator(kwargs_handlers=[accelerator_kwargs])
if accelerator.num_processes > 1:
self.accelerator = accelerator
else:
self.accelerator = None
model_kwargs = {}
if self.accelerator is not None:
model_kwargs.update({"device_map": {"": f"{self.accelerator.device}"}})
config = AutoConfig.from_pretrained(model_path, trust_remote_code=True)
if hasattr(config, "flash_attention"):
config.flash_attention = True
self.model = LLaDAModelLM.from_pretrained(
model_path,
trust_remote_code=True,
torch_dtype=torch.bfloat16,
config=config,
**model_kwargs,
)
self.model.eval()
self.device = torch.device(device)
if self.accelerator is not None:
self.model = self.model.to(self.accelerator.device)
self.device = torch.device(f"{self.accelerator.device}")
self._rank = self.accelerator.local_process_index
self._world_size = self.accelerator.num_processes
else:
self.model = self.model.to(device)
self._rank = 0
self._world_size = 1
self.mask_id = mask_id
if hasattr(config, "mask_token_id") and config.mask_token_id is not None:
self.mask_id = config.mask_token_id
self.tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
self.mc_num = mc_num
self.batch_size = int(batch_size)
assert mc_num % self.batch_size == 0
self.sampling_eps = 0.0
self.max_length = max_length
self.is_check_greedy = is_check_greedy
self.steps = steps
self.gen_length = gen_length
self.block_length = block_length
self.remasking = remasking
self.use_cache = use_cache
self.threshold = threshold
# self.is_instruct = True if ("instruct" in model_path.lower() or "1.5" in model_path.lower()) else False
self.is_instruct = True
self.save_dir = save_dir
self.stats_dir = stats_dir if stats_dir is not None else save_dir
self.show_speed = show_speed
self.dual_cache = dual_cache
self.multi_block = multi_block
self.block_add_threshold = block_add_threshold
self.decoded_token_threshold = decoded_token_threshold
self.early_stop = early_stop
self.task = task
self.cfg = 0
@property
def rank(self):
return self._rank
@property
def world_size(self):
return self._world_size
def _forward_process(self, batch, prompt_index):
b, l = batch.shape
target_len = (l - prompt_index.sum()).item()
k = torch.randint(1, target_len + 1, (), device=batch.device)
x = torch.round(torch.linspace(float(k), k + (b - 1) * (target_len / b), steps=b, device=batch.device)).long()
x = ((x - 1) % target_len) + 1
assert x.min() >= 1 and x.max() <= target_len
indices = torch.arange(target_len, device=batch.device).repeat(b, 1)
is_mask = indices < x.unsqueeze(1)
for i in range(b):
is_mask[i] = is_mask[i][torch.randperm(target_len)]
is_mask = torch.cat((torch.zeros(b, prompt_index.sum(), dtype=torch.bool, device=batch.device), is_mask), dim=1)
noisy_batch = torch.where(is_mask, self.mask_id, batch)
return noisy_batch, (x / target_len).unsqueeze(1).repeat(1, l)
@torch.no_grad()
def get_logits(self, batch, prompt_index):
if self.cfg > 0.0:
assert len(prompt_index) == batch.shape[1]
prompt_index = prompt_index.unsqueeze(0).repeat(batch.shape[0], 1)
un_batch = batch.clone()
un_batch[prompt_index] = self.mask_id
batch = torch.cat([batch, un_batch])
logits = self.model(batch).logits
if self.cfg > 0.0:
logits, un_logits = torch.chunk(logits, 2, dim=0)
logits = un_logits + (self.cfg + 1) * (logits - un_logits)
return logits[:, : batch.shape[1]]
@torch.no_grad()
def get_loglikelihood(self, prefix, target):
seq = torch.concatenate([prefix, target])[None, :]
seq = seq.repeat((self.batch_size, 1)).to(self.device)
prompt_index = torch.arange(seq.shape[1], device=self.device) < len(prefix)
loss_acc = []
for _ in range(self.mc_num // self.batch_size):
perturbed_seq, p_mask = self._forward_process(seq, prompt_index)
mask_indices = perturbed_seq == self.mask_id
logits = self.get_logits(perturbed_seq, prompt_index)
loss = F.cross_entropy(logits[mask_indices], seq[mask_indices], reduction="none") / p_mask[mask_indices]
loss = loss.sum() / self.batch_size
loss_acc.append(loss.item())
return -sum(loss_acc) / len(loss_acc)
@torch.no_grad()
def suffix_greedy_prediction(self, prefix, target):
if not self.is_check_greedy:
return False
seq = torch.full((1, len(prefix) + len(target)), self.mask_id, device=self.device)
prompt_index = torch.arange(seq.shape[1], device=self.device) < len(prefix)
prefix, target = prefix.to(self.device), target.to(self.device)
seq[0, : len(prefix)] = prefix
for _ in range(len(target)):
mask_index = seq == self.mask_id
logits = self.get_logits(seq, prompt_index)[mask_index]
x0 = torch.argmax(logits, dim=-1)
p = torch.softmax(logits.to(torch.float32), dim=-1)
confidence = torch.gather(p, dim=-1, index=torch.unsqueeze(x0, -1)).squeeze(dim=-1)
_, index = torch.sort(confidence, descending=True)
x0[index[1:]] = self.mask_id
seq[mask_index] = x0.clone()
correct = target == seq[0, len(prefix) :]
correct = torch.all(correct)
return correct
def _encode_pair(self, context, continuation):
n_spaces = len(context) - len(context.rstrip())
if n_spaces > 0:
continuation = context[-n_spaces:] + continuation
context = context[:-n_spaces]
whole_enc = self.tokenizer(context + continuation)["input_ids"]
context_enc = self.tokenizer(context)["input_ids"]
context_enc_len = len(context_enc)
continuation_enc = whole_enc[context_enc_len:]
return context_enc, continuation_enc
def loglikelihood(self, requests):
def _tokenize(e):
prefix, target = self._encode_pair(e["prefix"], e["target"])
return {
"prefix_text": e["prefix"],
"target_text": e["target"],
"prefix": prefix,
"target": target,
}
ds = [{"prefix": req.args[0], "target": req.args[1]} for req in requests]
ds = Dataset.from_list(ds)
ds = ds.map(_tokenize)
ds = ds.with_format("torch")
prompt_len = [len(x["prefix"]) + len(x["target"]) for x in ds]
assert max(prompt_len) <= 4096
out = []
with torch.no_grad():
for elem in tqdm(ds, desc="Computing likelihood..."):
prefix = elem["prefix"]
target = elem["target"]
ll = self.get_loglikelihood(prefix, target)
is_target_greedy_dec = self.suffix_greedy_prediction(prefix, target)
out.append((ll, 1.0 if is_target_greedy_dec else 0.0))
torch.cuda.empty_cache()
return out
def loglikelihood_rolling(self, requests):
raise NotImplementedError
def generate_until(self, requests):
output = []
num_tokens = 0
num_nfe = 0
processed_count = 0
start_time = time.time()
log_fh = None
if self.save_dir is not None:
os.makedirs(self.save_dir, exist_ok=True)
rank = self.rank
save_path = os.path.join(self.save_dir, f"rank_{rank}.jsonl")
print(f"save_path: {save_path}")
if os.path.exists(save_path):
print(f"load from {save_path}")
with open(save_path, "r", encoding="utf-8") as f:
output = [json.loads(line) for line in f]
processed_count = len(output)
print(f"processed_count: {processed_count}")
if self.stats_dir is not None:
os.makedirs(self.stats_dir, exist_ok=True)
stats_samples_path = os.path.join(self.stats_dir, f"rank_{self.rank}_samples.jsonl")
log_fh = open(stats_samples_path, "a", encoding="utf-8")
for i, req in enumerate(tqdm(requests, desc="Generating...")):
sample_start_time = time.time()
if i < processed_count:
continue
question = req.args[0]
if self.is_instruct:
tail = r" Please reason step by step, and put your final answer within \boxed{}."
if self.task == "gsm8k":
m = [{"role": "user", "content": question + tail}]
elif self.task == "humaneval" or self.task == "humaneval_plus":
start = "Write a solution to the following problem and make sure that it passes the tests:\n```python\n{{ prompt }}\n```\n "
question = start.replace("{{ prompt }}", question)
m = [{"role": "user", "content": question}]
else:
m = [{"role": "user", "content": question}]
user_input = self.tokenizer.apply_chat_template(m, add_generation_prompt=True, tokenize=False)
input_ids = self.tokenizer(user_input)["input_ids"]
else:
user_input = question
input_ids = self.tokenizer(user_input)["input_ids"]
stop_tokens = list(req.args[1]["until"])
if "<|eot_id|>" not in stop_tokens:
stop_tokens.append("<|eot_id|>")
input_ids = torch.tensor(input_ids).to(self.device).unsqueeze(0)
# Determine eos_token_id for early stopping
eos_token_id = self.tokenizer.eos_token_id if self.early_stop else None
if self.multi_block:
generated_answer, nfe = generate_multi_block(
self.model,
input_ids,
steps=self.steps,
max_new_tokens=self.gen_length,
block_size=self.block_length,
temperature=0,
remasking=self.remasking,
mask_id=self.mask_id,
threshold=self.threshold,
block_add_threshold=self.block_add_threshold,
decoded_token_threshold=self.decoded_token_threshold,
eos_token_id=eos_token_id,
)
elif self.use_cache:
if self.dual_cache:
generated_answer, nfe = generate_with_dual_cache(
self.model,
input_ids,
steps=self.steps,
gen_length=self.gen_length,
block_length=self.block_length,
temperature=0,
remasking=self.remasking,
mask_id=self.mask_id,
threshold=self.threshold,
eos_token_id=eos_token_id,
)
else:
generated_answer, nfe = generate_with_prefix_cache(
self.model,
input_ids,
steps=self.steps,
gen_length=self.gen_length,
block_length=self.block_length,
temperature=0,
remasking=self.remasking,
mask_id=self.mask_id,
threshold=self.threshold,
eos_token_id=eos_token_id,
)
else:
generated_answer, nfe = generate(
self.model,
input_ids,
steps=self.steps,
gen_length=self.gen_length,
block_length=self.block_length,
temperature=0,
remasking=self.remasking,
mask_id=self.mask_id,
threshold=self.threshold,
eos_token_id=eos_token_id,
)
if self.is_instruct and "task_id" in req.doc and str(req.doc["task_id"]).lower().startswith("humaneval"):
generated_answer = self.tokenizer.decode(generated_answer[0][input_ids.shape[1] :], skip_special_tokens=True)
generated_answer_ids = self.tokenizer(generated_answer)["input_ids"]
else:
generated_answer = self.tokenizer.decode(generated_answer[0][input_ids.shape[1] :], skip_special_tokens=False)
print(f"stop_tokens: {stop_tokens}")
for stop_seq in stop_tokens:
if stop_seq in generated_answer:
generated_answer = generated_answer.split(stop_seq)[0]
generated_answer_ids = torch.tensor(self.tokenizer(generated_answer)["input_ids"])
generated_answer = self.tokenizer.decode(generated_answer_ids, skip_special_tokens=True)
num_tokens += len(generated_answer_ids)
num_nfe += nfe
output.append(generated_answer)
if self.save_dir is not None:
with open(save_path, "a", encoding="utf-8") as f:
f.write(json.dumps(generated_answer, ensure_ascii=False) + "\n")
sample_end_time = time.time()
if log_fh is not None:
record = {
"sample_index": int(i),
"prompt_tokens": int(input_ids.shape[1]),
"generated_tokens": int(len(generated_answer_ids)),
"steps": int(self.steps),
"nfe": int(nfe),
"latency_seconds": float(sample_end_time - sample_start_time),
"timestamp": float(sample_end_time),
}
log_fh.write(json.dumps(record, ensure_ascii=False) + "\n")
print("=" * 20)
print("question: ", question)
print("answer: ", generated_answer)
print("=" * 20, end="\n\n")
if log_fh is not None:
log_fh.close()
total_time = time.time() - start_time
# Compute blocks per sample for logging
blocks_per_sample = int(self.gen_length // self.block_length) if getattr(self, "block_length", 0) > 0 else 0
if self.stats_dir is not None:
processed_samples = int(len(output))
avg_iters_per_sample = (float(num_nfe) / float(processed_samples)) if processed_samples > 0 else 0.0
avg_gen_len = (float(num_tokens) / float(processed_samples)) if processed_samples > 0 else 0.0
avg_gen_len_blocks_persample = (float(self.gen_length) / float(self.block_length)) if getattr(self, "block_length", 0) > 0 else 0.0
final_stats = {
"processed_samples": processed_samples,
"total_samples": int(len(requests)),
"total_tokens": int(num_tokens),
"total_nfe": int(num_nfe),
"total_time": float(total_time),
"tokens_per_second": (float(num_tokens) / float(total_time)) if total_time > 0 else 0.0,
"nfe_per_token": (float(num_nfe) / float(num_tokens)) if num_tokens > 0 else 0.0,
"tokens_per_forward": (float(num_tokens)) / (float(num_nfe)),
"blocks_per_sample": blocks_per_sample,
"gen_length": int(self.gen_length),
"block_length": int(self.block_length),
"avg_iters_per_sample": avg_iters_per_sample,
"avg_gen_len": avg_gen_len,
"avg_gen_len_blocks_persample": avg_gen_len_blocks_persample,
"timestamp": time.time(),
}
stats_path = os.path.join(self.stats_dir, f"rank_{self.rank}_final_stats.json")
with open(stats_path, "w", encoding="utf-8") as f:
json.dump(final_stats, f, ensure_ascii=False, indent=2)
if self.show_speed:
print(f"Total time taken: {total_time} seconds")
print(f"Total NFE is {num_nfe}")
if self.accelerator is not None and getattr(self, "world_size", 1) > 1:
local_stats = torch.tensor(
[
float(len(output)),
float(num_tokens),
float(total_time),
float(num_nfe),
],
dtype=torch.float64,
device=self.device,
)
gathered_stats = self.accelerator.gather(local_stats)
if self.accelerator.is_local_main_process:
gathered_stats = gathered_stats.view(self.world_size, -1)
total_samples_all = int(gathered_stats[:, 0].sum().item())
total_tokens_all = int(gathered_stats[:, 1].sum().item())
sum_time_all = gathered_stats[:, 2].sum().item()
total_nfe_all = int(gathered_stats[:, 3].sum().item())
overall_tps = (total_tokens_all / sum_time_all) if sum_time_all > 0 else 0.0
avg_iters_per_sample_all = (float(total_nfe_all) / float(total_samples_all)) if total_samples_all > 0 else 0.0
avg_gen_len_all = (float(total_tokens_all) / float(total_samples_all)) if total_samples_all > 0 else 0.0
if self.stats_dir is not None:
aggregated_stats = {
"total_processed_samples": total_samples_all,
"total_generated_tokens": total_tokens_all,
"total_wall_time": sum_time_all,
"overall_tokens_per_second": overall_tps,
"overall_nfe": total_nfe_all,
"overall_nfe_per_token": (float(total_nfe_all) / float(total_tokens_all)) if total_tokens_all > 0 else 0.0,
"overall_tokens_per_forward": (float(total_tokens_all)) / (float(total_nfe_all)),
"avg_iters_per_sample": avg_iters_per_sample_all,
"avg_gen_len": avg_gen_len_all,
"blocks_per_sample": blocks_per_sample,
"gen_length": int(self.gen_length),
"block_length": int(self.block_length),
"timestamp": time.time(),
}
all_ranks_stats_path = os.path.join(self.stats_dir, "all_ranks_final_stats.json")
with open(all_ranks_stats_path, "w", encoding="utf-8") as f:
json.dump(aggregated_stats, f, ensure_ascii=False, indent=2)
return output
if __name__ == "__main__":
set_seed(1234)
cli_evaluate()
|