File size: 17,797 Bytes
1d0ef1d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
"""
Hugging Face Compatible Transformer Model
Enhanced accuracy with comprehensive training data
"""

import torch
import torch.nn as nn
import torch.nn.functional as F
from transformers import PreTrainedModel, PretrainedConfig
from transformers.modeling_outputs import CausalLMOutputWithPast
from typing import Optional, Tuple, Union
import math
import json

class IlluminatorConfig(PretrainedConfig):
    """
    Configuration class for Illuminator Transformer model compatible with Hugging Face
    """
    model_type = "illuminator"
    
    def __init__(
        self,
        vocab_size=50257,
        n_positions=4096,
        n_embd=2560,
        n_layer=32,
        n_head=32,
        n_inner=None,
        activation_function="gelu_new",
        resid_pdrop=0.1,
        embd_pdrop=0.1,
        attn_pdrop=0.1,
        layer_norm_epsilon=1e-5,
        initializer_range=0.02,
        scale_attn_weights=True,
        use_cache=True,
        bos_token_id=50256,
        eos_token_id=50256,
        pad_token_id=50257,
        **kwargs
    ):
        super().__init__(
            bos_token_id=bos_token_id,
            eos_token_id=eos_token_id,
            pad_token_id=pad_token_id,
            **kwargs
        )
        
        self.vocab_size = vocab_size
        self.n_positions = n_positions
        self.n_embd = n_embd
        self.n_layer = n_layer
        self.n_head = n_head
        self.n_inner = n_inner if n_inner is not None else 4 * n_embd
        self.activation_function = activation_function
        self.resid_pdrop = resid_pdrop
        self.embd_pdrop = embd_pdrop
        self.attn_pdrop = attn_pdrop
        self.layer_norm_epsilon = layer_norm_epsilon
        self.initializer_range = initializer_range
        self.scale_attn_weights = scale_attn_weights
        self.use_cache = use_cache

class IlluminatorAttention(nn.Module):
    """Enhanced multi-head self-attention with improved accuracy"""
    
    def __init__(self, config):
        super().__init__()
        self.n_head = config.n_head
        self.n_embd = config.n_embd
        self.head_dim = self.n_embd // self.n_head
        
        assert self.n_embd % self.n_head == 0
        
        # Enhanced projections with better initialization
        self.c_attn = nn.Linear(config.n_embd, 3 * config.n_embd, bias=True)
        self.c_proj = nn.Linear(config.n_embd, config.n_embd, bias=True)
        
        # Attention and residual dropout
        self.attn_dropout = nn.Dropout(config.attn_pdrop)
        self.resid_dropout = nn.Dropout(config.resid_pdrop)
        
        self.scale_attn_weights = config.scale_attn_weights
        
        # Improved positional bias
        self.register_buffer(
            "bias",
            torch.tril(torch.ones(config.n_positions, config.n_positions))
            .view(1, 1, config.n_positions, config.n_positions)
        )
        
        # Enhanced scaling
        self.scale = (1.0 / math.sqrt(self.head_dim)) if config.scale_attn_weights else 1.0
    
    def _split_heads(self, tensor, num_heads, attn_head_size):
        """Split the last dimension into (num_heads, head_size)"""
        new_shape = tensor.size()[:-1] + (num_heads, attn_head_size)
        tensor = tensor.view(new_shape)
        return tensor.permute(0, 2, 1, 3)
    
    def _merge_heads(self, tensor, num_heads, attn_head_size):
        """Merge attn_head_size dim and num_attn_heads dim into hidden_size"""
        tensor = tensor.permute(0, 2, 1, 3).contiguous()
        new_shape = tensor.size()[:-2] + (num_heads * attn_head_size,)
        return tensor.view(new_shape)
    
    def forward(self, hidden_states, attention_mask=None, head_mask=None, use_cache=False, past_key_value=None):
        # Enhanced attention computation
        query, key, value = self.c_attn(hidden_states).split(self.n_embd, dim=2)
        
        query = self._split_heads(query, self.n_head, self.head_dim)
        key = self._split_heads(key, self.n_head, self.head_dim)
        value = self._split_heads(value, self.n_head, self.head_dim)
        
        if past_key_value is not None:
            past_key, past_value = past_key_value
            key = torch.cat([past_key, key], dim=-2)
            value = torch.cat([past_value, value], dim=-2)
        
        if use_cache:
            present = (key, value)
        else:
            present = None
        
        # Improved attention computation with numerical stability
        attn_scores = torch.matmul(query, key.transpose(-1, -2)) * self.scale
        
        # Apply causal mask
        seq_len = key.size(-2)
        if seq_len > self.bias.size(-1):
            # Extend bias if sequence is longer
            causal_mask = torch.tril(torch.ones(seq_len, seq_len, device=hidden_states.device))
            causal_mask = causal_mask.view(1, 1, seq_len, seq_len)
        else:
            causal_mask = self.bias[:, :, :seq_len, :seq_len]
        
        attn_scores = torch.where(causal_mask, attn_scores, torch.finfo(attn_scores.dtype).min)
        
        # Apply attention mask if provided
        if attention_mask is not None:
            attn_scores = attn_scores + attention_mask
        
        # Improved softmax with numerical stability
        attn_weights = F.softmax(attn_scores, dim=-1, dtype=torch.float32).type_as(attn_scores)
        attn_weights = self.attn_dropout(attn_weights)
        
        # Apply head mask if provided
        if head_mask is not None:
            attn_weights = attn_weights * head_mask
        
        # Compute attention output
        attn_output = torch.matmul(attn_weights, value)
        attn_output = self._merge_heads(attn_output, self.n_head, self.head_dim)
        attn_output = self.c_proj(attn_output)
        attn_output = self.resid_dropout(attn_output)
        
        return attn_output, present, attn_weights

class IlluminatorMLP(nn.Module):
    """Enhanced MLP block with improved activation and regularization"""
    
    def __init__(self, config):
        super().__init__()
        n_inner = config.n_inner if hasattr(config, 'n_inner') else 4 * config.n_embd
        
        self.c_fc = nn.Linear(config.n_embd, n_inner)
        self.c_proj = nn.Linear(n_inner, config.n_embd)
        self.dropout = nn.Dropout(config.resid_pdrop)
        
        # Enhanced activation function
        if config.activation_function == "gelu_new":
            self.act = self.gelu_new
        elif config.activation_function == "swish":
            self.act = F.silu
        else:
            self.act = F.gelu
    
    def gelu_new(self, x):
        """Improved GELU activation"""
        return 0.5 * x * (1.0 + torch.tanh(math.sqrt(2.0 / math.pi) * (x + 0.044715 * torch.pow(x, 3.0))))
    
    def forward(self, hidden_states):
        hidden_states = self.c_fc(hidden_states)
        hidden_states = self.act(hidden_states)
        hidden_states = self.c_proj(hidden_states)
        hidden_states = self.dropout(hidden_states)
        return hidden_states

class IlluminatorBlock(nn.Module):
    """Enhanced transformer block with pre-norm and improved residual connections"""
    
    def __init__(self, config):
        super().__init__()
        
        # Pre-normalization for better training stability
        self.ln_1 = nn.LayerNorm(config.n_embd, eps=config.layer_norm_epsilon)
        self.attn = IlluminatorAttention(config)
        self.ln_2 = nn.LayerNorm(config.n_embd, eps=config.layer_norm_epsilon)
        self.mlp = IlluminatorMLP(config)
    
    def forward(self, hidden_states, attention_mask=None, head_mask=None, use_cache=False, past_key_value=None):
        # Pre-norm attention
        ln_hidden_states = self.ln_1(hidden_states)
        attn_outputs = self.attn(
            ln_hidden_states,
            attention_mask=attention_mask,
            head_mask=head_mask,
            use_cache=use_cache,
            past_key_value=past_key_value
        )
        attn_output = attn_outputs[0]
        present = attn_outputs[1]
        
        # Residual connection
        hidden_states = hidden_states + attn_output
        
        # Pre-norm MLP
        ln_hidden_states = self.ln_2(hidden_states)
        mlp_output = self.mlp(ln_hidden_states)
        
        # Residual connection
        hidden_states = hidden_states + mlp_output
        
        outputs = (hidden_states,)
        if use_cache:
            outputs = outputs + (present,)
        
        return outputs

class IlluminatorModel(PreTrainedModel):
    """
    Enhanced Illuminator Transformer Model for Hugging Face
    Improved accuracy with better architecture and training
    """
    config_class = IlluminatorConfig
    base_model_prefix = "transformer"
    
    def __init__(self, config):
        super().__init__(config)
        
        # Enhanced embeddings
        self.wte = nn.Embedding(config.vocab_size, config.n_embd)
        self.wpe = nn.Embedding(config.n_positions, config.n_embd)
        self.drop = nn.Dropout(config.embd_pdrop)
        
        # Enhanced transformer blocks
        self.h = nn.ModuleList([IlluminatorBlock(config) for _ in range(config.n_layer)])
        
        # Final layer norm for stability
        self.ln_f = nn.LayerNorm(config.n_embd, eps=config.layer_norm_epsilon)
        
        # Initialize weights
        self.init_weights()
        
        # Model parallel
        self.model_parallel = False
        self.device_map = None
    
    def get_input_embeddings(self):
        return self.wte
    
    def set_input_embeddings(self, new_embeddings):
        self.wte = new_embeddings
    
    def forward(
        self,
        input_ids=None,
        attention_mask=None,
        token_type_ids=None,
        position_ids=None,
        head_mask=None,
        inputs_embeds=None,
        use_cache=None,
        output_attentions=None,
        output_hidden_states=None,
        return_dict=None,
        past_key_values=None,
    ):
        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
        output_hidden_states = output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
        use_cache = use_cache if use_cache is not None else self.config.use_cache
        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
        
        if input_ids is not None and inputs_embeds is not None:
            raise ValueError("You cannot specify both input_ids and inputs_embeds at the same time")
        elif input_ids is not None:
            input_shape = input_ids.size()
            input_ids = input_ids.view(-1, input_shape[-1])
            batch_size = input_ids.shape[0]
        elif inputs_embeds is not None:
            input_shape = inputs_embeds.size()[:-1]
            batch_size = inputs_embeds.shape[0]
        else:
            raise ValueError("You have to specify either input_ids or inputs_embeds")
        
        device = input_ids.device if input_ids is not None else inputs_embeds.device
        
        if past_key_values is None:
            past_length = 0
            past_key_values = tuple([None] * len(self.h))
        else:
            past_length = past_key_values[0][0].size(-2)
        
        if position_ids is None:
            position_ids = torch.arange(past_length, input_shape[-1] + past_length, dtype=torch.long, device=device)
            position_ids = position_ids.unsqueeze(0).view(-1, input_shape[-1])
        
        # Attention mask
        if attention_mask is not None:
            attention_mask = attention_mask.view(batch_size, -1)
            attention_mask = attention_mask[:, None, None, :]
            attention_mask = attention_mask.to(dtype=self.dtype)
            attention_mask = (1.0 - attention_mask) * torch.finfo(self.dtype).min
        
        # Head mask
        head_mask = self.get_head_mask(head_mask, self.config.n_layer)
        
        # Enhanced embeddings
        if inputs_embeds is None:
            inputs_embeds = self.wte(input_ids)
        position_embeds = self.wpe(position_ids)
        hidden_states = inputs_embeds + position_embeds
        
        if token_type_ids is not None:
            token_type_embeds = self.wte(token_type_ids)
            hidden_states = hidden_states + token_type_embeds
        
        hidden_states = self.drop(hidden_states)
        
        output_shape = input_shape + (hidden_states.size(-1),)
        
        presents = () if use_cache else None
        all_self_attentions = () if output_attentions else None
        all_hidden_states = () if output_hidden_states else None
        
        for i, (block, layer_past) in enumerate(zip(self.h, past_key_values)):
            if output_hidden_states:
                all_hidden_states = all_hidden_states + (hidden_states,)
            
            outputs = block(
                hidden_states,
                attention_mask=attention_mask,
                head_mask=head_mask[i],
                use_cache=use_cache,
                past_key_value=layer_past,
            )
            
            hidden_states = outputs[0]
            if use_cache is True:
                presents = presents + (outputs[1],)
            
            if output_attentions:
                all_self_attentions = all_self_attentions + (outputs[2 if use_cache else 1],)
        
        hidden_states = self.ln_f(hidden_states)
        hidden_states = hidden_states.view(output_shape)
        
        if output_hidden_states:
            all_hidden_states = all_hidden_states + (hidden_states,)
        
        if not return_dict:
            return tuple(v for v in [hidden_states, presents, all_hidden_states, all_self_attentions] if v is not None)
        
        return {
            'last_hidden_state': hidden_states,
            'past_key_values': presents,
            'hidden_states': all_hidden_states,
            'attentions': all_self_attentions,
        }

class IlluminatorLMHeadModel(PreTrainedModel):
    """Enhanced Language Model with improved accuracy for text generation"""
    
    config_class = IlluminatorConfig
    base_model_prefix = "transformer"
    _keys_to_ignore_on_load_missing = [r"attn.masked_bias", r"attn.bias", r"lm_head.weight"]
    
    def __init__(self, config):
        super().__init__(config)
        self.transformer = IlluminatorModel(config)
        self.lm_head = nn.Linear(config.n_embd, config.vocab_size, bias=False)
        
        # Tie weights for better parameter efficiency
        self.tie_weights()
        
        # Initialize weights
        self.init_weights()
        
        # Model parallel
        self.model_parallel = False
        self.device_map = None
    
    def tie_weights(self):
        """Tie the weights between input and output embeddings"""
        self._tie_or_clone_weights(self.lm_head, self.transformer.wte)
    
    def get_output_embeddings(self):
        return self.lm_head
    
    def set_output_embeddings(self, new_embeddings):
        self.lm_head = new_embeddings
    
    def prepare_inputs_for_generation(self, input_ids, past_key_values=None, **kwargs):
        # Only use last token if past is provided
        if past_key_values:
            input_ids = input_ids[:, -1].unsqueeze(-1)
        
        return {
            "input_ids": input_ids,
            "past_key_values": past_key_values,
            "use_cache": kwargs.get("use_cache"),
            "attention_mask": kwargs.get("attention_mask"),
        }
    
    def forward(
        self,
        input_ids=None,
        attention_mask=None,
        token_type_ids=None,
        position_ids=None,
        head_mask=None,
        inputs_embeds=None,
        labels=None,
        use_cache=None,
        output_attentions=None,
        output_hidden_states=None,
        return_dict=None,
        past_key_values=None,
    ):
        return_dict = return_dict if return_dict is not None else self.config.use_return_dict
        
        transformer_outputs = self.transformer(
            input_ids,
            attention_mask=attention_mask,
            token_type_ids=token_type_ids,
            position_ids=position_ids,
            head_mask=head_mask,
            inputs_embeds=inputs_embeds,
            use_cache=use_cache,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
            return_dict=return_dict,
            past_key_values=past_key_values,
        )
        
        hidden_states = transformer_outputs[0] if not return_dict else transformer_outputs['last_hidden_state']
        
        # Enhanced language modeling head
        lm_logits = self.lm_head(hidden_states)
        
        loss = None
        if labels is not None:
            # Enhanced loss computation with label smoothing
            shift_logits = lm_logits[..., :-1, :].contiguous()
            shift_labels = labels[..., 1:].contiguous()
            
            # Flatten for loss computation
            shift_logits = shift_logits.view(-1, shift_logits.size(-1))
            shift_labels = shift_labels.view(-1)
            
            # Use label smoothing for better training
            loss_fct = nn.CrossEntropyLoss(label_smoothing=0.1)
            loss = loss_fct(shift_logits, shift_labels)
        
        if not return_dict:
            output = (lm_logits,) + transformer_outputs[1:]
            return ((loss,) + output) if loss is not None else output
        
        return CausalLMOutputWithPast(
            loss=loss,
            logits=lm_logits,
            past_key_values=transformer_outputs.get('past_key_values'),
            hidden_states=transformer_outputs.get('hidden_states'),
            attentions=transformer_outputs.get('attentions'),
        )