cortexelus commited on
Commit
70aafed
·
verified ·
1 Parent(s): fc19859

cpu-amx: add dit_medium_bf16_core_manifest.txt (bf16 DiT engine)

Browse files
cpu-amx/dit_medium_bf16_core_manifest.txt ADDED
@@ -0,0 +1,434 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ b0.pre_g f32 0 1536 1536
2
+ b0.ff_g f32 6144 1536 1536
3
+ b0.cx_g f32 12288 1536 1536
4
+ b0.s_qkv.wt bf16 18432 11796480 1536 7680
5
+ b0.s_out.wt bf16 23611392 2359296 1536 1536
6
+ b0.s_qn f32 28329984 64 64
7
+ b0.s_kn f32 28330240 64 64
8
+ b0.c_q.wt bf16 28330496 4718592 1536 3072
9
+ b0.c_kv.wt bf16 37767680 7077888 1536 4608
10
+ b0.c_out.wt bf16 51923456 2359296 1536 1536
11
+ b0.c_qn f32 56642048 64 64
12
+ b0.c_kn f32 56642304 64 64
13
+ b0.ff0.wt bf16 56642560 18874368 1536 12288
14
+ b0.ff0.bias f32 94391296 12288 12288
15
+ b0.ff2.wt bf16 94440448 9437184 6144 1536
16
+ b0.ff2.bias f32 113314816 1536 1536
17
+ b0.ssg f32 113320960 9216 9216
18
+ b0.local f32 113357824 1536 1536
19
+ b1.pre_g f32 113363968 1536 1536
20
+ b1.ff_g f32 113370112 1536 1536
21
+ b1.cx_g f32 113376256 1536 1536
22
+ b1.s_qkv.wt bf16 113382400 11796480 1536 7680
23
+ b1.s_out.wt bf16 136975360 2359296 1536 1536
24
+ b1.s_qn f32 141693952 64 64
25
+ b1.s_kn f32 141694208 64 64
26
+ b1.c_q.wt bf16 141694464 4718592 1536 3072
27
+ b1.c_kv.wt bf16 151131648 7077888 1536 4608
28
+ b1.c_out.wt bf16 165287424 2359296 1536 1536
29
+ b1.c_qn f32 170006016 64 64
30
+ b1.c_kn f32 170006272 64 64
31
+ b1.ff0.wt bf16 170006528 18874368 1536 12288
32
+ b1.ff0.bias f32 207755264 12288 12288
33
+ b1.ff2.wt bf16 207804416 9437184 6144 1536
34
+ b1.ff2.bias f32 226678784 1536 1536
35
+ b1.ssg f32 226684928 9216 9216
36
+ b1.local f32 226721792 1536 1536
37
+ b2.pre_g f32 226727936 1536 1536
38
+ b2.ff_g f32 226734080 1536 1536
39
+ b2.cx_g f32 226740224 1536 1536
40
+ b2.s_qkv.wt bf16 226746368 11796480 1536 7680
41
+ b2.s_out.wt bf16 250339328 2359296 1536 1536
42
+ b2.s_qn f32 255057920 64 64
43
+ b2.s_kn f32 255058176 64 64
44
+ b2.c_q.wt bf16 255058432 4718592 1536 3072
45
+ b2.c_kv.wt bf16 264495616 7077888 1536 4608
46
+ b2.c_out.wt bf16 278651392 2359296 1536 1536
47
+ b2.c_qn f32 283369984 64 64
48
+ b2.c_kn f32 283370240 64 64
49
+ b2.ff0.wt bf16 283370496 18874368 1536 12288
50
+ b2.ff0.bias f32 321119232 12288 12288
51
+ b2.ff2.wt bf16 321168384 9437184 6144 1536
52
+ b2.ff2.bias f32 340042752 1536 1536
53
+ b2.ssg f32 340048896 9216 9216
54
+ b2.local f32 340085760 1536 1536
55
+ b3.pre_g f32 340091904 1536 1536
56
+ b3.ff_g f32 340098048 1536 1536
57
+ b3.cx_g f32 340104192 1536 1536
58
+ b3.s_qkv.wt bf16 340110336 11796480 1536 7680
59
+ b3.s_out.wt bf16 363703296 2359296 1536 1536
60
+ b3.s_qn f32 368421888 64 64
61
+ b3.s_kn f32 368422144 64 64
62
+ b3.c_q.wt bf16 368422400 4718592 1536 3072
63
+ b3.c_kv.wt bf16 377859584 7077888 1536 4608
64
+ b3.c_out.wt bf16 392015360 2359296 1536 1536
65
+ b3.c_qn f32 396733952 64 64
66
+ b3.c_kn f32 396734208 64 64
67
+ b3.ff0.wt bf16 396734464 18874368 1536 12288
68
+ b3.ff0.bias f32 434483200 12288 12288
69
+ b3.ff2.wt bf16 434532352 9437184 6144 1536
70
+ b3.ff2.bias f32 453406720 1536 1536
71
+ b3.ssg f32 453412864 9216 9216
72
+ b3.local f32 453449728 1536 1536
73
+ b4.pre_g f32 453455872 1536 1536
74
+ b4.ff_g f32 453462016 1536 1536
75
+ b4.cx_g f32 453468160 1536 1536
76
+ b4.s_qkv.wt bf16 453474304 11796480 1536 7680
77
+ b4.s_out.wt bf16 477067264 2359296 1536 1536
78
+ b4.s_qn f32 481785856 64 64
79
+ b4.s_kn f32 481786112 64 64
80
+ b4.c_q.wt bf16 481786368 4718592 1536 3072
81
+ b4.c_kv.wt bf16 491223552 7077888 1536 4608
82
+ b4.c_out.wt bf16 505379328 2359296 1536 1536
83
+ b4.c_qn f32 510097920 64 64
84
+ b4.c_kn f32 510098176 64 64
85
+ b4.ff0.wt bf16 510098432 18874368 1536 12288
86
+ b4.ff0.bias f32 547847168 12288 12288
87
+ b4.ff2.wt bf16 547896320 9437184 6144 1536
88
+ b4.ff2.bias f32 566770688 1536 1536
89
+ b4.ssg f32 566776832 9216 9216
90
+ b4.local f32 566813696 1536 1536
91
+ b5.pre_g f32 566819840 1536 1536
92
+ b5.ff_g f32 566825984 1536 1536
93
+ b5.cx_g f32 566832128 1536 1536
94
+ b5.s_qkv.wt bf16 566838272 11796480 1536 7680
95
+ b5.s_out.wt bf16 590431232 2359296 1536 1536
96
+ b5.s_qn f32 595149824 64 64
97
+ b5.s_kn f32 595150080 64 64
98
+ b5.c_q.wt bf16 595150336 4718592 1536 3072
99
+ b5.c_kv.wt bf16 604587520 7077888 1536 4608
100
+ b5.c_out.wt bf16 618743296 2359296 1536 1536
101
+ b5.c_qn f32 623461888 64 64
102
+ b5.c_kn f32 623462144 64 64
103
+ b5.ff0.wt bf16 623462400 18874368 1536 12288
104
+ b5.ff0.bias f32 661211136 12288 12288
105
+ b5.ff2.wt bf16 661260288 9437184 6144 1536
106
+ b5.ff2.bias f32 680134656 1536 1536
107
+ b5.ssg f32 680140800 9216 9216
108
+ b5.local f32 680177664 1536 1536
109
+ b6.pre_g f32 680183808 1536 1536
110
+ b6.ff_g f32 680189952 1536 1536
111
+ b6.cx_g f32 680196096 1536 1536
112
+ b6.s_qkv.wt bf16 680202240 11796480 1536 7680
113
+ b6.s_out.wt bf16 703795200 2359296 1536 1536
114
+ b6.s_qn f32 708513792 64 64
115
+ b6.s_kn f32 708514048 64 64
116
+ b6.c_q.wt bf16 708514304 4718592 1536 3072
117
+ b6.c_kv.wt bf16 717951488 7077888 1536 4608
118
+ b6.c_out.wt bf16 732107264 2359296 1536 1536
119
+ b6.c_qn f32 736825856 64 64
120
+ b6.c_kn f32 736826112 64 64
121
+ b6.ff0.wt bf16 736826368 18874368 1536 12288
122
+ b6.ff0.bias f32 774575104 12288 12288
123
+ b6.ff2.wt bf16 774624256 9437184 6144 1536
124
+ b6.ff2.bias f32 793498624 1536 1536
125
+ b6.ssg f32 793504768 9216 9216
126
+ b6.local f32 793541632 1536 1536
127
+ b7.pre_g f32 793547776 1536 1536
128
+ b7.ff_g f32 793553920 1536 1536
129
+ b7.cx_g f32 793560064 1536 1536
130
+ b7.s_qkv.wt bf16 793566208 11796480 1536 7680
131
+ b7.s_out.wt bf16 817159168 2359296 1536 1536
132
+ b7.s_qn f32 821877760 64 64
133
+ b7.s_kn f32 821878016 64 64
134
+ b7.c_q.wt bf16 821878272 4718592 1536 3072
135
+ b7.c_kv.wt bf16 831315456 7077888 1536 4608
136
+ b7.c_out.wt bf16 845471232 2359296 1536 1536
137
+ b7.c_qn f32 850189824 64 64
138
+ b7.c_kn f32 850190080 64 64
139
+ b7.ff0.wt bf16 850190336 18874368 1536 12288
140
+ b7.ff0.bias f32 887939072 12288 12288
141
+ b7.ff2.wt bf16 887988224 9437184 6144 1536
142
+ b7.ff2.bias f32 906862592 1536 1536
143
+ b7.ssg f32 906868736 9216 9216
144
+ b7.local f32 906905600 1536 1536
145
+ b8.pre_g f32 906911744 1536 1536
146
+ b8.ff_g f32 906917888 1536 1536
147
+ b8.cx_g f32 906924032 1536 1536
148
+ b8.s_qkv.wt bf16 906930176 11796480 1536 7680
149
+ b8.s_out.wt bf16 930523136 2359296 1536 1536
150
+ b8.s_qn f32 935241728 64 64
151
+ b8.s_kn f32 935241984 64 64
152
+ b8.c_q.wt bf16 935242240 4718592 1536 3072
153
+ b8.c_kv.wt bf16 944679424 7077888 1536 4608
154
+ b8.c_out.wt bf16 958835200 2359296 1536 1536
155
+ b8.c_qn f32 963553792 64 64
156
+ b8.c_kn f32 963554048 64 64
157
+ b8.ff0.wt bf16 963554304 18874368 1536 12288
158
+ b8.ff0.bias f32 1001303040 12288 12288
159
+ b8.ff2.wt bf16 1001352192 9437184 6144 1536
160
+ b8.ff2.bias f32 1020226560 1536 1536
161
+ b8.ssg f32 1020232704 9216 9216
162
+ b8.local f32 1020269568 1536 1536
163
+ b9.pre_g f32 1020275712 1536 1536
164
+ b9.ff_g f32 1020281856 1536 1536
165
+ b9.cx_g f32 1020288000 1536 1536
166
+ b9.s_qkv.wt bf16 1020294144 11796480 1536 7680
167
+ b9.s_out.wt bf16 1043887104 2359296 1536 1536
168
+ b9.s_qn f32 1048605696 64 64
169
+ b9.s_kn f32 1048605952 64 64
170
+ b9.c_q.wt bf16 1048606208 4718592 1536 3072
171
+ b9.c_kv.wt bf16 1058043392 7077888 1536 4608
172
+ b9.c_out.wt bf16 1072199168 2359296 1536 1536
173
+ b9.c_qn f32 1076917760 64 64
174
+ b9.c_kn f32 1076918016 64 64
175
+ b9.ff0.wt bf16 1076918272 18874368 1536 12288
176
+ b9.ff0.bias f32 1114667008 12288 12288
177
+ b9.ff2.wt bf16 1114716160 9437184 6144 1536
178
+ b9.ff2.bias f32 1133590528 1536 1536
179
+ b9.ssg f32 1133596672 9216 9216
180
+ b9.local f32 1133633536 1536 1536
181
+ b10.pre_g f32 1133639680 1536 1536
182
+ b10.ff_g f32 1133645824 1536 1536
183
+ b10.cx_g f32 1133651968 1536 1536
184
+ b10.s_qkv.wt bf16 1133658112 11796480 1536 7680
185
+ b10.s_out.wt bf16 1157251072 2359296 1536 1536
186
+ b10.s_qn f32 1161969664 64 64
187
+ b10.s_kn f32 1161969920 64 64
188
+ b10.c_q.wt bf16 1161970176 4718592 1536 3072
189
+ b10.c_kv.wt bf16 1171407360 7077888 1536 4608
190
+ b10.c_out.wt bf16 1185563136 2359296 1536 1536
191
+ b10.c_qn f32 1190281728 64 64
192
+ b10.c_kn f32 1190281984 64 64
193
+ b10.ff0.wt bf16 1190282240 18874368 1536 12288
194
+ b10.ff0.bias f32 1228030976 12288 12288
195
+ b10.ff2.wt bf16 1228080128 9437184 6144 1536
196
+ b10.ff2.bias f32 1246954496 1536 1536
197
+ b10.ssg f32 1246960640 9216 9216
198
+ b10.local f32 1246997504 1536 1536
199
+ b11.pre_g f32 1247003648 1536 1536
200
+ b11.ff_g f32 1247009792 1536 1536
201
+ b11.cx_g f32 1247015936 1536 1536
202
+ b11.s_qkv.wt bf16 1247022080 11796480 1536 7680
203
+ b11.s_out.wt bf16 1270615040 2359296 1536 1536
204
+ b11.s_qn f32 1275333632 64 64
205
+ b11.s_kn f32 1275333888 64 64
206
+ b11.c_q.wt bf16 1275334144 4718592 1536 3072
207
+ b11.c_kv.wt bf16 1284771328 7077888 1536 4608
208
+ b11.c_out.wt bf16 1298927104 2359296 1536 1536
209
+ b11.c_qn f32 1303645696 64 64
210
+ b11.c_kn f32 1303645952 64 64
211
+ b11.ff0.wt bf16 1303646208 18874368 1536 12288
212
+ b11.ff0.bias f32 1341394944 12288 12288
213
+ b11.ff2.wt bf16 1341444096 9437184 6144 1536
214
+ b11.ff2.bias f32 1360318464 1536 1536
215
+ b11.ssg f32 1360324608 9216 9216
216
+ b11.local f32 1360361472 1536 1536
217
+ b12.pre_g f32 1360367616 1536 1536
218
+ b12.ff_g f32 1360373760 1536 1536
219
+ b12.cx_g f32 1360379904 1536 1536
220
+ b12.s_qkv.wt bf16 1360386048 11796480 1536 7680
221
+ b12.s_out.wt bf16 1383979008 2359296 1536 1536
222
+ b12.s_qn f32 1388697600 64 64
223
+ b12.s_kn f32 1388697856 64 64
224
+ b12.c_q.wt bf16 1388698112 4718592 1536 3072
225
+ b12.c_kv.wt bf16 1398135296 7077888 1536 4608
226
+ b12.c_out.wt bf16 1412291072 2359296 1536 1536
227
+ b12.c_qn f32 1417009664 64 64
228
+ b12.c_kn f32 1417009920 64 64
229
+ b12.ff0.wt bf16 1417010176 18874368 1536 12288
230
+ b12.ff0.bias f32 1454758912 12288 12288
231
+ b12.ff2.wt bf16 1454808064 9437184 6144 1536
232
+ b12.ff2.bias f32 1473682432 1536 1536
233
+ b12.ssg f32 1473688576 9216 9216
234
+ b12.local f32 1473725440 1536 1536
235
+ b13.pre_g f32 1473731584 1536 1536
236
+ b13.ff_g f32 1473737728 1536 1536
237
+ b13.cx_g f32 1473743872 1536 1536
238
+ b13.s_qkv.wt bf16 1473750016 11796480 1536 7680
239
+ b13.s_out.wt bf16 1497342976 2359296 1536 1536
240
+ b13.s_qn f32 1502061568 64 64
241
+ b13.s_kn f32 1502061824 64 64
242
+ b13.c_q.wt bf16 1502062080 4718592 1536 3072
243
+ b13.c_kv.wt bf16 1511499264 7077888 1536 4608
244
+ b13.c_out.wt bf16 1525655040 2359296 1536 1536
245
+ b13.c_qn f32 1530373632 64 64
246
+ b13.c_kn f32 1530373888 64 64
247
+ b13.ff0.wt bf16 1530374144 18874368 1536 12288
248
+ b13.ff0.bias f32 1568122880 12288 12288
249
+ b13.ff2.wt bf16 1568172032 9437184 6144 1536
250
+ b13.ff2.bias f32 1587046400 1536 1536
251
+ b13.ssg f32 1587052544 9216 9216
252
+ b13.local f32 1587089408 1536 1536
253
+ b14.pre_g f32 1587095552 1536 1536
254
+ b14.ff_g f32 1587101696 1536 1536
255
+ b14.cx_g f32 1587107840 1536 1536
256
+ b14.s_qkv.wt bf16 1587113984 11796480 1536 7680
257
+ b14.s_out.wt bf16 1610706944 2359296 1536 1536
258
+ b14.s_qn f32 1615425536 64 64
259
+ b14.s_kn f32 1615425792 64 64
260
+ b14.c_q.wt bf16 1615426048 4718592 1536 3072
261
+ b14.c_kv.wt bf16 1624863232 7077888 1536 4608
262
+ b14.c_out.wt bf16 1639019008 2359296 1536 1536
263
+ b14.c_qn f32 1643737600 64 64
264
+ b14.c_kn f32 1643737856 64 64
265
+ b14.ff0.wt bf16 1643738112 18874368 1536 12288
266
+ b14.ff0.bias f32 1681486848 12288 12288
267
+ b14.ff2.wt bf16 1681536000 9437184 6144 1536
268
+ b14.ff2.bias f32 1700410368 1536 1536
269
+ b14.ssg f32 1700416512 9216 9216
270
+ b14.local f32 1700453376 1536 1536
271
+ b15.pre_g f32 1700459520 1536 1536
272
+ b15.ff_g f32 1700465664 1536 1536
273
+ b15.cx_g f32 1700471808 1536 1536
274
+ b15.s_qkv.wt bf16 1700477952 11796480 1536 7680
275
+ b15.s_out.wt bf16 1724070912 2359296 1536 1536
276
+ b15.s_qn f32 1728789504 64 64
277
+ b15.s_kn f32 1728789760 64 64
278
+ b15.c_q.wt bf16 1728790016 4718592 1536 3072
279
+ b15.c_kv.wt bf16 1738227200 7077888 1536 4608
280
+ b15.c_out.wt bf16 1752382976 2359296 1536 1536
281
+ b15.c_qn f32 1757101568 64 64
282
+ b15.c_kn f32 1757101824 64 64
283
+ b15.ff0.wt bf16 1757102080 18874368 1536 12288
284
+ b15.ff0.bias f32 1794850816 12288 12288
285
+ b15.ff2.wt bf16 1794899968 9437184 6144 1536
286
+ b15.ff2.bias f32 1813774336 1536 1536
287
+ b15.ssg f32 1813780480 9216 9216
288
+ b15.local f32 1813817344 1536 1536
289
+ b16.pre_g f32 1813823488 1536 1536
290
+ b16.ff_g f32 1813829632 1536 1536
291
+ b16.cx_g f32 1813835776 1536 1536
292
+ b16.s_qkv.wt bf16 1813841920 11796480 1536 7680
293
+ b16.s_out.wt bf16 1837434880 2359296 1536 1536
294
+ b16.s_qn f32 1842153472 64 64
295
+ b16.s_kn f32 1842153728 64 64
296
+ b16.c_q.wt bf16 1842153984 4718592 1536 3072
297
+ b16.c_kv.wt bf16 1851591168 7077888 1536 4608
298
+ b16.c_out.wt bf16 1865746944 2359296 1536 1536
299
+ b16.c_qn f32 1870465536 64 64
300
+ b16.c_kn f32 1870465792 64 64
301
+ b16.ff0.wt bf16 1870466048 18874368 1536 12288
302
+ b16.ff0.bias f32 1908214784 12288 12288
303
+ b16.ff2.wt bf16 1908263936 9437184 6144 1536
304
+ b16.ff2.bias f32 1927138304 1536 1536
305
+ b16.ssg f32 1927144448 9216 9216
306
+ b16.local f32 1927181312 1536 1536
307
+ b17.pre_g f32 1927187456 1536 1536
308
+ b17.ff_g f32 1927193600 1536 1536
309
+ b17.cx_g f32 1927199744 1536 1536
310
+ b17.s_qkv.wt bf16 1927205888 11796480 1536 7680
311
+ b17.s_out.wt bf16 1950798848 2359296 1536 1536
312
+ b17.s_qn f32 1955517440 64 64
313
+ b17.s_kn f32 1955517696 64 64
314
+ b17.c_q.wt bf16 1955517952 4718592 1536 3072
315
+ b17.c_kv.wt bf16 1964955136 7077888 1536 4608
316
+ b17.c_out.wt bf16 1979110912 2359296 1536 1536
317
+ b17.c_qn f32 1983829504 64 64
318
+ b17.c_kn f32 1983829760 64 64
319
+ b17.ff0.wt bf16 1983830016 18874368 1536 12288
320
+ b17.ff0.bias f32 2021578752 12288 12288
321
+ b17.ff2.wt bf16 2021627904 9437184 6144 1536
322
+ b17.ff2.bias f32 2040502272 1536 1536
323
+ b17.ssg f32 2040508416 9216 9216
324
+ b17.local f32 2040545280 1536 1536
325
+ b18.pre_g f32 2040551424 1536 1536
326
+ b18.ff_g f32 2040557568 1536 1536
327
+ b18.cx_g f32 2040563712 1536 1536
328
+ b18.s_qkv.wt bf16 2040569856 11796480 1536 7680
329
+ b18.s_out.wt bf16 2064162816 2359296 1536 1536
330
+ b18.s_qn f32 2068881408 64 64
331
+ b18.s_kn f32 2068881664 64 64
332
+ b18.c_q.wt bf16 2068881920 4718592 1536 3072
333
+ b18.c_kv.wt bf16 2078319104 7077888 1536 4608
334
+ b18.c_out.wt bf16 2092474880 2359296 1536 1536
335
+ b18.c_qn f32 2097193472 64 64
336
+ b18.c_kn f32 2097193728 64 64
337
+ b18.ff0.wt bf16 2097193984 18874368 1536 12288
338
+ b18.ff0.bias f32 2134942720 12288 12288
339
+ b18.ff2.wt bf16 2134991872 9437184 6144 1536
340
+ b18.ff2.bias f32 2153866240 1536 1536
341
+ b18.ssg f32 2153872384 9216 9216
342
+ b18.local f32 2153909248 1536 1536
343
+ b19.pre_g f32 2153915392 1536 1536
344
+ b19.ff_g f32 2153921536 1536 1536
345
+ b19.cx_g f32 2153927680 1536 1536
346
+ b19.s_qkv.wt bf16 2153933824 11796480 1536 7680
347
+ b19.s_out.wt bf16 2177526784 2359296 1536 1536
348
+ b19.s_qn f32 2182245376 64 64
349
+ b19.s_kn f32 2182245632 64 64
350
+ b19.c_q.wt bf16 2182245888 4718592 1536 3072
351
+ b19.c_kv.wt bf16 2191683072 7077888 1536 4608
352
+ b19.c_out.wt bf16 2205838848 2359296 1536 1536
353
+ b19.c_qn f32 2210557440 64 64
354
+ b19.c_kn f32 2210557696 64 64
355
+ b19.ff0.wt bf16 2210557952 18874368 1536 12288
356
+ b19.ff0.bias f32 2248306688 12288 12288
357
+ b19.ff2.wt bf16 2248355840 9437184 6144 1536
358
+ b19.ff2.bias f32 2267230208 1536 1536
359
+ b19.ssg f32 2267236352 9216 9216
360
+ b19.local f32 2267273216 1536 1536
361
+ b20.pre_g f32 2267279360 1536 1536
362
+ b20.ff_g f32 2267285504 1536 1536
363
+ b20.cx_g f32 2267291648 1536 1536
364
+ b20.s_qkv.wt bf16 2267297792 11796480 1536 7680
365
+ b20.s_out.wt bf16 2290890752 2359296 1536 1536
366
+ b20.s_qn f32 2295609344 64 64
367
+ b20.s_kn f32 2295609600 64 64
368
+ b20.c_q.wt bf16 2295609856 4718592 1536 3072
369
+ b20.c_kv.wt bf16 2305047040 7077888 1536 4608
370
+ b20.c_out.wt bf16 2319202816 2359296 1536 1536
371
+ b20.c_qn f32 2323921408 64 64
372
+ b20.c_kn f32 2323921664 64 64
373
+ b20.ff0.wt bf16 2323921920 18874368 1536 12288
374
+ b20.ff0.bias f32 2361670656 12288 12288
375
+ b20.ff2.wt bf16 2361719808 9437184 6144 1536
376
+ b20.ff2.bias f32 2380594176 1536 1536
377
+ b20.ssg f32 2380600320 9216 9216
378
+ b20.local f32 2380637184 1536 1536
379
+ b21.pre_g f32 2380643328 1536 1536
380
+ b21.ff_g f32 2380649472 1536 1536
381
+ b21.cx_g f32 2380655616 1536 1536
382
+ b21.s_qkv.wt bf16 2380661760 11796480 1536 7680
383
+ b21.s_out.wt bf16 2404254720 2359296 1536 1536
384
+ b21.s_qn f32 2408973312 64 64
385
+ b21.s_kn f32 2408973568 64 64
386
+ b21.c_q.wt bf16 2408973824 4718592 1536 3072
387
+ b21.c_kv.wt bf16 2418411008 7077888 1536 4608
388
+ b21.c_out.wt bf16 2432566784 2359296 1536 1536
389
+ b21.c_qn f32 2437285376 64 64
390
+ b21.c_kn f32 2437285632 64 64
391
+ b21.ff0.wt bf16 2437285888 18874368 1536 12288
392
+ b21.ff0.bias f32 2475034624 12288 12288
393
+ b21.ff2.wt bf16 2475083776 9437184 6144 1536
394
+ b21.ff2.bias f32 2493958144 1536 1536
395
+ b21.ssg f32 2493964288 9216 9216
396
+ b21.local f32 2494001152 1536 1536
397
+ b22.pre_g f32 2494007296 1536 1536
398
+ b22.ff_g f32 2494013440 1536 1536
399
+ b22.cx_g f32 2494019584 1536 1536
400
+ b22.s_qkv.wt bf16 2494025728 11796480 1536 7680
401
+ b22.s_out.wt bf16 2517618688 2359296 1536 1536
402
+ b22.s_qn f32 2522337280 64 64
403
+ b22.s_kn f32 2522337536 64 64
404
+ b22.c_q.wt bf16 2522337792 4718592 1536 3072
405
+ b22.c_kv.wt bf16 2531774976 7077888 1536 4608
406
+ b22.c_out.wt bf16 2545930752 2359296 1536 1536
407
+ b22.c_qn f32 2550649344 64 64
408
+ b22.c_kn f32 2550649600 64 64
409
+ b22.ff0.wt bf16 2550649856 18874368 1536 12288
410
+ b22.ff0.bias f32 2588398592 12288 12288
411
+ b22.ff2.wt bf16 2588447744 9437184 6144 1536
412
+ b22.ff2.bias f32 2607322112 1536 1536
413
+ b22.ssg f32 2607328256 9216 9216
414
+ b22.local f32 2607365120 1536 1536
415
+ b23.pre_g f32 2607371264 1536 1536
416
+ b23.ff_g f32 2607377408 1536 1536
417
+ b23.cx_g f32 2607383552 1536 1536
418
+ b23.s_qkv.wt bf16 2607389696 11796480 1536 7680
419
+ b23.s_out.wt bf16 2630982656 2359296 1536 1536
420
+ b23.s_qn f32 2635701248 64 64
421
+ b23.s_kn f32 2635701504 64 64
422
+ b23.c_q.wt bf16 2635701760 4718592 1536 3072
423
+ b23.c_kv.wt bf16 2645138944 7077888 1536 4608
424
+ b23.c_out.wt bf16 2659294720 2359296 1536 1536
425
+ b23.c_qn f32 2664013312 64 64
426
+ b23.c_kn f32 2664013568 64 64
427
+ b23.ff0.wt bf16 2664013824 18874368 1536 12288
428
+ b23.ff0.bias f32 2701762560 12288 12288
429
+ b23.ff2.wt bf16 2701811712 9437184 6144 1536
430
+ b23.ff2.bias f32 2720686080 1536 1536
431
+ b23.ssg f32 2720692224 9216 9216
432
+ b23.local f32 2720729088 1536 1536
433
+ pout.wt bf16 2720735232 393216 1536 256
434
+ Wpost.wt f32 2721521664 65536 256 256