File size: 30,100 Bytes
b84549f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
import torch
from torch import nn
from copy import deepcopy
from abc import ABC, abstractmethod
from methods.elasticdnn.api.model import ElasticDNN_OfflineFMModel, ElasticDNN_OfflineMDModel
from utils.dl.common.model import get_model_device, get_model_latency, get_model_size, set_module, get_module
from utils.common.log import logger
from transformers import GPTNeoForCausalLM
from utils.dl.common.model import set_module
from torch import nn
import torch
from einops import rearrange, repeat
from einops.layers.torch import Rearrange
from utils.common.log import logger
from methods.elasticdnn.pipeline.offline.fm_to_md.base import FM_to_MD_Util
from methods.elasticdnn.pipeline.offline.fm_lora.base import FMLoRA_Util, LoRA
from methods.elasticdnn.model.base import Abs, KTakesAll, ElasticDNNUtil, Layer_WrappedWithFBS
import tqdm
from transformers import GPT2Tokenizer
import os
import nltk
from nltk.translate.bleu_score import sentence_bleu, corpus_bleu
from nltk.translate.bleu_score import SmoothingFunction
import json

def collate_fn(batch):
    dict = {}
    input_ids = []
    attention_mask = []
    token_type_ids = []
    labels = []
    return_dict = True
    lenli = []
    for item in batch:
        if len(item) == 1 or len(item['labels']) == 0:
            continue
        input_ids.append(item['input_ids'].unsqueeze(0))
        if 'attention_mask' in item.keys():
            attention_mask.append(item['attention_mask'].unsqueeze(0))
        if 'token_type_ids' in item.keys():
            token_type_ids.append(item['token_type_ids'].unsqueeze(0))
        labels.append(item['labels'].unsqueeze(0))
        if 'len' in item.keys():
            lenli.append(item['len'])

    dict['return_dict'] = batch[0]['return_dict']
    if len(input_ids) > 0:
        dict['input_ids'] = torch.cat(input_ids, dim=0)
    else:
        return {}, torch.Tensor([0])
    if len(attention_mask) > 0:
        dict['attention_mask'] = torch.cat(attention_mask, dim=0)
    if len(token_type_ids) > 0:
        dict['token_type_ids'] = torch.cat(token_type_ids, dim=0)
    dict['labels'] = torch.cat(labels, dim=0)
    if len(lenli) > 0:
        dict['len'] = lenli
    return dict, torch.Tensor([0])

class GPTNeoForNLG(nn.Module):
    def __init__(self, series_id):
        super(GPTNeoForNLG, self).__init__()
        
        # logger.info(f'init bert for sen cls (using {bert_model_tag})')
        self.model = GPTNeoForCausalLM.from_pretrained(f'experiments/elasticdnn/gpt_neo/{series_id}')
        self.config = self.model.config
        
        self.config.pad_token_id = self.config.eos_token_id
    
    def generate(self, x, pad_id=None):
        return self.model.generate(x['input_ids'], max_new_tokens=128, num_beams=3, early_stopping=True, pad_token_id=pad_id)
        
    def forward(self, **x):
        x['return_dict'] = True

        output = self.model(**x)
        
        if self.training == True:
            return output['loss']
        else:
            return output['logits']

class ToQKV_WrappedWithLoRA(nn.Module):
    def __init__(self, fc: nn.Linear, ab_r: int):
        super(ToQKV_WrappedWithLoRA, self).__init__()
        
        self.fc = fc
        self.ab = self.create_ab_as_linear(fc.weight.data, ab_r)
        
    def create_ab_as_linear(self, fc_weight: torch.Tensor, ab_r: int):
        res = nn.Sequential(
            LoRA(fc_weight.size(1), fc_weight.size(0) // ab_r, bias=False),
            LoRA(fc_weight.size(0) // ab_r, fc_weight.size(0), bias=False)
        ).to(fc_weight.device)
        nn.init.kaiming_uniform_(res[0].weight, a=5 ** 0.5)
        nn.init.zeros_(res[1].weight)
        return res
        
    def forward(self, x):
        x1 = self.fc(x)
        x2 = self.ab(x)
        return x1 + x2
    

class FMLoRA_GPT_Util(FMLoRA_Util):
    
    @torch.no_grad()
    def add_lora_ab_to_fm(self, fm: nn.Module, ab_r: int, samples: dict):
        fm.eval()
        
        o1 = fm(**samples)
        
        for name, module in fm.named_modules():
            if name.endswith(('k_proj', 'v_proj', 'q_proj')):
                set_module(fm, name, ToQKV_WrappedWithLoRA(module, ab_r))
        
        o2 = fm(**samples)
        
        if isinstance(o1, tuple):
            o1 = o1[-1]
            o2 = o2[-1]
        output_diff = ((o1 - o2) ** 2).sum()
        assert output_diff < 1e-5
        
        return fm
    
    @torch.no_grad()
    def absorb_lora_and_recover_net_structure(self, fm: nn.Module, samples: dict):       
        fm.eval()
        # print('absorb lora before')
        o1 = fm(**samples)
        
        for name, module in fm.named_modules():
            if not isinstance(module, ToQKV_WrappedWithLoRA):
                continue
            
            fc = module.fc
            ab = module.ab

            fc.weight.add_(ab[1].weight @ ab[0].weight)
            
            set_module(fm, name, fc)
        
        # print('absorb lora after')
        o2 = fm(**samples)
        
        if isinstance(o1, tuple):
            o1 = o1[-1]
            o2 = o2[-1]
        output_diff = ((o1 - o2) ** 2).sum()
        # assert output_diff < 1e-6, output_diff
        
        return fm

def getTokenizer():
    tokenizer = GPT2Tokenizer.from_pretrained(f'experiments/elasticdnn/gpt_neo/{os.environ["gpt_neo_series_id"]}', padding_side='left')
    # tokenizer.pad_token = tokenizer.eos_token
    tokenizer.sep_token = tokenizer.eos_token
    special_tokens = {"pad_token":"<|pad|>"}#, "sep_token":"<|sep|>", "bos_token":"<|bos|>"}
    tokenizer.add_special_tokens(special_tokens)
    tokenizer.pad_token = "<|pad|>"
    # tokenizer.bos_token = "<|bos|>"
    # tokenizer.sep_token = "<|sep|>"
    return tokenizer

class ElasticDNN_OfflineTextGenFMModel(ElasticDNN_OfflineFMModel):
    def get_accuracy(self, test_loader, *args, **kwargs):
        acc = 0
        sample_num = 0
        tokenizer = getTokenizer()
        self.to_eval_mode()
        pred_txt = []
        true_txt = []
        res = []
        with torch.no_grad():
            pbar = tqdm.tqdm(enumerate(test_loader), total=len(test_loader), dynamic_ncols=True, leave=False)
            for batch_index, (x, _) in pbar:
                if len(x) == 0:
                    continue
                # if batch_index > 10:
                #     break
                for k, v in x.items():
                    if isinstance(v, torch.Tensor):
                        x[k] = v.to(self.device)
                # input_ids = []
                inputlen = x['len']
                y = x['labels']
                x['labels'] = None
                outputs = self.models_dict['main'].generate(x, pad_id=tokenizer.pad_token_id)
                
                for i, op in enumerate(outputs):
                    op = op.tolist()
                    op = list(filter(lambda x: x != tokenizer.pad_token_id, op))
                    txt = tokenizer.decode(op)
                    txt = txt.replace(tokenizer.pad_token, "")
                    res.append(txt)
                    txt = txt[inputlen[i]:]
                    pred_txt.append(nltk.word_tokenize(txt))
                for tp in y:
                    true_txt.append(nltk.word_tokenize(tokenizer.decode(tp).replace(tokenizer.pad_token, '')))
                # pred = F.softmax(output, dim=1).argmax(dim=1)
                # correct = torch.eq(pred, y).sum().item()
                # acc += correct
                sample_num += len(y)
                
                # pbar.set_description(f'cur_batch_total: {len(y)}, cur_batch_correct: {correct}, '
                #                      f'cur_batch_acc: {(correct / len(y)):.4f}')
        json.dump(res, open("./gpt_generation.json", "w"))
        smooth = SmoothingFunction()
        score = 0.
        for pred, true in zip(pred_txt, true_txt):
            score += sentence_bleu([true], pred, weights=(0.25, 0.25, 0.25, 0.25), smoothing_function=smooth.method1)
        score /= sample_num
        return score
    
    def infer(self, x, *args, **kwargs):
        return self.models_dict['main'].forward(**x)

class FM_to_MD_GPT_Util(FM_to_MD_Util):
    def init_md_from_fm_by_reducing_width(self, fm: nn.Module, reducing_width_ratio: int, sparsity=0.0) -> nn.Module:
        # sparsity: It is mainly used to make a distilled model used in the baseline algorithm, and this parameter can ensure that the model has the same size as the model used in the online algorithm.
        fm_vit = deepcopy(fm)
        
        for block in fm_vit.model.transformer.h:
            tmp = get_module(block, 'attn.attention')
            tmp.head_dim = max(tmp.head_dim // reducing_width_ratio, 16)
            set_module(block, 'attn.attention', tmp)
        #ddw
        def _f(n):
            return int(n // reducing_width_ratio)
        
        # def _rand_indexes(n):
            # return torch.randperm(n)[0: int(n // reducing_width_ratio)]
        
        def l1_max_indexes_for_qkv(p: torch.Tensor, dim=0):
            assert dim in [0, 1]
            assert p.dim() in [1, 2, 4]
            
            if dim == 1:
                p = p.T
            
            p_norm = p.abs().contiguous().view(p.size(0), -1).sum(dim=1)
            n = p.size(0)
            return p_norm.argsort(descending=True)[0: max(256, int(n // reducing_width_ratio))].sort()[0]

        def l1_max_indexes(p: torch.Tensor, dim=0):
            assert dim in [0, 1]
            assert p.dim() in [1, 2, 4]
            
            if dim == 1:
                p = p.T
            
            p_norm = p.abs().contiguous().view(p.size(0), -1).sum(dim=1)
            n = p.size(0)
            return p_norm.argsort(descending=True)[0: int(n // reducing_width_ratio)].sort()[0]
        
        def l1_max_indexes_with_sparsity(p: torch.Tensor, dim=0):
            assert dim in [0, 1]
            assert p.dim() in [1, 2, 4]
            
            if dim == 1:
                p = p.T
            
            p_norm = p.abs().contiguous().view(p.size(0), -1).sum(dim=1)
            n = p.size(0)
            return p_norm.argsort(descending=True)[0: int(n // reducing_width_ratio * (1 - sparsity))].sort()[0]

        for block_i, block in enumerate(fm_vit.model.transformer.h):
            for k in ['k_proj', 'v_proj', 'q_proj']:
                qkv = get_module(block, f'attn.attention.{k}')

                new_qkv = nn.Linear(qkv.in_features, max(256, _f(qkv.out_features)), 
                                    qkv.bias is not None, qkv.weight.device)
                indexes = l1_max_indexes_for_qkv(qkv.weight.data, 0)
                
                new_qkv.weight.data.copy_(qkv.weight.data[indexes])
                if qkv.bias is not None:
                    new_qkv.bias.data.copy_(qkv.bias.data[indexes])
                set_module(block, f'attn.attention.{k}', new_qkv)
            
            proj = get_module(block, f'attn.attention.out_proj')
            new_proj = nn.Linear(max(256, _f(proj.in_features)), proj.out_features, 
                                proj.bias is not None, proj.weight.device)
            new_proj.weight.data.copy_(proj.weight.data[:, l1_max_indexes_for_qkv(proj.weight.data, 1)])
            if proj.bias is not None:
                new_proj.bias.data.copy_(proj.bias.data)
            set_module(block, f'attn.attention.out_proj', new_proj)
            
            fc1 = get_module(block, f'mlp.c_fc')
            new_fc1 = nn.Linear(fc1.in_features, int(_f(fc1.out_features) * (1 - sparsity)), 
                                fc1.bias is not None, fc1.weight.device)
            indexes = l1_max_indexes_with_sparsity(fc1.weight.data, 0)
            new_fc1.weight.data.copy_(fc1.weight.data[indexes])
            if fc1.bias is not None:
                new_fc1.bias.data.copy_(fc1.bias.data[indexes])
            set_module(block, f'mlp.c_fc', new_fc1)

            fc2 = get_module(block, f'mlp.c_proj')
            new_fc2 = nn.Linear(int(_f(fc2.in_features) * (1 - sparsity)), fc2.out_features, 
                                fc2.bias is not None, fc2.weight.device)
            new_fc2.weight.data.copy_(fc2.weight.data[:, l1_max_indexes_with_sparsity(fc2.weight.data, 1)])
            if fc2.bias is not None:
                new_fc2.bias.data.copy_(fc2.bias.data)
            set_module(block, f'mlp.c_proj', new_fc2)
            
        return fm_vit
    
    def init_md_from_fm_by_reducing_width_with_perf_test(self, fm: nn.Module, reducing_width_ratio: int,
                                                         samples: torch.Tensor) -> nn.Module:
        fm_size = get_model_size(fm, True)
        fm_latency = self._get_model_latency(fm, samples, 20, 
                                               get_model_device(fm), 20, False)
        
        master_dnn = self.init_md_from_fm_by_reducing_width(fm, reducing_width_ratio)
        master_dnn_size = get_model_size(master_dnn, True)
        logger.debug(f'inited master DNN: {master_dnn}')
        master_dnn_latency = self._get_model_latency(master_dnn, samples, 20, 
                                               get_model_device(master_dnn), 20, False)

        logger.info(f'init master DNN (w/o FBS yet) by reducing foundation model\'s width (by {reducing_width_ratio:d}x)')
        logger.info(f'foundation model ({fm_size:.3f}MB, {fm_latency:.4f}s/sample) -> '
                    f'master DNN ({master_dnn_size:.3f}MB, {master_dnn_latency:.4f}s/sample)\n'
                    f'(model size: ↓ {(fm_size / master_dnn_size):.2f}x, '
                    f'latency: ↓ {(fm_latency / master_dnn_latency):.2f}x)')
        
        return master_dnn
    
    def init_md_from_fm_by_reducing_layers(self, fm: nn.Module, layers: list) -> nn.Module:
        fm_vit = deepcopy(fm)
        tmp_h = []
        for block_i, block in enumerate(fm_vit.model.transformer.h):
            if block_i in layers:
                tmp_h.append(block)
        
        tmp_h = nn.ModuleList(tmp_h)
        set_module(fm_vit, f'model.transformer.h', tmp_h)
            
        return fm_vit
    
    def init_md_from_fm_by_reducing_layers_with_perf_test(self, fm: nn.Module, layers: list,
                                                         samples: torch.Tensor) -> nn.Module:
        fm_size = get_model_size(fm, True)
        fm_latency = self._get_model_latency(fm, samples, 20, 
                                               get_model_device(fm), 20, False)
        
        master_dnn = self.init_md_from_fm_by_reducing_layers(fm, layers)
        master_dnn_size = get_model_size(master_dnn, True)
        logger.debug(f'inited master DNN: {master_dnn}')
        master_dnn_latency = self._get_model_latency(master_dnn, samples, 20, 
                                               get_model_device(master_dnn), 20, False)

        logger.info(f'init master DNN (w/o FBS yet) by reducing foundation model\'s layers (to {len(layers):d} layers)')
        logger.info(f'foundation model ({fm_size:.3f}MB, {fm_latency:.4f}s/sample) -> '
                    f'master DNN ({master_dnn_size:.3f}MB, {master_dnn_latency:.4f}s/sample)\n'
                    f'(model size: ↓ {(fm_size / master_dnn_size):.2f}x, '
                    f'latency: ↓ {(fm_latency / master_dnn_latency):.2f}x)')
        
        return master_dnn

    def _get_model_latency(self, model: torch.nn.Module, model_input_size, sample_num: int, 
                           device: str, warmup_sample_num: int, return_detail=False):
        import time
        
        if isinstance(model_input_size, tuple):
            dummy_input = torch.rand(model_input_size).to(device)
        else:
            dummy_input = model_input_size
            
        model = model.to(device)
        model.eval()
        
        # warm up
        with torch.no_grad():
            for _ in range(warmup_sample_num):
                model(**dummy_input)
                
        infer_time_list = []
                
        if device == 'cuda' or 'cuda' in str(device):
            with torch.no_grad():
                for _ in range(sample_num):
                    s, e = torch.cuda.Event(enable_timing=True), torch.cuda.Event(enable_timing=True)
                    s.record()
                    model(**dummy_input)
                    e.record()
                    torch.cuda.synchronize()
                    cur_model_infer_time = s.elapsed_time(e) / 1000.
                    infer_time_list += [cur_model_infer_time]

        else:
            with torch.no_grad():
                for _ in range(sample_num):
                    start = time.time()
                    model(**dummy_input)
                    cur_model_infer_time = time.time() - start
                    infer_time_list += [cur_model_infer_time]
                    
        avg_infer_time = sum(infer_time_list) / sample_num

        if return_detail:
            return avg_infer_time, infer_time_list
        return avg_infer_time

class ElasticDNN_OfflineTextGenMDModel(ElasticDNN_OfflineMDModel):
    def get_accuracy(self, test_loader, *args, **kwargs):
        acc = 0
        sample_num = 0
        tokenizer = getTokenizer()
        self.to_eval_mode()
        pred_txt = []
        true_txt = []
        res = []
        with torch.no_grad():
            pbar = tqdm.tqdm(enumerate(test_loader), total=len(test_loader), dynamic_ncols=True, leave=False)
            for batch_index, (x, _) in pbar:
                if len(x) == 0:
                    continue
                # if batch_index > 10:
                #     break
                for k, v in x.items():
                    if isinstance(v, torch.Tensor):
                        x[k] = v.to(self.device)
                # input_ids = []
                inputlen = x['len']
                y = x['labels']
                x['labels'] = None
                outputs = self.models_dict['main'].generate(x, pad_id=tokenizer.pad_token_id)
                
                for i, op in enumerate(outputs):
                    op = op.tolist()
                    op = list(filter(lambda x: x != tokenizer.pad_token_id, op))
                    txt = tokenizer.decode(op)
                    txt = txt.replace(tokenizer.pad_token, "")
                    res.append(txt)
                    txt = txt[inputlen[i]:]
                    pred_txt.append(nltk.word_tokenize(txt))
                for tp in y:
                    true_txt.append(nltk.word_tokenize(tokenizer.decode(tp).replace(tokenizer.pad_token, '')))
                # pred = F.softmax(output, dim=1).argmax(dim=1)
                # correct = torch.eq(pred, y).sum().item()
                # acc += correct
                sample_num += len(y)
                
                # pbar.set_description(f'cur_batch_total: {len(y)}, cur_batch_correct: {correct}, '
                #                      f'cur_batch_acc: {(correct / len(y)):.4f}')
        json.dump(res, open("./gpt_generation.json", "w"))
        smooth = SmoothingFunction()
        score = 0.
        for pred, true in zip(pred_txt, true_txt):
            score += sentence_bleu([true], pred, weights=(0.25, 0.25, 0.25, 0.25), smoothing_function=smooth.method1)
        score /= sample_num
        return score
    
    def infer(self, x, *args, **kwargs):
        return self.models_dict['main'].forward(**x)

class SqueezeLast(nn.Module):
    def __init__(self):
        super(SqueezeLast, self).__init__()
    
    def forward(self, x):
        return x.squeeze(-1)

class StaticFBS(nn.Module):
    def __init__(self, static_channel_attention):
        super(StaticFBS, self).__init__()
        assert static_channel_attention.dim() == 2 and static_channel_attention.size(0) == 1
        self.static_channel_attention = nn.Parameter(static_channel_attention, requires_grad=False) # (1, dim)
        
    def forward(self, x):
        # print('staticfbs', x, self.static_channel_attention.unsqueeze(1))
        return x * self.static_channel_attention.unsqueeze(1)

class Linear_WrappedWithFBS(Layer_WrappedWithFBS):
    def __init__(self, linear: nn.Linear, r):
        super(Linear_WrappedWithFBS, self).__init__()
        
        self.linear = linear
        
        # for conv: (B, C_in, H, W) -> (B, C_in) -> (B, C_out)
        # for mlp in ViT: (B, #patches, D: dim of patches embedding) -> (B, D) -> (B, C_out)
        self.fbs = nn.Sequential(
            Rearrange('b n d -> b d n'),
            Abs(),
            nn.AdaptiveAvgPool1d(1),
            SqueezeLast(),
            nn.Linear(linear.in_features, max(linear.out_features // r, 36)),
            nn.ReLU(),
            nn.Linear(max(linear.out_features // r, 36), linear.out_features),
            nn.ReLU()
        )
        
        nn.init.constant_(self.fbs[6].bias, 1.)
        nn.init.kaiming_normal_(self.fbs[6].weight)
        
    
    def forward(self, x):
        if self.use_cached_channel_attention and self.cached_channel_attention is not None:
            channel_attention = self.cached_channel_attention
        else:
            self.cached_raw_channel_attention = self.fbs(x)
            self.cached_channel_attention = self.k_takes_all(self.cached_raw_channel_attention)
            
            channel_attention = self.cached_channel_attention
        
        raw_res = self.linear(x)
        
        return channel_attention.unsqueeze(1) * raw_res

class ElasticGPTUtil(ElasticDNNUtil):
    def convert_raw_dnn_to_master_dnn(self, raw_dnn: nn.Module, r: float, ignore_layers=[]):
        assert len(ignore_layers) == 0, 'not supported yet'

        raw_vit = deepcopy(raw_dnn)
        
        # set_module(module, 'patch_embed.proj', ProjConv_WrappedWithFBS(module.patch_embed.proj, r))
                
        for name, module in raw_vit.named_modules():
            # if name.endswith('attn'):
            #     set_module(module, 'qkv', ToQKV_WrappedWithFBS(module.qkv, r))
            if name.endswith('intermediate'):
                set_module(module, 'dense', Linear_WrappedWithFBS(module.dense, r))
            elif name.endswith('mlp'):
                set_module(module, 'c_fc', Linear_WrappedWithFBS(module.c_fc, r))


        return raw_vit
    
    def set_master_dnn_sparsity(self, master_dnn: nn.Module, sparsity: float):
        # for name, module in master_dnn.named_modules():
        #     if not name.endswith('attn'):
        #         continue
            
        #     q_features = module.qkv.to_qk.out_features // 2
            
        #     if (q_features - int(q_features * sparsity)) % module.num_heads != 0:
        #         # tune sparsity to ensure #unpruned channel % num_heads == 0
        #         # so that the pruning seems to reduce the dim_head of each head
        #         tuned_sparsity = 1. - int((q_features - int(q_features * sparsity)) / module.num_heads) * module.num_heads / q_features
        #         logger.debug(f'tune sparsity from {sparsity:.2f} to {tuned_sparsity}')
        #         sparsity = tuned_sparsity
        #         break
        
        return super().set_master_dnn_sparsity(master_dnn, sparsity)
    
    def select_most_rep_sample(self, master_dnn: nn.Module, samples: torch.Tensor):
        # print(samples)
        # return samples[0].unsqueeze(0)
        ret_di = samples['return_dict']
        del samples['return_dict']
        res = {k: v[0: 1] for k, v in samples.items()}
        res['return_dict'] = ret_di
        return res
        
    def extract_surrogate_dnn_via_samples(self, master_dnn: nn.Module, samples: torch.Tensor, return_detail=False):
        sample = self.select_most_rep_sample(master_dnn, samples)
        # assert sample.dim() == 4 and sample.size(0) == 1
        
        # print('before')
        master_dnn.eval()
        self.clear_cached_channel_attention_in_master_dnn(master_dnn)
        with torch.no_grad():
            master_dnn_output = master_dnn(**sample)
            
        # print('after')
        
        boosted_vit = deepcopy(master_dnn)
        
        def get_unpruned_indexes_from_channel_attn(channel_attn: torch.Tensor, k):
            assert channel_attn.size(0) == 1, 'use A representative sample to generate channel attentions'
            
            # print('attn_in_unpruned', channel_attn[0][0: 10])
            
            res = channel_attn[0].nonzero(as_tuple=True)[0] # should be one-dim

            # res = channel_attn[0].argsort(descending=True)[0: -int(channel_attn.size(1) * k)].sort()[0]
            
            # g = channel_attn
            # k = g.size(1) - int(g.size(1) * k)
            # res = g.topk(k, 1)[1][0].sort()[0]
            
            return res
        
        unpruned_indexes_of_layers = {}
        
        # for attn, ff in boosted_vit.transformer.layers:
        # for block_i, block in enumerate(boosted_vit.blocks):
        for block_i, block in enumerate(boosted_vit.model.transformer.h):
            # attn = block.attn
            # ff = block.mlp
            
            ff_0 = get_module(block, f'mlp.c_fc')
            # ff_0_unpruned_indexes = get_unpruned_indexes_from_channel_attn(ff_0.cached_channel_attention, k)
            ff_0_pruned_indexes = ff_0.k_takes_all.cached_i[0].sort()[0]
            ff_0_unpruned_indexes = torch.LongTensor([ii for ii in range(ff_0.cached_channel_attention.size(1)) if ii not in ff_0_pruned_indexes])
            new_ff_0 = nn.Linear(ff_0.linear.in_features, ff_0_unpruned_indexes.size(0), ff_0.linear.bias is not None)
            new_ff_0.weight.data.copy_(ff_0.linear.weight.data[ff_0_unpruned_indexes])
            if ff_0.linear.bias is not None:
                new_ff_0.bias.data.copy_(ff_0.linear.bias.data[ff_0_unpruned_indexes])
            set_module(block, 'mlp.c_fc', nn.Sequential(new_ff_0, StaticFBS(ff_0.cached_channel_attention[:, ff_0_unpruned_indexes])))
            
            ff_1 = get_module(block, f'mlp.c_proj')
            new_ff_1 = nn.Linear(ff_0_unpruned_indexes.size(0), ff_1.out_features, ff_1.bias is not None)
            new_ff_1.weight.data.copy_(ff_1.weight.data[:, ff_0_unpruned_indexes])
            if ff_1.bias is not None:
                new_ff_1.bias.data.copy_(ff_1.bias.data)
            set_module(block, 'mlp.c_proj', new_ff_1)
            
            unpruned_indexes_of_layers[f'model.transformer.h.{block_i}.mlp.c_fc.0.weight'] = ff_0_unpruned_indexes
        
        surrogate_dnn = boosted_vit
        surrogate_dnn.eval()
        surrogate_dnn = surrogate_dnn.to(get_model_device(master_dnn))
        # logger.debug(surrogate_dnn)
        with torch.no_grad():
            surrogate_dnn_output = surrogate_dnn(**sample)
            
        output_diff = ((surrogate_dnn_output - master_dnn_output) ** 2).sum()
        # assert output_diff < 1e-4, output_diff
        logger.info(f'output diff of master and surrogate DNN: {output_diff}')
        logger.debug(f'example output of master/surrogate: {master_dnn_output.sum(0)[0: 10]}, {surrogate_dnn_output.sum(0)[0: 10]}')
        # logger.info(f'\nonly prune mlp!!!!\n')
        # logger.info(f'\nonly prune mlp!!!!\n')
        
        if return_detail:
            return boosted_vit, unpruned_indexes_of_layers
        
        return boosted_vit
    
    def extract_surrogate_dnn_via_samples_with_perf_test(self, master_dnn: nn.Module, samples: torch.Tensor, return_detail=False):
        master_dnn_size = get_model_size(master_dnn, True)
        master_dnn_latency = self._get_model_latency(master_dnn, samples, 50, 
                                               get_model_device(master_dnn), 50, False)
        
        res = self.extract_surrogate_dnn_via_samples(master_dnn, samples, return_detail)
        if not return_detail:
            surrogate_dnn = res
        else:
            surrogate_dnn, unpruned_indexes_of_layers = res
        surrogate_dnn_size = get_model_size(surrogate_dnn, True)
        surrogate_dnn_latency = self._get_model_latency(master_dnn, samples, 50, 
                                               get_model_device(master_dnn), 50, False)

        logger.info(f'master DNN ({master_dnn_size:.3f}MB, {master_dnn_latency:.4f}s/sample) -> '
                    f'surrogate DNN ({surrogate_dnn_size:.3f}MB, {surrogate_dnn_latency:.4f}s/sample)\n'
                    f'(model size: ↓ {(master_dnn_size / surrogate_dnn_size):.2f}x, '
                    f'latency: ↓ {(master_dnn_latency / surrogate_dnn_latency):.2f}x)')
        
        return res
    
    def _get_model_latency(self, model: torch.nn.Module, model_input_size, sample_num: int, 
                           device: str, warmup_sample_num: int, return_detail=False):
        import time
        
        if isinstance(model_input_size, tuple):
            dummy_input = torch.rand(model_input_size).to(device)
        else:
            dummy_input = model_input_size
            
        model = model.to(device)
        model.eval()
        
        # warm up
        with torch.no_grad():
            for _ in range(warmup_sample_num):
                model(**dummy_input)
                
        infer_time_list = []
                
        if device == 'cuda' or 'cuda' in str(device):
            with torch.no_grad():
                for _ in range(sample_num):
                    s, e = torch.cuda.Event(enable_timing=True), torch.cuda.Event(enable_timing=True)
                    s.record()
                    model(**dummy_input)
                    e.record()
                    torch.cuda.synchronize()
                    cur_model_infer_time = s.elapsed_time(e) / 1000.
                    infer_time_list += [cur_model_infer_time]

        else:
            with torch.no_grad():
                for _ in range(sample_num):
                    start = time.time()
                    model(**dummy_input)
                    cur_model_infer_time = time.time() - start
                    infer_time_list += [cur_model_infer_time]
                    
        avg_infer_time = sum(infer_time_list) / sample_num

        if return_detail:
            return avg_infer_time, infer_time_list
        return avg_infer_time