patrickbdevaney commited on
Commit
866ebe7
·
verified ·
1 Parent(s): d89bb7b

archive auto-beta0p05: eval.log

Browse files
Files changed (1) hide show
  1. arms/auto-beta0p05/eval.log +1099 -0
arms/auto-beta0p05/eval.log ADDED
@@ -0,0 +1,1099 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [decode] loading /home/patrickd/models/DeepSeek-V4-Flash-0731-REAP ... s=6 NDEC=8 seqmax=297
2
+ [decode] loaded 100.40 GiB, 45821 tensors (weights in MANAGED device-preferred memory)
3
+ [decode] KV layout: fp32, 2048 B/row, 0.024 GiB for 297 rows x 43 layers
4
+ [decode] building 43 layer structs once (persistent)...
5
+ [decode] structs built. mem 108.5/122.8 GiB
6
+ [decode] prefill 5 positions...
7
+ [f32mk] tile 4x4 (DSV4_F32MK_TILE=unset)
8
+ [decode] PREFILL: 5 positions in 196.2 ms = 25.5 tok/s (39.237 ms/tok)
9
+ [decode] prefill done. caches populated. starting decode.
10
+ step 0 pos 5 -> token 11111 (67.7 ms warmup)
11
+ step 1 pos 6 -> token 16 (66.9 ms)
12
+ step 2 pos 7 -> token 455 (73.3 ms)
13
+ step 3 pos 8 -> token 6102 (66.7 ms)
14
+ step 4 pos 9 -> token 294 (67.3 ms)
15
+ step 5 pos 10 -> token 16603 (66.9 ms)
16
+ step 6 pos 11 -> token 344 (72.5 ms)
17
+ step 7 pos 12 -> token 29168 (66.3 ms)
18
+
19
+ [decode] first decoded token argmax = 11111 (expect 11111) -> GATE PASS
20
+ [decode] generated: 11111 16 455 6102 294 16603 344 29168
21
+ [decode] WARM decode: 68.6 ms/tok = 14.59 tok/s (M=1 steady state, 7-step avg)
22
+
23
+ [spec-verify] M=5 verify in ONE forward: 115.0 ms (= 23.0 ms/tok if all accepted vs 68.6 M=1 -> 2.98x)
24
+ [spec-verify] verify argmax: 11111 16 455 6102 294
25
+ [spec-verify] decode tokens: 11111 16 455 6102 294
26
+ [spec-verify] MATCH 5/5 -> PASS (M=K verify == K sequential decodes; diffs = MoE-atomic near-ties)
27
+
28
+ [spec] loading EXTERNAL DSpark head /home/patrickd/s5-capture/auto-beta0p05/head ...
29
+ [spec] head loaded 6.53 GiB, 2977 mtp tensors (SEPARATE store)
30
+ [spec] NSTAGE=3 head-experts=160 BLK sweep: 5 5 5 5 5 5 5 5 5
31
+ [spec] head built. mem 114.7/122.8 GiB
32
+ [decode] PREFILL: 5 positions in 137.9 ms = 36.3 tok/s (27.576 ms/tok)
33
+ [spec] decoding 200 tokens (block=5, draft passes=1, adaptK=1.50, prompt=0 s=6)...
34
+ [arena] first draft: draft footprint 210.72 MB, global hwm 210.72 MB / cap 512.00 MB (41.2%), draft path = ARENA
35
+ [margins] 1.78 1.27 0.59 0.55 0.73 K=2
36
+ verify 1: accepted 1/1 (K=2) + correction -> +2 tokens (97.7 ms) cpos=7
37
+ [margins] 0.40 1.25 0.33 0.24 0.12 K=2
38
+ verify 2: accepted 0/1 (K=2) + correction -> +1 tokens (102.9 ms) cpos=8
39
+ [margins] 1.47 0.95 1.82 0.19 0.02 K=2
40
+ verify 3: accepted 1/1 (K=2) + correction -> +2 tokens (96.1 ms) cpos=10
41
+ [margins] 2.41 0.35 3.92 0.46 1.38 K=2
42
+ verify 4: accepted 0/1 (K=2) + correction -> +1 tokens (98.6 ms) cpos=11
43
+ [margins] 6.50 0.04 0.47 1.69 2.49 K=2
44
+ verify 5: accepted 1/1 (K=2) + correction -> +2 tokens (99.3 ms) cpos=13
45
+ [margins] 4.27 2.74 5.19 2.16 0.56 K=5
46
+ verify 6: accepted 4/4 (K=5) + correction -> +5 tokens (139.0 ms) cpos=18
47
+ [margins] 5.11 2.59 2.07 0.74 0.97 K=4
48
+ verify 7: accepted 1/3 (K=4) + correction -> +2 tokens (125.9 ms) cpos=20
49
+ [margins] 2.48 2.20 5.09 2.24 2.02 K=6
50
+ verify 8: accepted 4/5 (K=6) + correction -> +5 tokens (154.9 ms) cpos=25
51
+ [margins] 7.11 2.12 2.00 3.19 2.72 K=6
52
+ verify 9: accepted 1/5 (K=6) + correction -> +2 tokens (156.5 ms) cpos=27
53
+ [margins] 3.65 3.44 6.11 2.64 0.58 K=5
54
+ verify 10: accepted 4/4 (K=5) + correction -> +5 tokens (144.2 ms) cpos=32
55
+ [margins] 7.22 0.10 3.64 1.81 1.36 K=2
56
+ verify 11: accepted 1/1 (K=2) + correction -> +2 tokens (95.9 ms) cpos=34
57
+ [margins] 2.08 4.13 4.98 1.61 1.25 K=5
58
+ verify 12: accepted 4/4 (K=5) + correction -> +5 tokens (141.2 ms) cpos=39
59
+ [margins] 4.58 0.04 4.53 2.87 3.30 K=2
60
+ verify 13: accepted 1/1 (K=2) + correction -> +2 tokens (100.8 ms) cpos=41
61
+ [margins] 3.48 3.71 6.25 1.63 0.95 K=5
62
+ verify 14: accepted 4/4 (K=5) + correction -> +5 tokens (140.1 ms) cpos=46
63
+ [margins] 1.71 1.31 4.51 0.36 3.39 K=2
64
+ verify 15: accepted 1/1 (K=2) + correction -> +2 tokens (100.9 ms) cpos=48
65
+ [margins] 0.04 0.63 1.05 1.62 0.59 K=2
66
+ verify 16: accepted 0/1 (K=2) + correction -> +1 tokens (97.9 ms) cpos=49
67
+ [margins] 6.54 4.74 3.67 5.07 1.17 K=5
68
+ verify 17: accepted 4/4 (K=5) + correction -> +5 tokens (143.5 ms) cpos=54
69
+ [margins] 2.99 1.81 7.06 0.20 2.48 K=4
70
+ verify 18: accepted 3/3 (K=4) + correction -> +4 tokens (127.1 ms) cpos=58
71
+ [margins] 1.15 5.07 5.46 6.35 1.40 K=5
72
+ verify 19: accepted 0/4 (K=5) + correction -> +1 tokens (141.5 ms) cpos=59
73
+ [margins] 5.86 6.07 6.88 0.12 5.17 K=4
74
+ verify 20: accepted 3/3 (K=4) + correction -> +4 tokens (129.2 ms) cpos=63
75
+ [margins] 0.59 0.99 6.98 0.98 3.68 K=2
76
+ verify 21: accepted 0/1 (K=2) + correction -> +1 tokens (101.5 ms) cpos=64
77
+ [margins] 6.40 0.39 3.23 2.63 3.85 K=2
78
+ verify 22: accepted 1/1 (K=2) + correction -> +2 tokens (98.3 ms) cpos=66
79
+ [margins] 4.65 4.25 5.10 2.71 0.10 K=5
80
+ verify 23: accepted 4/4 (K=5) + correction -> +5 tokens (142.4 ms) cpos=71
81
+ [margins] 6.75 1.83 2.17 1.43 2.09 K=4
82
+ verify 24: accepted 3/3 (K=4) + correction -> +4 tokens (129.8 ms) cpos=75
83
+ [margins] 7.30 5.32 2.92 0.67 0.99 K=4
84
+ verify 25: accepted 2/3 (K=4) + correction -> +3 tokens (128.1 ms) cpos=78
85
+ [margins] 6.03 0.43 1.75 5.10 3.03 K=2
86
+ verify 26: accepted 1/1 (K=2) + correction -> +2 tokens (103.6 ms) cpos=80
87
+ [margins] 5.10 6.46 5.84 3.61 1.82 K=6
88
+ verify 27: accepted 4/5 (K=6) + correction -> +5 tokens (160.1 ms) cpos=85
89
+ [margins] 7.77 0.97 2.77 3.74 3.69 K=2
90
+ verify 28: accepted 1/1 (K=2) + correction -> +2 tokens (99.4 ms) cpos=87
91
+ [margins] 5.87 5.72 7.38 2.59 2.67 K=6
92
+ verify 29: accepted 4/5 (K=6) + correction -> +5 tokens (166.1 ms) cpos=92
93
+ [margins] 7.98 1.08 0.03 3.31 5.27 K=2
94
+ verify 30: accepted 1/1 (K=2) + correction -> +2 tokens (100.8 ms) cpos=94
95
+ [margins] 4.94 4.93 4.79 3.42 2.61 K=6
96
+ verify 31: accepted 4/5 (K=6) + correction -> +5 tokens (167.3 ms) cpos=99
97
+ [margins] 5.87 1.64 0.79 3.02 4.38 K=3
98
+ verify 32: accepted 1/2 (K=3) + correction -> +2 tokens (117.8 ms) cpos=101
99
+ [margins] 6.95 5.88 6.01 0.06 1.39 K=4
100
+ verify 33: accepted 3/3 (K=4) + correction -> +4 tokens (130.5 ms) cpos=105
101
+ [margins] 2.14 7.57 0.40 6.10 3.82 K=3
102
+ verify 34: accepted 2/2 (K=3) + correction -> +3 tokens (117.6 ms) cpos=108
103
+ [margins] 6.94 4.68 6.05 3.08 2.20 K=6
104
+ verify 35: accepted 4/5 (K=6) + correction -> +5 tokens (162.3 ms) cpos=113
105
+ [margins] 8.08 1.57 0.14 3.90 3.20 K=3
106
+ verify 36: accepted 1/2 (K=3) + correction -> +2 tokens (116.9 ms) cpos=115
107
+ [margins] 3.51 1.96 7.10 2.70 0.24 K=5
108
+ verify 37: accepted 4/4 (K=5) + correction -> +5 tokens (149.7 ms) cpos=120
109
+ [margins] 7.49 1.61 0.73 3.65 1.69 K=3
110
+ verify 38: accepted 1/2 (K=3) + correction -> +2 tokens (113.6 ms) cpos=122
111
+ [margins] 6.05 4.28 5.85 4.61 1.27 K=5
112
+ verify 39: accepted 4/4 (K=5) + correction -> +5 tokens (145.3 ms) cpos=127
113
+ [margins] 7.76 1.47 0.85 4.04 4.62 K=2
114
+ verify 40: accepted 1/1 (K=2) + correction -> +2 tokens (119.6 ms) cpos=129
115
+ [margins] 8.14 6.31 8.32 1.67 1.67 K=6
116
+ verify 41: accepted 4/5 (K=6) + correction -> +5 tokens (163.2 ms) cpos=134
117
+ [margins] 6.72 1.50 1.15 2.16 2.92 K=2
118
+ verify 42: accepted 1/1 (K=2) + correction -> +2 tokens (107.3 ms) cpos=136
119
+ [margins] 7.96 4.12 5.14 2.70 1.59 K=6
120
+ verify 43: accepted 5/5 (K=6) + correction -> +6 tokens (161.9 ms) cpos=142
121
+ [margins] 9.05 9.01 2.81 2.13 5.57 K=6
122
+ verify 44: accepted 2/5 (K=6) + correction -> +3 tokens (165.4 ms) cpos=145
123
+ [margins] 6.76 6.16 7.15 1.85 2.30 K=6
124
+ verify 45: accepted 4/5 (K=6) + correction -> +5 tokens (160.1 ms) cpos=150
125
+ [margins] 7.17 1.11 1.79 5.62 4.95 K=2
126
+ verify 46: accepted 1/1 (K=2) + correction -> +2 tokens (105.5 ms) cpos=152
127
+ [margins] 14.20 3.06 7.80 7.91 9.52 K=6
128
+ verify 47: accepted 1/5 (K=6) + correction -> +2 tokens (161.5 ms) cpos=154
129
+ [margins] 3.84 8.79 7.62 4.40 3.45 K=6
130
+ verify 48: accepted 4/5 (K=6) + correction -> +5 tokens (166.6 ms) cpos=159
131
+ [margins] 3.64 0.03 0.07 4.58 4.72 K=2
132
+ verify 49: accepted 1/1 (K=2) + correction -> +2 tokens (105.0 ms) cpos=161
133
+ [margins] 9.60 6.63 7.90 7.66 3.30 K=6
134
+ verify 50: accepted 5/5 (K=6) + correction -> +6 tokens (163.3 ms) cpos=167
135
+ [margins] 6.10 1.08 2.41 4.54 4.70 K=2
136
+ verify 51: accepted 1/1 (K=2) + correction -> +2 tokens (104.6 ms) cpos=169
137
+ [margins] 2.68 5.00 6.46 3.71 2.40 K=6
138
+ verify 52: accepted 4/5 (K=6) + correction -> +5 tokens (160.5 ms) cpos=174
139
+ [margins] 3.77 1.89 1.30 4.66 5.84 K=3
140
+ verify 53: accepted 1/2 (K=3) + correction -> +2 tokens (117.3 ms) cpos=176
141
+ [margins] 4.47 5.67 6.88 2.45 1.00 K=5
142
+ verify 54: accepted 4/4 (K=5) + correction -> +5 tokens (144.5 ms) cpos=181
143
+ [margins] 7.00 2.12 0.10 2.06 6.08 K=3
144
+ verify 55: accepted 1/2 (K=3) + correction -> +2 tokens (118.0 ms) cpos=183
145
+ [margins] 3.98 6.39 8.25 2.56 0.80 K=5
146
+ verify 56: accepted 4/4 (K=5) + correction -> +5 tokens (150.7 ms) cpos=188
147
+ [margins] 2.18 0.48 1.23 1.02 6.36 K=2
148
+ verify 57: accepted 1/1 (K=2) + correction -> +2 tokens (102.1 ms) cpos=190
149
+ [margins] 9.29 2.10 0.10 3.20 2.63 K=3
150
+ verify 58: accepted 1/2 (K=3) + correction -> +2 tokens (118.4 ms) cpos=192
151
+ [margins] 6.37 2.54 6.50 6.31 8.64 K=6
152
+ verify 59: accepted 5/5 (K=6) + correction -> +6 tokens (162.3 ms) cpos=198
153
+ [margins] 1.49 6.37 0.71 1.36 4.14 K=3
154
+ verify 60: accepted 2/2 (K=3) + correction -> +3 tokens (119.1 ms) cpos=201
155
+ [margins] 5.93 7.40 8.64 7.61 1.16 K=5
156
+ verify 61: accepted 4/4 (K=5) + correction -> +5 tokens (144.9 ms) cpos=206
157
+
158
+ [spec] generated 201 tokens over 61 verifies: mean tokens/verify = 3.30 (block=5, max 6)
159
+ [spec] tokens: 11111 16 455 6102 294 16603 344 29168 16 455 6102 294 14251 344 16235 16 455 6102 294 29585 344 76405 16 455 6102 294 25062 344 40977 16 455 6102 294 10322 344 17575 16 455 6102 294
160
+ [spec] LOSSLESS GATE: first 8 tokens match base AR -> PASS
161
+ [spec] SPEC-DECODE: 39.2 ms/tok = 25.49 tok/s (vs base M=1 68.6 ms/tok = 14.59 tok/s -> 1.75x)
162
+ [decode] PREFILL: 34 positions in 602.5 ms = 56.4 tok/s (17.719 ms/tok)
163
+ [spec] decoding 200 tokens (block=5, draft passes=1, adaptK=1.50, prompt=1 s=35)...
164
+ [margins] 2.24 1.71 1.69 1.93 0.55 K=5
165
+ verify 1: accepted 0/4 (K=5) + correction -> +1 tokens (134.6 ms) cpos=35
166
+ [margins] 2.92 2.58 2.04 2.26 4.93 K=6
167
+ verify 2: accepted 0/5 (K=6) + correction -> +1 tokens (159.2 ms) cpos=36
168
+ [margins] 1.64 1.47 0.83 3.22 0.63 K=2
169
+ verify 3: accepted 0/1 (K=2) + correction -> +1 tokens (95.3 ms) cpos=37
170
+ [margins] 1.84 2.35 1.70 1.42 0.67 K=4
171
+ verify 4: accepted 2/3 (K=4) + correction -> +3 tokens (122.7 ms) cpos=40
172
+ [margins] 0.54 0.54 0.89 0.21 1.20 K=2
173
+ verify 5: accepted 0/1 (K=2) + correction -> +1 tokens (96.9 ms) cpos=41
174
+ [margins] 1.27 3.26 0.14 1.65 0.80 K=3
175
+ verify 6: accepted 2/2 (K=3) + correction -> +3 tokens (112.5 ms) cpos=44
176
+ [margins] 0.75 5.42 11.16 2.28 1.77 K=6
177
+ verify 7: accepted 5/5 (K=6) + correction -> +6 tokens (155.0 ms) cpos=50
178
+ [margins] 3.35 1.89 0.20 0.43 1.48 K=3
179
+ verify 8: accepted 1/2 (K=3) + correction -> +2 tokens (112.7 ms) cpos=52
180
+ [margins] 4.79 5.33 7.55 1.93 0.53 K=5
181
+ verify 9: accepted 3/4 (K=5) + correction -> +4 tokens (140.9 ms) cpos=56
182
+ [margins] 3.71 2.60 0.02 2.21 5.88 K=3
183
+ verify 10: accepted 2/2 (K=3) + correction -> +3 tokens (107.9 ms) cpos=59
184
+ [margins] 0.59 0.65 3.50 0.64 1.69 K=2
185
+ verify 11: accepted 0/1 (K=2) + correction -> +1 tokens (101.9 ms) cpos=60
186
+ [margins] 1.28 1.76 5.35 0.64 1.14 K=4
187
+ verify 12: accepted 0/3 (K=4) + correction -> +1 tokens (126.4 ms) cpos=61
188
+ [margins] 0.92 4.28 0.09 1.29 0.50 K=3
189
+ verify 13: accepted 1/2 (K=3) + correction -> +2 tokens (114.8 ms) cpos=63
190
+ [margins] 2.01 4.74 0.61 5.25 0.05 K=3
191
+ verify 14: accepted 0/2 (K=3) + correction -> +1 tokens (115.3 ms) cpos=64
192
+ [margins] 4.28 0.24 0.44 6.76 1.65 K=2
193
+ verify 15: accepted 1/1 (K=2) + correction -> +2 tokens (97.7 ms) cpos=66
194
+ [margins] 1.19 1.54 3.10 5.01 0.23 K=5
195
+ verify 16: accepted 4/4 (K=5) + correction -> +5 tokens (139.4 ms) cpos=71
196
+ [margins] 2.72 1.50 1.35 4.77 0.07 K=3
197
+ verify 17: accepted 2/2 (K=3) + correction -> +3 tokens (114.8 ms) cpos=74
198
+ [margins] 7.48 1.99 3.12 2.66 3.12 K=6
199
+ verify 18: accepted 5/5 (K=6) + correction -> +6 tokens (160.9 ms) cpos=80
200
+ [margins] 5.42 5.40 5.25 2.20 6.83 K=6
201
+ verify 19: accepted 5/5 (K=6) + correction -> +6 tokens (158.5 ms) cpos=86
202
+ [margins] 8.67 3.68 2.89 5.85 0.81 K=5
203
+ verify 20: accepted 4/4 (K=5) + correction -> +5 tokens (143.4 ms) cpos=91
204
+ [margins] 7.39 2.12 3.56 8.58 0.87 K=5
205
+ verify 21: accepted 4/4 (K=5) + correction -> +5 tokens (146.5 ms) cpos=96
206
+ [margins] 9.43 5.74 4.35 8.40 2.36 K=6
207
+ verify 22: accepted 5/5 (K=6) + correction -> +6 tokens (158.5 ms) cpos=102
208
+ [margins] 9.71 3.28 7.47 8.31 6.14 K=6
209
+ verify 23: accepted 5/5 (K=6) + correction -> +6 tokens (165.6 ms) cpos=108
210
+ [margins] 2.99 1.22 0.09 6.81 4.74 K=2
211
+ verify 24: accepted 0/1 (K=2) + correction -> +1 tokens (102.1 ms) cpos=109
212
+ [margins] 9.18 6.65 5.25 5.60 5.45 K=6
213
+ verify 25: accepted 5/5 (K=6) + correction -> +6 tokens (159.6 ms) cpos=115
214
+ [margins] 9.54 8.86 10.03 3.05 7.03 K=6
215
+ verify 26: accepted 5/5 (K=6) + correction -> +6 tokens (167.5 ms) cpos=121
216
+ [margins] 9.55 5.18 3.92 8.76 0.53 K=5
217
+ verify 27: accepted 4/4 (K=5) + correction -> +5 tokens (147.6 ms) cpos=126
218
+ [margins] 8.75 6.54 7.43 1.05 0.84 K=4
219
+ verify 28: accepted 3/3 (K=4) + correction -> +4 tokens (148.5 ms) cpos=130
220
+ [margins] 8.36 7.62 5.46 3.39 8.36 K=6
221
+ verify 29: accepted 5/5 (K=6) + correction -> +6 tokens (166.5 ms) cpos=136
222
+ [margins] 7.84 5.34 6.76 9.26 2.97 K=6
223
+ verify 30: accepted 5/5 (K=6) + correction -> +6 tokens (162.1 ms) cpos=142
224
+ [margins] 12.89 6.34 10.41 12.46 9.40 K=6
225
+ verify 31: accepted 5/5 (K=6) + correction -> +6 tokens (168.4 ms) cpos=148
226
+ [margins] 5.15 0.24 7.81 2.95 3.29 K=2
227
+ verify 32: accepted 0/1 (K=2) + correction -> +1 tokens (103.5 ms) cpos=149
228
+ [margins] 9.02 10.93 8.21 8.15 4.64 K=6
229
+ verify 33: accepted 5/5 (K=6) + correction -> +6 tokens (160.9 ms) cpos=155
230
+ [margins] 10.01 8.48 8.20 2.98 8.31 K=6
231
+ verify 34: accepted 5/5 (K=6) + correction -> +6 tokens (167.7 ms) cpos=161
232
+ [margins] 10.27 6.65 5.29 9.18 7.17 K=6
233
+ verify 35: accepted 5/5 (K=6) + correction -> +6 tokens (161.7 ms) cpos=167
234
+ [margins] 13.86 5.60 10.62 9.73 10.35 K=6
235
+ verify 36: accepted 5/5 (K=6) + correction -> +6 tokens (168.6 ms) cpos=173
236
+ [margins] 1.49 8.57 9.58 6.94 6.28 K=6
237
+ verify 37: accepted 5/5 (K=6) + correction -> +6 tokens (160.2 ms) cpos=179
238
+ [margins] 8.99 10.53 7.23 8.60 5.99 K=6
239
+ verify 38: accepted 5/5 (K=6) + correction -> +6 tokens (165.7 ms) cpos=185
240
+ [margins] 10.59 10.94 9.91 8.31 9.65 K=6
241
+ verify 39: accepted 5/5 (K=6) + correction -> +6 tokens (161.8 ms) cpos=191
242
+ [margins] 12.61 9.80 6.13 9.36 0.35 K=5
243
+ verify 40: accepted 4/4 (K=5) + correction -> +5 tokens (152.0 ms) cpos=196
244
+ [margins] 13.52 7.58 10.52 9.55 6.88 K=6
245
+ verify 41: accepted 5/5 (K=6) + correction -> +6 tokens (162.6 ms) cpos=202
246
+ [margins] 14.03 6.04 8.30 6.70 6.92 K=6
247
+ verify 42: accepted 5/5 (K=6) + correction -> +6 tokens (168.7 ms) cpos=208
248
+ [margins] 5.26 0.44 0.37 1.15 0.45 K=2
249
+ verify 43: accepted 0/1 (K=2) + correction -> +1 tokens (103.8 ms) cpos=209
250
+ [margins] 9.77 7.33 7.40 8.25 8.72 K=6
251
+ verify 44: accepted 5/5 (K=6) + correction -> +6 tokens (162.3 ms) cpos=215
252
+ [margins] 8.80 13.01 9.91 8.99 9.31 K=6
253
+ verify 45: accepted 5/5 (K=6) + correction -> +6 tokens (169.5 ms) cpos=221
254
+ [margins] 11.03 3.52 10.01 9.72 1.10 K=5
255
+ verify 46: accepted 4/4 (K=5) + correction -> +5 tokens (148.6 ms) cpos=226
256
+ [margins] 13.39 6.07 8.14 7.26 6.80 K=6
257
+ verify 47: accepted 5/5 (K=6) + correction -> +6 tokens (168.4 ms) cpos=232
258
+ [margins] 13.87 6.27 11.28 8.69 9.10 K=6
259
+ verify 48: accepted 5/5 (K=6) + correction -> +6 tokens (165.9 ms) cpos=238
260
+
261
+ [spec] generated 204 tokens over 48 verifies: mean tokens/verify = 4.25 (block=5, max 6)
262
+ [spec] tokens: 4147 554 1926 270 4105 339 43 1114 1926 270 1178 65 50219 4105 418 270 9517 28 4593 1281 106239 3305 1 1 14 223 19 16 23 14 223 20 16 23 14 223 21 16 23 14
263
+ [spec] SPEC-DECODE: 33.1 ms/tok = 30.20 tok/s (vs base M=1 68.6 ms/tok = 14.59 tok/s -> 2.07x)
264
+ [decode] PREFILL: 41 positions in 692.6 ms = 59.2 tok/s (16.892 ms/tok)
265
+ [spec] decoding 200 tokens (block=5, draft passes=1, adaptK=1.50, prompt=2 s=42)...
266
+ [margins] 7.50 0.32 0.94 2.81 0.51 K=2
267
+ verify 1: accepted 1/1 (K=2) + correction -> +2 tokens (99.6 ms) cpos=43
268
+ [margins] 0.63 3.93 2.14 1.16 0.80 K=4
269
+ verify 2: accepted 3/3 (K=4) + correction -> +4 tokens (114.8 ms) cpos=47
270
+ [margins] 0.22 2.54 1.15 0.22 0.26 K=3
271
+ verify 3: accepted 0/2 (K=3) + correction -> +1 tokens (104.4 ms) cpos=48
272
+ [margins] 3.96 5.99 0.35 0.18 0.63 K=3
273
+ verify 4: accepted 0/2 (K=3) + correction -> +1 tokens (109.0 ms) cpos=49
274
+ [margins] 3.00 8.12 1.19 1.44 3.22 K=3
275
+ verify 5: accepted 0/2 (K=3) + correction -> +1 tokens (113.9 ms) cpos=50
276
+ [margins] 1.68 1.40 0.19 0.64 2.19 K=2
277
+ verify 6: accepted 0/1 (K=2) + correction -> +1 tokens (94.5 ms) cpos=51
278
+ [margins] 7.22 4.74 0.62 0.64 0.33 K=3
279
+ verify 7: accepted 1/2 (K=3) + correction -> +2 tokens (110.2 ms) cpos=53
280
+ [margins] 0.40 4.16 6.52 1.19 0.35 K=4
281
+ verify 8: accepted 0/3 (K=4) + correction -> +1 tokens (126.8 ms) cpos=54
282
+ [margins] 0.57 6.93 1.88 1.56 0.12 K=5
283
+ verify 9: accepted 2/4 (K=5) + correction -> +3 tokens (133.1 ms) cpos=57
284
+ [margins] 0.79 4.54 8.52 0.64 0.06 K=4
285
+ verify 10: accepted 0/3 (K=4) + correction -> +1 tokens (126.9 ms) cpos=58
286
+ [margins] 4.49 4.57 2.70 3.76 0.10 K=5
287
+ verify 11: accepted 2/4 (K=5) + correction -> +3 tokens (133.6 ms) cpos=61
288
+ [margins] 2.15 6.35 8.39 4.79 5.58 K=6
289
+ verify 12: accepted 3/5 (K=6) + correction -> +4 tokens (153.2 ms) cpos=65
290
+ [margins] 1.53 4.97 6.37 2.95 6.16 K=6
291
+ verify 13: accepted 3/5 (K=6) + correction -> +4 tokens (152.4 ms) cpos=69
292
+ [margins] 4.49 5.95 5.98 0.08 1.84 K=4
293
+ verify 14: accepted 2/3 (K=4) + correction -> +3 tokens (129.3 ms) cpos=72
294
+ [margins] 3.42 0.06 1.81 3.92 1.14 K=2
295
+ verify 15: accepted 1/1 (K=2) + correction -> +2 tokens (97.6 ms) cpos=74
296
+ [margins] 3.13 5.93 2.12 0.89 4.19 K=4
297
+ verify 16: accepted 1/3 (K=4) + correction -> +2 tokens (126.1 ms) cpos=76
298
+ [margins] 9.25 5.70 4.88 4.23 0.71 K=5
299
+ verify 17: accepted 2/4 (K=5) + correction -> +3 tokens (140.4 ms) cpos=79
300
+ [margins] 17.19 8.77 6.01 5.13 0.36 K=5
301
+ verify 18: accepted 4/4 (K=5) + correction -> +5 tokens (145.6 ms) cpos=84
302
+ [margins] 17.79 10.33 3.38 0.55 1.49 K=4
303
+ verify 19: accepted 2/3 (K=4) + correction -> +3 tokens (130.1 ms) cpos=87
304
+ [margins] 11.62 15.11 13.48 3.45 11.57 K=6
305
+ verify 20: accepted 5/5 (K=6) + correction -> +6 tokens (161.6 ms) cpos=93
306
+ [margins] 8.83 10.90 12.36 7.99 1.87 K=6
307
+ verify 21: accepted 4/5 (K=6) + correction -> +5 tokens (156.9 ms) cpos=98
308
+ [margins] 10.54 8.63 5.43 3.42 2.01 K=6
309
+ verify 22: accepted 4/5 (K=6) + correction -> +5 tokens (164.2 ms) cpos=103
310
+ [margins] 17.76 13.99 10.36 4.17 3.39 K=6
311
+ verify 23: accepted 5/5 (K=6) + correction -> +6 tokens (164.5 ms) cpos=109
312
+ [margins] 9.61 2.43 3.11 6.60 3.91 K=6
313
+ verify 24: accepted 5/5 (K=6) + correction -> +6 tokens (154.6 ms) cpos=115
314
+ [margins] 11.79 9.22 11.77 0.78 0.00 K=4
315
+ verify 25: accepted 3/3 (K=4) + correction -> +4 tokens (130.7 ms) cpos=119
316
+ [margins] 5.08 7.57 0.21 3.23 0.50 K=3
317
+ verify 26: accepted 2/2 (K=3) + correction -> +3 tokens (116.5 ms) cpos=122
318
+ [margins] 15.32 3.70 3.52 9.99 1.52 K=6
319
+ verify 27: accepted 5/5 (K=6) + correction -> +6 tokens (180.0 ms) cpos=128
320
+ [margins] 9.19 11.66 12.86 7.08 4.95 K=6
321
+ verify 28: accepted 5/5 (K=6) + correction -> +6 tokens (161.7 ms) cpos=134
322
+ [margins] 1.91 1.41 10.80 6.17 4.15 K=2
323
+ verify 29: accepted 0/1 (K=2) + correction -> +1 tokens (108.2 ms) cpos=135
324
+ [margins] 14.96 10.05 2.21 2.74 4.62 K=6
325
+ verify 30: accepted 5/5 (K=6) + correction -> +6 tokens (166.3 ms) cpos=141
326
+ [margins] 14.47 12.82 11.24 13.45 4.57 K=6
327
+ verify 31: accepted 5/5 (K=6) + correction -> +6 tokens (158.9 ms) cpos=147
328
+ [margins] 5.95 16.25 10.47 3.69 2.45 K=6
329
+ verify 32: accepted 5/5 (K=6) + correction -> +6 tokens (166.8 ms) cpos=153
330
+ [margins] 1.91 8.67 12.43 7.05 5.84 K=6
331
+ verify 33: accepted 4/5 (K=6) + correction -> +5 tokens (160.9 ms) cpos=158
332
+ [margins] 12.03 6.24 8.87 0.19 0.56 K=4
333
+ verify 34: accepted 3/3 (K=4) + correction -> +4 tokens (129.2 ms) cpos=162
334
+ [margins] 16.46 0.39 1.81 4.16 2.61 K=2
335
+ verify 35: accepted 1/1 (K=2) + correction -> +2 tokens (103.3 ms) cpos=164
336
+ [margins] 12.70 8.44 2.42 0.91 1.04 K=4
337
+ verify 36: accepted 2/3 (K=4) + correction -> +3 tokens (132.2 ms) cpos=167
338
+ [margins] 10.51 14.09 11.70 7.08 2.69 K=6
339
+ verify 37: accepted 4/5 (K=6) + correction -> +5 tokens (165.3 ms) cpos=172
340
+ [margins] 16.98 7.15 11.49 6.23 0.82 K=5
341
+ verify 38: accepted 4/4 (K=5) + correction -> +5 tokens (144.6 ms) cpos=177
342
+ [margins] 15.90 14.98 13.42 9.87 3.89 K=6
343
+ verify 39: accepted 5/5 (K=6) + correction -> +6 tokens (160.6 ms) cpos=183
344
+ [margins] 15.08 11.26 9.24 1.24 5.92 K=4
345
+ verify 40: accepted 3/3 (K=4) + correction -> +4 tokens (133.0 ms) cpos=187
346
+ [margins] 13.34 11.51 10.98 5.44 0.32 K=5
347
+ verify 41: accepted 4/4 (K=5) + correction -> +5 tokens (152.8 ms) cpos=192
348
+ [margins] 13.53 9.12 7.16 7.33 2.39 K=6
349
+ verify 42: accepted 5/5 (K=6) + correction -> +6 tokens (162.7 ms) cpos=198
350
+ [margins] 14.48 8.19 8.62 0.50 0.72 K=4
351
+ verify 43: accepted 3/3 (K=4) + correction -> +4 tokens (134.3 ms) cpos=202
352
+ [margins] 17.64 12.38 13.30 8.46 1.13 K=5
353
+ verify 44: accepted 4/4 (K=5) + correction -> +5 tokens (146.9 ms) cpos=207
354
+ [margins] 10.77 7.71 7.15 5.29 4.61 K=6
355
+ verify 45: accepted 5/5 (K=6) + correction -> +6 tokens (161.5 ms) cpos=213
356
+ [margins] 20.90 11.71 5.12 12.46 5.50 K=6
357
+ verify 46: accepted 5/5 (K=6) + correction -> +6 tokens (159.8 ms) cpos=219
358
+ [margins] 20.42 10.61 19.62 5.72 14.36 K=6
359
+ verify 47: accepted 5/5 (K=6) + correction -> +6 tokens (170.1 ms) cpos=225
360
+ [margins] 8.01 24.34 7.75 14.77 9.32 K=6
361
+ verify 48: accepted 5/5 (K=6) + correction -> +6 tokens (163.2 ms) cpos=231
362
+ [margins] 27.61 10.29 15.97 4.38 7.61 K=6
363
+ verify 49: accepted 5/5 (K=6) + correction -> +6 tokens (170.1 ms) cpos=237
364
+ [margins] 11.95 13.21 6.85 0.98 7.47 K=4
365
+ verify 50: accepted 3/3 (K=4) + correction -> +4 tokens (134.7 ms) cpos=241
366
+
367
+ [spec] generated 200 tokens over 50 verifies: mean tokens/verify = 4.00 (block=5, max 6)
368
+ [spec] tokens: 1 1 1 1 1 1 50 1697 1 16 1 50 1697 16 1 50 1697 16 1 50 1697 16 1 50 1697 16 1 50 1697 16 20692 50 1697 16 791 50 1697 4394 1015 50
369
+ [spec] SPEC-DECODE: 34.9 ms/tok = 28.62 tok/s (vs base M=1 68.6 ms/tok = 14.59 tok/s -> 1.96x)
370
+ [decode] PREFILL: 26 positions in 454.0 ms = 57.3 tok/s (17.460 ms/tok)
371
+ [spec] decoding 200 tokens (block=5, draft passes=1, adaptK=1.50, prompt=3 s=27)...
372
+ [margins] 1.81 2.01 0.97 0.54 0.62 K=3
373
+ verify 1: accepted 1/2 (K=3) + correction -> +2 tokens (114.6 ms) cpos=28
374
+ [margins] 4.69 2.59 0.65 1.53 0.57 K=3
375
+ verify 2: accepted 1/2 (K=3) + correction -> +2 tokens (110.0 ms) cpos=30
376
+ [margins] 2.84 0.76 1.05 0.11 0.16 K=2
377
+ verify 3: accepted 1/1 (K=2) + correction -> +2 tokens (102.9 ms) cpos=32
378
+ [margins] 3.02 2.88 2.09 0.38 0.68 K=4
379
+ verify 4: accepted 3/3 (K=4) + correction -> +4 tokens (118.0 ms) cpos=36
380
+ [margins] 2.69 6.57 1.99 1.51 1.50 K=6
381
+ verify 5: accepted 5/5 (K=6) + correction -> +6 tokens (143.0 ms) cpos=42
382
+ [margins] 1.87 9.82 1.85 4.64 1.52 K=6
383
+ verify 6: accepted 5/5 (K=6) + correction -> +6 tokens (149.5 ms) cpos=48
384
+ [margins] 0.19 12.95 2.80 7.99 3.05 K=6
385
+ verify 7: accepted 0/5 (K=6) + correction -> +1 tokens (148.3 ms) cpos=49
386
+ [margins] 8.19 2.23 8.80 2.00 6.75 K=6
387
+ verify 8: accepted 5/5 (K=6) + correction -> +6 tokens (144.1 ms) cpos=55
388
+ [margins] 7.28 0.46 8.02 3.21 6.54 K=2
389
+ verify 9: accepted 1/1 (K=2) + correction -> +2 tokens (102.5 ms) cpos=57
390
+ [margins] 8.02 1.59 9.09 2.61 6.00 K=6
391
+ verify 10: accepted 5/5 (K=6) + correction -> +6 tokens (145.3 ms) cpos=63
392
+ [margins] 6.93 0.98 7.96 3.27 6.06 K=2
393
+ verify 11: accepted 1/1 (K=2) + correction -> +2 tokens (103.4 ms) cpos=65
394
+ [margins] 6.73 2.32 8.54 2.97 6.16 K=6
395
+ verify 12: accepted 5/5 (K=6) + correction -> +6 tokens (144.7 ms) cpos=71
396
+ [margins] 8.04 0.74 9.61 2.75 6.56 K=2
397
+ verify 13: accepted 1/1 (K=2) + correction -> +2 tokens (103.5 ms) cpos=73
398
+ [margins] 7.70 0.62 9.84 3.24 6.83 K=2
399
+ verify 14: accepted 1/1 (K=2) + correction -> +2 tokens (99.6 ms) cpos=75
400
+ [margins] 8.65 2.08 7.54 2.09 4.32 K=6
401
+ verify 15: accepted 0/5 (K=6) + correction -> +1 tokens (155.6 ms) cpos=76
402
+ [margins] 2.09 6.23 0.29 1.14 0.15 K=3
403
+ verify 16: accepted 0/2 (K=3) + correction -> +1 tokens (110.5 ms) cpos=77
404
+ [margins] 0.19 1.45 1.88 1.37 1.83 K=2
405
+ verify 17: accepted 0/1 (K=2) + correction -> +1 tokens (97.8 ms) cpos=78
406
+ [margins] 1.90 2.92 1.11 2.68 1.86 K=3
407
+ verify 18: accepted 2/2 (K=3) + correction -> +3 tokens (113.2 ms) cpos=81
408
+ [margins] 2.16 0.11 1.20 0.32 0.64 K=2
409
+ verify 19: accepted 1/1 (K=2) + correction -> +2 tokens (96.8 ms) cpos=83
410
+ [margins] 1.07 0.27 0.61 0.21 0.54 K=2
411
+ verify 20: accepted 1/1 (K=2) + correction -> +2 tokens (101.4 ms) cpos=85
412
+ [margins] 4.84 1.92 0.46 0.66 1.37 K=3
413
+ verify 21: accepted 1/2 (K=3) + correction -> +2 tokens (112.4 ms) cpos=87
414
+ [margins] 3.66 3.10 1.47 2.70 0.72 K=3
415
+ verify 22: accepted 1/2 (K=3) + correction -> +2 tokens (115.0 ms) cpos=89
416
+ [margins] 1.41 1.18 0.76 1.49 0.34 K=2
417
+ verify 23: accepted 1/1 (K=2) + correction -> +2 tokens (98.6 ms) cpos=91
418
+ [margins] 4.98 0.33 1.75 2.50 0.12 K=2
419
+ verify 24: accepted 1/1 (K=2) + correction -> +2 tokens (101.4 ms) cpos=93
420
+ [margins] 3.48 0.12 4.38 3.31 0.31 K=2
421
+ verify 25: accepted 1/1 (K=2) + correction -> +2 tokens (96.8 ms) cpos=95
422
+ [margins] 7.46 1.18 3.63 2.02 0.06 K=2
423
+ verify 26: accepted 1/1 (K=2) + correction -> +2 tokens (102.1 ms) cpos=97
424
+ [margins] 2.36 2.19 0.27 3.31 0.14 K=3
425
+ verify 27: accepted 2/2 (K=3) + correction -> +3 tokens (117.7 ms) cpos=100
426
+ [margins] 0.19 0.09 2.48 1.59 0.91 K=2
427
+ verify 28: accepted 0/1 (K=2) + correction -> +1 tokens (98.2 ms) cpos=101
428
+ [margins] 3.61 0.93 0.19 3.50 2.16 K=2
429
+ verify 29: accepted 1/1 (K=2) + correction -> +2 tokens (100.5 ms) cpos=103
430
+ [margins] 5.84 7.10 1.14 4.16 6.96 K=3
431
+ verify 30: accepted 2/2 (K=3) + correction -> +3 tokens (112.5 ms) cpos=106
432
+ [margins] 7.16 5.43 0.47 5.96 1.75 K=3
433
+ verify 31: accepted 2/2 (K=3) + correction -> +3 tokens (112.8 ms) cpos=109
434
+ [margins] 10.09 1.44 2.74 1.42 1.25 K=2
435
+ verify 32: accepted 1/1 (K=2) + correction -> +2 tokens (97.8 ms) cpos=111
436
+ [margins] 2.17 6.56 0.43 0.33 1.52 K=3
437
+ verify 33: accepted 2/2 (K=3) + correction -> +3 tokens (116.4 ms) cpos=114
438
+ [margins] 0.81 10.17 2.07 1.53 2.11 K=6
439
+ verify 34: accepted 2/5 (K=6) + correction -> +3 tokens (161.1 ms) cpos=117
440
+ [margins] 1.10 5.43 4.06 3.21 9.54 K=6
441
+ verify 35: accepted 5/5 (K=6) + correction -> +6 tokens (161.5 ms) cpos=123
442
+ [margins] 9.90 11.81 4.67 12.77 4.09 K=6
443
+ verify 36: accepted 5/5 (K=6) + correction -> +6 tokens (177.3 ms) cpos=129
444
+ [margins] 11.37 12.48 11.74 8.61 3.31 K=6
445
+ verify 37: accepted 5/5 (K=6) + correction -> +6 tokens (160.2 ms) cpos=135
446
+ [margins] 3.77 1.85 0.99 0.76 3.05 K=3
447
+ verify 38: accepted 1/2 (K=3) + correction -> +2 tokens (114.5 ms) cpos=137
448
+ [margins] 2.78 0.20 1.82 1.09 2.98 K=2
449
+ verify 39: accepted 0/1 (K=2) + correction -> +1 tokens (101.3 ms) cpos=138
450
+ [margins] 0.39 1.48 1.11 1.96 0.24 K=2
451
+ verify 40: accepted 1/1 (K=2) + correction -> +2 tokens (103.2 ms) cpos=140
452
+ [margins] 3.68 1.54 0.49 0.49 0.51 K=3
453
+ verify 41: accepted 0/2 (K=3) + correction -> +1 tokens (112.2 ms) cpos=141
454
+ [margins] 1.83 0.43 1.11 0.56 0.77 K=2
455
+ verify 42: accepted 1/1 (K=2) + correction -> +2 tokens (99.2 ms) cpos=143
456
+ [margins] 0.41 0.21 0.84 0.13 0.87 K=2
457
+ verify 43: accepted 0/1 (K=2) + correction -> +1 tokens (105.1 ms) cpos=144
458
+ [margins] 1.31 2.40 1.46 2.92 0.41 K=3
459
+ verify 44: accepted 1/2 (K=3) + correction -> +2 tokens (115.7 ms) cpos=146
460
+ [margins] 7.00 2.07 1.94 1.72 0.24 K=5
461
+ verify 45: accepted 2/4 (K=5) + correction -> +3 tokens (140.1 ms) cpos=149
462
+ [margins] 0.91 2.29 3.09 0.83 1.36 K=4
463
+ verify 46: accepted 0/3 (K=4) + correction -> +1 tokens (128.8 ms) cpos=150
464
+ [margins] 1.21 0.91 0.16 2.17 0.00 K=2
465
+ verify 47: accepted 1/1 (K=2) + correction -> +2 tokens (101.8 ms) cpos=152
466
+ [margins] 4.04 0.59 0.89 1.42 1.17 K=2
467
+ verify 48: accepted 1/1 (K=2) + correction -> +2 tokens (99.1 ms) cpos=154
468
+ [margins] 0.96 0.92 3.13 0.48 0.28 K=2
469
+ verify 49: accepted 1/1 (K=2) + correction -> +2 tokens (105.0 ms) cpos=156
470
+ [margins] 4.76 2.16 0.83 0.27 0.41 K=3
471
+ verify 50: accepted 1/2 (K=3) + correction -> +2 tokens (110.4 ms) cpos=158
472
+ [margins] 3.40 1.79 1.87 0.70 0.85 K=4
473
+ verify 51: accepted 3/3 (K=4) + correction -> +4 tokens (131.1 ms) cpos=162
474
+ [margins] 5.42 3.08 0.38 0.19 0.87 K=3
475
+ verify 52: accepted 2/2 (K=3) + correction -> +3 tokens (114.6 ms) cpos=165
476
+ [margins] 3.36 1.57 1.95 0.98 2.39 K=4
477
+ verify 53: accepted 2/3 (K=4) + correction -> +3 tokens (128.8 ms) cpos=168
478
+ [margins] 4.59 1.87 6.81 2.60 0.30 K=5
479
+ verify 54: accepted 4/4 (K=5) + correction -> +5 tokens (143.9 ms) cpos=173
480
+ [margins] 0.05 1.14 3.35 0.06 0.20 K=2
481
+ verify 55: accepted 0/1 (K=2) + correction -> +1 tokens (100.9 ms) cpos=174
482
+ [margins] 0.06 0.46 1.69 3.08 0.87 K=2
483
+ verify 56: accepted 0/1 (K=2) + correction -> +1 tokens (104.3 ms) cpos=175
484
+ [margins] 0.24 1.83 1.19 0.52 1.96 K=3
485
+ verify 57: accepted 1/2 (K=3) + correction -> +2 tokens (118.6 ms) cpos=177
486
+ [margins] 5.29 1.47 2.24 1.71 2.80 K=2
487
+ verify 58: accepted 1/1 (K=2) + correction -> +2 tokens (101.0 ms) cpos=179
488
+ [margins] 3.37 3.22 3.07 1.45 0.89 K=4
489
+ verify 59: accepted 1/3 (K=4) + correction -> +2 tokens (130.8 ms) cpos=181
490
+ [margins] 5.26 0.82 0.40 1.47 1.37 K=2
491
+ verify 60: accepted 1/1 (K=2) + correction -> +2 tokens (99.7 ms) cpos=183
492
+ [margins] 2.70 3.66 5.48 2.95 3.14 K=6
493
+ verify 61: accepted 2/5 (K=6) + correction -> +3 tokens (160.2 ms) cpos=186
494
+ [margins] 3.20 3.05 0.34 0.49 0.64 K=3
495
+ verify 62: accepted 2/2 (K=3) + correction -> +3 tokens (117.9 ms) cpos=189
496
+ [margins] 3.54 1.52 0.43 0.18 0.80 K=3
497
+ verify 63: accepted 2/2 (K=3) + correction -> +3 tokens (119.4 ms) cpos=192
498
+ [margins] 6.39 0.67 0.28 0.97 0.18 K=2
499
+ verify 64: accepted 1/1 (K=2) + correction -> +2 tokens (100.2 ms) cpos=194
500
+ [margins] 0.83 1.55 3.25 1.28 0.17 K=4
501
+ verify 65: accepted 0/3 (K=4) + correction -> +1 tokens (131.2 ms) cpos=195
502
+ [margins] 1.85 1.14 1.72 3.84 0.87 K=2
503
+ verify 66: accepted 1/1 (K=2) + correction -> +2 tokens (103.5 ms) cpos=197
504
+ [margins] 1.15 0.92 6.52 1.54 2.74 K=2
505
+ verify 67: accepted 0/1 (K=2) + correction -> +1 tokens (99.9 ms) cpos=198
506
+ [margins] 1.92 1.17 1.27 4.42 1.07 K=2
507
+ verify 68: accepted 0/1 (K=2) + correction -> +1 tokens (105.6 ms) cpos=199
508
+ [margins] 4.06 7.58 7.47 1.67 5.25 K=6
509
+ verify 69: accepted 0/5 (K=6) + correction -> +1 tokens (165.7 ms) cpos=200
510
+ [margins] 7.45 3.49 0.82 3.05 2.49 K=3
511
+ verify 70: accepted 2/2 (K=3) + correction -> +3 tokens (112.5 ms) cpos=203
512
+ [margins] 4.33 3.72 3.21 2.49 1.29 K=5
513
+ verify 71: accepted 1/4 (K=5) + correction -> +2 tokens (143.6 ms) cpos=205
514
+ [margins] 2.44 0.73 0.11 0.42 1.18 K=2
515
+ verify 72: accepted 1/1 (K=2) + correction -> +2 tokens (99.6 ms) cpos=207
516
+ [margins] 0.13 1.15 2.02 3.37 4.36 K=2
517
+ verify 73: accepted 0/1 (K=2) + correction -> +1 tokens (103.8 ms) cpos=208
518
+ [margins] 11.24 1.00 1.72 4.77 1.02 K=2
519
+ verify 74: accepted 1/1 (K=2) + correction -> +2 tokens (101.3 ms) cpos=210
520
+ [margins] 3.14 6.22 1.72 1.90 5.02 K=6
521
+ verify 75: accepted 2/5 (K=6) + correction -> +3 tokens (159.6 ms) cpos=213
522
+ [margins] 1.00 0.01 0.29 1.02 0.39 K=2
523
+ verify 76: accepted 1/1 (K=2) + correction -> +2 tokens (99.9 ms) cpos=215
524
+ [margins] 3.74 2.06 4.65 5.00 2.81 K=6
525
+ verify 77: accepted 4/5 (K=6) + correction -> +5 tokens (166.4 ms) cpos=220
526
+ [margins] 4.02 1.59 2.76 0.36 1.83 K=4
527
+ verify 78: accepted 0/3 (K=4) + correction -> +1 tokens (133.1 ms) cpos=221
528
+ [margins] 11.63 2.60 2.77 0.07 4.97 K=4
529
+ verify 79: accepted 3/3 (K=4) + correction -> +4 tokens (131.7 ms) cpos=225
530
+ [margins] 11.06 7.13 0.94 0.99 0.24 K=3
531
+ verify 80: accepted 2/2 (K=3) + correction -> +3 tokens (120.2 ms) cpos=228
532
+
533
+ [spec] generated 202 tokens over 80 verifies: mean tokens/verify = 2.52 (block=5, max 6)
534
+ [spec] tokens: 1 16 1 16 1 16 1 16 1 16 1 16 1 16 1 16 1 16 1 16 1 16 1 16 1 16 1 16 1 16 1 16 1 16 1 16 1 16 1 16
535
+ [spec] SPEC-DECODE: 47.1 ms/tok = 21.24 tok/s (vs base M=1 68.6 ms/tok = 14.59 tok/s -> 1.46x)
536
+ [decode] PREFILL: 14 positions in 284.8 ms = 49.2 tok/s (20.341 ms/tok)
537
+ [spec] decoding 200 tokens (block=5, draft passes=1, adaptK=1.50, prompt=4 s=15)...
538
+ [margins] 0.16 0.38 1.56 1.71 0.08 K=2
539
+ verify 1: accepted 0/1 (K=2) + correction -> +1 tokens (102.1 ms) cpos=15
540
+ [margins] 0.46 1.61 0.33 0.85 0.22 K=3
541
+ verify 2: accepted 1/2 (K=3) + correction -> +2 tokens (111.7 ms) cpos=17
542
+ [margins] 0.37 0.53 0.59 0.92 0.69 K=2
543
+ verify 3: accepted 0/1 (K=2) + correction -> +1 tokens (95.4 ms) cpos=18
544
+ [margins] 1.32 0.81 0.33 0.40 1.13 K=2
545
+ verify 4: accepted 1/1 (K=2) + correction -> +2 tokens (99.4 ms) cpos=20
546
+ [margins] 0.84 1.20 0.59 0.51 1.36 K=2
547
+ verify 5: accepted 1/1 (K=2) + correction -> +2 tokens (94.4 ms) cpos=22
548
+ [margins] 1.78 1.01 0.65 0.37 0.42 K=2
549
+ verify 6: accepted 1/1 (K=2) + correction -> +2 tokens (100.8 ms) cpos=24
550
+ [margins] 0.17 0.60 0.21 0.70 3.89 K=2
551
+ verify 7: accepted 1/1 (K=2) + correction -> +2 tokens (96.6 ms) cpos=26
552
+ [margins] 0.53 0.13 0.62 0.51 0.57 K=2
553
+ verify 8: accepted 0/1 (K=2) + correction -> +1 tokens (100.7 ms) cpos=27
554
+ [margins] 0.21 0.42 0.30 1.04 0.31 K=2
555
+ verify 9: accepted 1/1 (K=2) + correction -> +2 tokens (103.5 ms) cpos=29
556
+ [margins] 2.61 2.68 0.59 3.31 0.58 K=3
557
+ verify 10: accepted 2/2 (K=3) + correction -> +3 tokens (112.8 ms) cpos=32
558
+ [margins] 1.03 1.84 0.10 0.09 2.52 K=3
559
+ verify 11: accepted 1/2 (K=3) + correction -> +2 tokens (107.3 ms) cpos=34
560
+ [margins] 0.50 0.15 0.05 1.12 0.28 K=2
561
+ verify 12: accepted 1/1 (K=2) + correction -> +2 tokens (99.1 ms) cpos=36
562
+ [margins] 0.65 1.84 12.52 1.18 2.14 K=4
563
+ verify 13: accepted 0/3 (K=4) + correction -> +1 tokens (122.9 ms) cpos=37
564
+ [margins] 0.63 2.84 0.19 0.67 0.20 K=3
565
+ verify 14: accepted 1/2 (K=3) + correction -> +2 tokens (112.8 ms) cpos=39
566
+ [margins] 2.88 0.96 0.07 0.16 0.10 K=2
567
+ verify 15: accepted 0/1 (K=2) + correction -> +1 tokens (99.4 ms) cpos=40
568
+ [margins] 3.05 3.41 3.29 1.35 0.82 K=4
569
+ verify 16: accepted 1/3 (K=4) + correction -> +2 tokens (127.1 ms) cpos=42
570
+ [margins] 0.58 0.40 1.48 0.44 0.74 K=2
571
+ verify 17: accepted 1/1 (K=2) + correction -> +2 tokens (101.4 ms) cpos=44
572
+ [margins] 1.88 8.17 0.74 1.11 2.30 K=3
573
+ verify 18: accepted 2/2 (K=3) + correction -> +3 tokens (109.5 ms) cpos=47
574
+ [margins] 3.57 7.41 2.36 1.21 2.88 K=4
575
+ verify 19: accepted 3/3 (K=4) + correction -> +4 tokens (125.7 ms) cpos=51
576
+ [margins] 2.84 0.23 0.04 1.91 0.13 K=2
577
+ verify 20: accepted 0/1 (K=2) + correction -> +1 tokens (101.0 ms) cpos=52
578
+ [margins] 2.00 1.79 1.28 0.68 3.93 K=3
579
+ verify 21: accepted 0/2 (K=3) + correction -> +1 tokens (108.4 ms) cpos=53
580
+ [margins] 7.63 3.19 2.09 0.74 2.61 K=4
581
+ verify 22: accepted 2/3 (K=4) + correction -> +3 tokens (129.2 ms) cpos=56
582
+ [margins] 4.48 1.80 3.86 3.96 0.90 K=5
583
+ verify 23: accepted 4/4 (K=5) + correction -> +5 tokens (136.4 ms) cpos=61
584
+ [margins] 1.77 1.43 0.18 1.26 0.11 K=2
585
+ verify 24: accepted 0/1 (K=2) + correction -> +1 tokens (98.1 ms) cpos=62
586
+ [margins] 1.27 0.51 1.71 2.00 2.49 K=2
587
+ verify 25: accepted 0/1 (K=2) + correction -> +1 tokens (100.1 ms) cpos=63
588
+ [margins] 1.05 1.54 0.98 1.31 0.02 K=3
589
+ verify 26: accepted 2/2 (K=3) + correction -> +3 tokens (114.4 ms) cpos=66
590
+ [margins] 0.98 3.76 0.73 0.20 0.32 K=3
591
+ verify 27: accepted 0/2 (K=3) + correction -> +1 tokens (112.9 ms) cpos=67
592
+ [margins] 1.68 0.41 0.79 2.20 0.99 K=2
593
+ verify 28: accepted 1/1 (K=2) + correction -> +2 tokens (102.9 ms) cpos=69
594
+ [margins] 0.05 0.91 1.19 0.32 0.19 K=2
595
+ verify 29: accepted 0/1 (K=2) + correction -> +1 tokens (99.2 ms) cpos=70
596
+ [margins] 1.85 0.79 0.66 0.32 0.70 K=2
597
+ verify 30: accepted 0/1 (K=2) + correction -> +1 tokens (102.1 ms) cpos=71
598
+ [margins] 0.27 1.16 2.62 0.25 0.21 K=2
599
+ verify 31: accepted 0/1 (K=2) + correction -> +1 tokens (103.1 ms) cpos=72
600
+ [margins] 1.59 0.07 2.23 0.65 0.25 K=2
601
+ verify 32: accepted 0/1 (K=2) + correction -> +1 tokens (97.7 ms) cpos=73
602
+ [margins] 6.14 1.82 0.57 0.49 0.85 K=3
603
+ verify 33: accepted 2/2 (K=3) + correction -> +3 tokens (113.2 ms) cpos=76
604
+ [margins] 2.01 4.25 1.67 1.33 1.41 K=4
605
+ verify 34: accepted 0/3 (K=4) + correction -> +1 tokens (127.5 ms) cpos=77
606
+ [margins] 4.76 1.75 1.65 2.62 0.84 K=5
607
+ verify 35: accepted 2/4 (K=5) + correction -> +3 tokens (139.8 ms) cpos=80
608
+ [margins] 1.74 5.40 1.91 2.61 8.86 K=6
609
+ verify 36: accepted 5/5 (K=6) + correction -> +6 tokens (162.7 ms) cpos=86
610
+ [margins] 4.97 5.21 4.16 3.07 0.20 K=5
611
+ verify 37: accepted 2/4 (K=5) + correction -> +3 tokens (129.5 ms) cpos=89
612
+ [margins] 8.80 3.99 2.99 1.08 0.46 K=4
613
+ verify 38: accepted 1/3 (K=4) + correction -> +2 tokens (122.7 ms) cpos=91
614
+ [margins] 5.98 3.97 0.31 1.52 3.71 K=3
615
+ verify 39: accepted 1/2 (K=3) + correction -> +2 tokens (111.3 ms) cpos=93
616
+ [margins] 3.78 0.59 5.38 4.01 0.42 K=2
617
+ verify 40: accepted 1/1 (K=2) + correction -> +2 tokens (100.5 ms) cpos=95
618
+ [margins] 4.26 1.06 2.54 2.08 4.23 K=2
619
+ verify 41: accepted 1/1 (K=2) + correction -> +2 tokens (105.6 ms) cpos=97
620
+ [margins] 2.80 5.80 2.25 0.88 2.75 K=4
621
+ verify 42: accepted 3/3 (K=4) + correction -> +4 tokens (122.7 ms) cpos=101
622
+ [margins] 2.43 8.37 2.70 4.64 3.36 K=6
623
+ verify 43: accepted 5/5 (K=6) + correction -> +6 tokens (147.2 ms) cpos=107
624
+ [margins] 2.13 8.97 2.57 5.45 3.53 K=6
625
+ verify 44: accepted 5/5 (K=6) + correction -> +6 tokens (153.8 ms) cpos=113
626
+ [margins] 1.40 7.89 2.50 4.33 3.13 K=6
627
+ verify 45: accepted 5/5 (K=6) + correction -> +6 tokens (148.2 ms) cpos=119
628
+ [margins] 3.02 7.91 3.43 4.42 2.92 K=6
629
+ verify 46: accepted 5/5 (K=6) + correction -> +6 tokens (156.2 ms) cpos=125
630
+ [margins] 3.49 9.27 3.86 5.00 3.51 K=6
631
+ verify 47: accepted 5/5 (K=6) + correction -> +6 tokens (163.1 ms) cpos=131
632
+ [margins] 2.52 7.72 3.63 4.55 3.58 K=6
633
+ verify 48: accepted 5/5 (K=6) + correction -> +6 tokens (155.8 ms) cpos=137
634
+ [margins] 3.35 9.43 3.33 3.90 3.50 K=6
635
+ verify 49: accepted 5/5 (K=6) + correction -> +6 tokens (149.7 ms) cpos=143
636
+ [margins] 4.02 9.34 3.34 4.46 2.77 K=6
637
+ verify 50: accepted 5/5 (K=6) + correction -> +6 tokens (156.7 ms) cpos=149
638
+ [margins] 2.48 10.25 2.61 5.45 3.30 K=6
639
+ verify 51: accepted 5/5 (K=6) + correction -> +6 tokens (151.5 ms) cpos=155
640
+ [margins] 3.50 10.01 2.75 4.97 3.05 K=6
641
+ verify 52: accepted 0/5 (K=6) + correction -> +1 tokens (157.0 ms) cpos=156
642
+ [margins] 5.71 2.68 6.91 2.15 1.43 K=5
643
+ verify 53: accepted 1/4 (K=5) + correction -> +2 tokens (138.7 ms) cpos=158
644
+ [margins] 5.00 1.16 1.39 0.25 7.19 K=2
645
+ verify 54: accepted 0/1 (K=2) + correction -> +1 tokens (104.3 ms) cpos=159
646
+ [margins] 7.64 1.30 0.25 0.64 0.35 K=2
647
+ verify 55: accepted 1/1 (K=2) + correction -> +2 tokens (104.5 ms) cpos=161
648
+ [margins] 8.47 5.58 1.63 3.76 1.63 K=6
649
+ verify 56: accepted 4/5 (K=6) + correction -> +5 tokens (157.1 ms) cpos=166
650
+ [margins] 1.01 0.25 5.45 0.39 1.11 K=2
651
+ verify 57: accepted 0/1 (K=2) + correction -> +1 tokens (108.1 ms) cpos=167
652
+ [margins] 10.05 3.56 7.92 8.11 4.05 K=6
653
+ verify 58: accepted 5/5 (K=6) + correction -> +6 tokens (165.8 ms) cpos=173
654
+ [margins] 5.40 1.66 1.48 3.13 0.77 K=3
655
+ verify 59: accepted 1/2 (K=3) + correction -> +2 tokens (118.5 ms) cpos=175
656
+ [margins] 7.48 8.47 0.91 0.56 4.98 K=3
657
+ verify 60: accepted 2/2 (K=3) + correction -> +3 tokens (118.5 ms) cpos=178
658
+ [margins] 16.85 13.37 11.71 10.05 8.02 K=6
659
+ verify 61: accepted 5/5 (K=6) + correction -> +6 tokens (167.2 ms) cpos=184
660
+ [margins] 5.08 7.59 10.11 12.57 5.04 K=6
661
+ verify 62: accepted 5/5 (K=6) + correction -> +6 tokens (162.2 ms) cpos=190
662
+ [margins] 9.89 5.14 8.92 7.33 10.45 K=6
663
+ verify 63: accepted 5/5 (K=6) + correction -> +6 tokens (169.4 ms) cpos=196
664
+ [margins] 13.08 9.14 8.38 9.30 2.09 K=6
665
+ verify 64: accepted 4/5 (K=6) + correction -> +5 tokens (165.1 ms) cpos=201
666
+ [margins] 3.62 10.15 10.85 15.04 7.39 K=6
667
+ verify 65: accepted 5/5 (K=6) + correction -> +6 tokens (163.0 ms) cpos=207
668
+ [margins] 10.30 6.17 8.72 10.41 11.32 K=6
669
+ verify 66: accepted 5/5 (K=6) + correction -> +6 tokens (167.5 ms) cpos=213
670
+ [margins] 12.93 14.42 12.35 9.42 4.25 K=6
671
+ verify 67: accepted 4/5 (K=6) + correction -> +5 tokens (163.5 ms) cpos=218
672
+
673
+ [spec] generated 204 tokens over 67 verifies: mean tokens/verify = 3.04 (block=5, max 6)
674
+ [spec] tokens: 5715 14 295 834 10175 14 8470 3939 436 344 2239 362 4197 339 35 126664 23809 16765 270 3077 305 1990 23714 538 9479 21121 35977 2184 47270 34788 14 832 851 588 366 99782 1082 20351 14319 35977
675
+ [spec] SPEC-DECODE: 40.8 ms/tok = 24.53 tok/s (vs base M=1 68.6 ms/tok = 14.59 tok/s -> 1.68x)
676
+ [decode] PREFILL: 82 positions in 1289.8 ms = 63.6 tok/s (15.729 ms/tok)
677
+ [spec] decoding 200 tokens (block=5, draft passes=1, adaptK=1.50, prompt=5 s=83)...
678
+ [margins] 6.43 3.13 1.20 0.45 3.60 K=3
679
+ verify 1: accepted 1/2 (K=3) + correction -> +2 tokens (117.9 ms) cpos=84
680
+ [margins] 0.08 0.12 0.52 0.85 1.56 K=2
681
+ verify 2: accepted 0/1 (K=2) + correction -> +1 tokens (98.5 ms) cpos=85
682
+ [margins] 0.18 0.56 0.30 0.75 1.01 K=2
683
+ verify 3: accepted 1/1 (K=2) + correction -> +2 tokens (98.9 ms) cpos=87
684
+ [margins] 5.70 2.01 4.26 0.55 0.64 K=4
685
+ verify 4: accepted 3/3 (K=4) + correction -> +4 tokens (129.0 ms) cpos=91
686
+ [margins] 12.37 6.46 6.89 4.06 3.38 K=6
687
+ verify 5: accepted 5/5 (K=6) + correction -> +6 tokens (163.3 ms) cpos=97
688
+ [margins] 8.79 3.44 2.08 0.63 1.47 K=4
689
+ verify 6: accepted 3/3 (K=4) + correction -> +4 tokens (128.7 ms) cpos=101
690
+ [margins] 5.17 0.42 0.22 0.39 0.06 K=2
691
+ verify 7: accepted 1/1 (K=2) + correction -> +2 tokens (102.0 ms) cpos=103
692
+ [margins] 2.64 0.88 1.77 0.63 1.76 K=2
693
+ verify 8: accepted 1/1 (K=2) + correction -> +2 tokens (104.3 ms) cpos=105
694
+ [margins] 8.81 7.34 6.17 5.27 0.83 K=5
695
+ verify 9: accepted 4/4 (K=5) + correction -> +5 tokens (141.2 ms) cpos=110
696
+ [margins] 13.15 7.81 8.61 4.36 13.05 K=6
697
+ verify 10: accepted 5/5 (K=6) + correction -> +6 tokens (166.1 ms) cpos=116
698
+ [margins] 10.19 14.47 7.96 1.10 1.75 K=4
699
+ verify 11: accepted 3/3 (K=4) + correction -> +4 tokens (131.9 ms) cpos=120
700
+ [margins] 3.09 1.56 1.42 0.24 3.29 K=3
701
+ verify 12: accepted 1/2 (K=3) + correction -> +2 tokens (114.3 ms) cpos=122
702
+ [margins] 8.35 4.18 7.93 2.96 2.01 K=6
703
+ verify 13: accepted 3/5 (K=6) + correction -> +4 tokens (181.2 ms) cpos=126
704
+ [margins] 12.86 5.66 4.70 10.93 7.44 K=6
705
+ verify 14: accepted 5/5 (K=6) + correction -> +6 tokens (181.0 ms) cpos=132
706
+ [margins] 12.24 14.59 1.62 9.31 9.81 K=6
707
+ verify 15: accepted 2/5 (K=6) + correction -> +3 tokens (158.8 ms) cpos=135
708
+ [margins] 12.92 14.93 9.17 7.25 0.93 K=5
709
+ verify 16: accepted 4/4 (K=5) + correction -> +5 tokens (148.3 ms) cpos=140
710
+ [margins] 7.78 2.13 6.93 0.33 5.57 K=4
711
+ verify 17: accepted 0/3 (K=4) + correction -> +1 tokens (127.9 ms) cpos=141
712
+ [margins] 7.67 5.76 2.76 0.11 1.62 K=4
713
+ verify 18: accepted 2/3 (K=4) + correction -> +3 tokens (128.5 ms) cpos=144
714
+ [margins] 10.85 5.86 4.33 5.05 2.86 K=6
715
+ verify 19: accepted 4/5 (K=6) + correction -> +5 tokens (161.5 ms) cpos=149
716
+ [margins] 13.69 11.24 4.49 6.33 10.06 K=6
717
+ verify 20: accepted 5/5 (K=6) + correction -> +6 tokens (160.7 ms) cpos=155
718
+ [margins] 14.97 10.81 11.49 8.17 1.11 K=5
719
+ verify 21: accepted 4/4 (K=5) + correction -> +5 tokens (148.4 ms) cpos=160
720
+ [margins] 8.21 1.04 10.85 6.02 4.27 K=2
721
+ verify 22: accepted 0/1 (K=2) + correction -> +1 tokens (100.6 ms) cpos=161
722
+ [margins] 9.44 6.04 1.53 4.17 0.87 K=5
723
+ verify 23: accepted 2/4 (K=5) + correction -> +3 tokens (140.7 ms) cpos=164
724
+ [margins] 9.58 7.03 2.74 9.25 6.31 K=6
725
+ verify 24: accepted 5/5 (K=6) + correction -> +6 tokens (161.2 ms) cpos=170
726
+ [margins] 1.52 7.41 10.73 10.58 11.20 K=6
727
+ verify 25: accepted 0/5 (K=6) + correction -> +1 tokens (166.6 ms) cpos=171
728
+ [margins] 1.77 11.35 7.98 1.96 8.53 K=6
729
+ verify 26: accepted 3/5 (K=6) + correction -> +4 tokens (167.1 ms) cpos=175
730
+ [margins] 13.99 13.45 11.89 8.00 1.15 K=5
731
+ verify 27: accepted 4/4 (K=5) + correction -> +5 tokens (149.2 ms) cpos=180
732
+ [margins] 16.37 16.38 5.45 2.80 7.66 K=6
733
+ verify 28: accepted 5/5 (K=6) + correction -> +6 tokens (160.3 ms) cpos=186
734
+ [margins] 5.00 9.08 4.87 7.00 1.55 K=6
735
+ verify 29: accepted 5/5 (K=6) + correction -> +6 tokens (166.6 ms) cpos=192
736
+ [margins] 13.04 16.25 7.42 9.97 11.10 K=6
737
+ verify 30: accepted 5/5 (K=6) + correction -> +6 tokens (161.3 ms) cpos=198
738
+ [margins] 7.97 11.58 11.85 12.80 2.60 K=6
739
+ verify 31: accepted 5/5 (K=6) + correction -> +6 tokens (163.7 ms) cpos=204
740
+ [margins] 9.12 2.91 3.21 6.09 0.97 K=5
741
+ verify 32: accepted 4/4 (K=5) + correction -> +5 tokens (142.6 ms) cpos=209
742
+ [margins] 14.06 11.71 6.65 8.81 11.51 K=6
743
+ verify 33: accepted 5/5 (K=6) + correction -> +6 tokens (161.4 ms) cpos=215
744
+ [margins] 16.08 14.28 10.40 8.70 0.35 K=5
745
+ verify 34: accepted 4/4 (K=5) + correction -> +5 tokens (150.7 ms) cpos=220
746
+ [margins] 19.48 13.46 2.26 0.75 6.07 K=4
747
+ verify 35: accepted 3/3 (K=4) + correction -> +4 tokens (129.3 ms) cpos=224
748
+ [margins] 11.27 4.49 2.60 8.39 1.23 K=5
749
+ verify 36: accepted 4/4 (K=5) + correction -> +5 tokens (144.8 ms) cpos=229
750
+ [margins] 18.88 10.79 7.33 10.19 9.57 K=6
751
+ verify 37: accepted 5/5 (K=6) + correction -> +6 tokens (165.6 ms) cpos=235
752
+ [margins] 7.19 10.65 7.83 5.92 0.85 K=5
753
+ verify 38: accepted 4/4 (K=5) + correction -> +5 tokens (151.2 ms) cpos=240
754
+ [margins] 19.11 14.53 5.12 2.87 6.08 K=6
755
+ verify 39: accepted 5/5 (K=6) + correction -> +6 tokens (162.2 ms) cpos=246
756
+ [margins] 10.52 5.20 0.76 8.95 10.63 K=3
757
+ verify 40: accepted 2/2 (K=3) + correction -> +3 tokens (122.2 ms) cpos=249
758
+ [margins] 14.68 11.09 6.78 10.81 11.23 K=6
759
+ verify 41: accepted 5/5 (K=6) + correction -> +6 tokens (161.0 ms) cpos=255
760
+ [margins] 14.68 13.66 11.48 7.06 6.84 K=6
761
+ verify 42: accepted 5/5 (K=6) + correction -> +6 tokens (182.5 ms) cpos=261
762
+ [margins] 14.71 9.23 5.94 7.58 1.54 K=6
763
+ verify 43: accepted 4/5 (K=6) + correction -> +5 tokens (162.3 ms) cpos=266
764
+ [margins] 12.39 10.52 1.09 4.89 6.97 K=3
765
+ verify 44: accepted 2/2 (K=3) + correction -> +3 tokens (120.2 ms) cpos=269
766
+ [margins] 14.69 12.46 6.68 10.59 9.66 K=6
767
+ verify 45: accepted 5/5 (K=6) + correction -> +6 tokens (165.3 ms) cpos=275
768
+ [margins] 13.00 16.06 11.22 8.15 8.80 K=6
769
+ verify 46: accepted 5/5 (K=6) + correction -> +6 tokens (168.4 ms) cpos=281
770
+ [margins] 14.44 6.50 6.47 7.37 0.36 K=5
771
+ verify 47: accepted 4/4 (K=5) + correction -> +5 tokens (145.3 ms) cpos=286
772
+
773
+ [spec] generated 204 tokens over 47 verifies: mean tokens/verify = 4.34 (block=5, max 6)
774
+ [spec] tokens: 1 50 201 671 15020 4181 362 2775 14817 14922 344 223 25 22765 7975 55 15 3180 16 1 50 201 671 15020 4181 362 2775 14817 14922 344 223 25 22765 7975 55 15 3180 16 1 50
775
+ [spec] SPEC-DECODE: 33.4 ms/tok = 29.94 tok/s (vs base M=1 68.6 ms/tok = 14.59 tok/s -> 2.05x)
776
+ [decode] PREFILL: 14 positions in 285.5 ms = 49.0 tok/s (20.392 ms/tok)
777
+ [spec] decoding 200 tokens (block=5, draft passes=1, adaptK=1.50, prompt=6 s=15)...
778
+ [margins] 0.25 4.42 0.71 0.94 0.59 K=3
779
+ verify 1: accepted 1/2 (K=3) + correction -> +2 tokens (115.1 ms) cpos=16
780
+ [margins] 3.22 0.57 1.14 1.92 2.93 K=2
781
+ verify 2: accepted 1/1 (K=2) + correction -> +2 tokens (98.9 ms) cpos=18
782
+ [margins] 2.29 0.38 0.85 2.01 3.46 K=2
783
+ verify 3: accepted 1/1 (K=2) + correction -> +2 tokens (102.9 ms) cpos=20
784
+ [margins] 2.27 1.12 1.82 2.99 1.82 K=2
785
+ verify 4: accepted 1/1 (K=2) + correction -> +2 tokens (97.5 ms) cpos=22
786
+ [margins] 3.58 1.14 1.97 3.32 2.57 K=2
787
+ verify 5: accepted 1/1 (K=2) + correction -> +2 tokens (101.8 ms) cpos=24
788
+ [margins] 3.14 3.84 1.90 1.15 4.28 K=4
789
+ verify 6: accepted 3/3 (K=4) + correction -> +4 tokens (117.2 ms) cpos=28
790
+ [margins] 2.17 4.72 3.34 2.04 3.00 K=6
791
+ verify 7: accepted 2/5 (K=6) + correction -> +3 tokens (143.5 ms) cpos=31
792
+ [margins] 0.49 3.10 3.63 3.48 0.26 K=5
793
+ verify 8: accepted 0/4 (K=5) + correction -> +1 tokens (137.4 ms) cpos=32
794
+ [margins] 6.77 1.18 0.23 7.44 2.39 K=2
795
+ verify 9: accepted 1/1 (K=2) + correction -> +2 tokens (96.9 ms) cpos=34
796
+ [margins] 2.67 5.87 0.57 0.12 5.66 K=3
797
+ verify 10: accepted 2/2 (K=3) + correction -> +3 tokens (110.3 ms) cpos=37
798
+ [margins] 11.60 6.83 3.32 4.36 6.25 K=6
799
+ verify 11: accepted 5/5 (K=6) + correction -> +6 tokens (148.8 ms) cpos=43
800
+ [margins] 13.34 8.08 8.66 11.85 6.97 K=6
801
+ verify 12: accepted 5/5 (K=6) + correction -> +6 tokens (156.2 ms) cpos=49
802
+ [margins] 14.24 10.58 9.88 10.07 10.45 K=6
803
+ verify 13: accepted 5/5 (K=6) + correction -> +6 tokens (151.9 ms) cpos=55
804
+ [margins] 13.81 9.82 12.52 7.65 9.54 K=6
805
+ verify 14: accepted 5/5 (K=6) + correction -> +6 tokens (155.2 ms) cpos=61
806
+ [margins] 17.49 10.53 12.91 10.16 10.48 K=6
807
+ verify 15: accepted 5/5 (K=6) + correction -> +6 tokens (152.2 ms) cpos=67
808
+ [margins] 16.59 10.84 9.23 10.44 9.02 K=6
809
+ verify 16: accepted 5/5 (K=6) + correction -> +6 tokens (159.2 ms) cpos=73
810
+ [margins] 14.93 11.09 11.95 8.04 9.28 K=6
811
+ verify 17: accepted 5/5 (K=6) + correction -> +6 tokens (153.0 ms) cpos=79
812
+ [margins] 16.19 10.92 12.85 12.61 12.42 K=6
813
+ verify 18: accepted 5/5 (K=6) + correction -> +6 tokens (157.8 ms) cpos=85
814
+ [margins] 14.46 10.88 11.01 11.56 9.83 K=6
815
+ verify 19: accepted 5/5 (K=6) + correction -> +6 tokens (155.5 ms) cpos=91
816
+ [margins] 16.56 11.10 12.72 9.53 11.35 K=6
817
+ verify 20: accepted 5/5 (K=6) + correction -> +6 tokens (161.8 ms) cpos=97
818
+ [margins] 16.74 11.42 13.92 11.04 12.02 K=6
819
+ verify 21: accepted 5/5 (K=6) + correction -> +6 tokens (153.4 ms) cpos=103
820
+ [margins] 13.99 11.07 12.39 12.04 9.34 K=6
821
+ verify 22: accepted 5/5 (K=6) + correction -> +6 tokens (162.1 ms) cpos=109
822
+ [margins] 15.63 10.88 13.10 13.12 10.23 K=6
823
+ verify 23: accepted 5/5 (K=6) + correction -> +6 tokens (156.3 ms) cpos=115
824
+ [margins] 13.26 10.39 13.38 11.83 11.12 K=6
825
+ verify 24: accepted 5/5 (K=6) + correction -> +6 tokens (160.5 ms) cpos=121
826
+ [margins] 14.76 10.59 13.47 9.99 11.08 K=6
827
+ verify 25: accepted 5/5 (K=6) + correction -> +6 tokens (157.7 ms) cpos=127
828
+ [margins] 16.90 11.26 14.16 11.79 10.59 K=6
829
+ verify 26: accepted 5/5 (K=6) + correction -> +6 tokens (176.1 ms) cpos=133
830
+ [margins] 14.92 9.76 13.04 9.45 10.02 K=6
831
+ verify 27: accepted 5/5 (K=6) + correction -> +6 tokens (153.6 ms) cpos=139
832
+ [margins] 13.75 9.82 13.81 12.25 10.47 K=6
833
+ verify 28: accepted 5/5 (K=6) + correction -> +6 tokens (160.7 ms) cpos=145
834
+ [margins] 15.96 9.74 12.75 10.44 9.18 K=6
835
+ verify 29: accepted 5/5 (K=6) + correction -> +6 tokens (156.0 ms) cpos=151
836
+ [margins] 14.86 10.12 13.09 8.78 11.65 K=6
837
+ verify 30: accepted 5/5 (K=6) + correction -> +6 tokens (160.3 ms) cpos=157
838
+ [margins] 14.51 9.35 11.11 12.79 9.11 K=6
839
+ verify 31: accepted 5/5 (K=6) + correction -> +6 tokens (159.2 ms) cpos=163
840
+ [margins] 17.49 10.05 12.74 10.73 8.97 K=6
841
+ verify 32: accepted 5/5 (K=6) + correction -> +6 tokens (160.9 ms) cpos=169
842
+ [margins] 15.84 10.99 13.12 7.96 12.79 K=6
843
+ verify 33: accepted 5/5 (K=6) + correction -> +6 tokens (158.7 ms) cpos=175
844
+ [margins] 15.49 9.94 12.40 6.64 10.56 K=6
845
+ verify 34: accepted 5/5 (K=6) + correction -> +6 tokens (162.2 ms) cpos=181
846
+ [margins] 17.16 10.47 12.95 9.33 11.14 K=6
847
+ verify 35: accepted 5/5 (K=6) + correction -> +6 tokens (158.8 ms) cpos=187
848
+ [margins] 16.75 10.98 13.72 10.89 8.72 K=6
849
+ verify 36: accepted 5/5 (K=6) + correction -> +6 tokens (163.9 ms) cpos=193
850
+ [margins] 16.93 10.23 12.95 10.38 8.27 K=6
851
+ verify 37: accepted 5/5 (K=6) + correction -> +6 tokens (157.7 ms) cpos=199
852
+ [margins] 19.10 10.19 12.96 10.60 7.98 K=6
853
+ verify 38: accepted 5/5 (K=6) + correction -> +6 tokens (163.0 ms) cpos=205
854
+ [margins] 16.51 10.76 13.36 10.81 9.53 K=6
855
+ verify 39: accepted 5/5 (K=6) + correction -> +6 tokens (158.2 ms) cpos=211
856
+ [margins] 17.49 9.63 12.79 5.35 10.07 K=6
857
+ verify 40: accepted 5/5 (K=6) + correction -> +6 tokens (163.2 ms) cpos=217
858
+
859
+ [spec] generated 203 tokens over 40 verifies: mean tokens/verify = 5.08 (block=5, max 6)
860
+ [spec] tokens: 1 16 1 16 1 16 1 16 1 16 1 16 1 16 1 16 403 31310 16 403 31310 16 403 31310 16 403 31310 16 403 31310 16 403 31310 16 403 31310 16 403 31310 16
861
+ [spec] SPEC-DECODE: 28.7 ms/tok = 34.89 tok/s (vs base M=1 68.6 ms/tok = 14.59 tok/s -> 2.39x)
862
+ [decode] PREFILL: 37 positions in 642.5 ms = 57.6 tok/s (17.364 ms/tok)
863
+ [spec] decoding 200 tokens (block=5, draft passes=1, adaptK=1.50, prompt=7 s=38)...
864
+ [margins] 0.90 2.24 0.11 0.27 0.46 K=3
865
+ verify 1: accepted 0/2 (K=3) + correction -> +1 tokens (116.1 ms) cpos=38
866
+ [margins] 0.10 6.02 0.50 1.81 0.24 K=3
867
+ verify 2: accepted 0/2 (K=3) + correction -> +1 tokens (111.4 ms) cpos=39
868
+ [margins] 1.62 0.20 5.08 0.50 0.73 K=2
869
+ verify 3: accepted 1/1 (K=2) + correction -> +2 tokens (99.4 ms) cpos=41
870
+ [margins] 2.83 0.15 1.53 6.03 0.06 K=2
871
+ verify 4: accepted 1/1 (K=2) + correction -> +2 tokens (95.0 ms) cpos=43
872
+ [margins] 1.13 1.96 0.23 0.05 0.20 K=3
873
+ verify 5: accepted 0/2 (K=3) + correction -> +1 tokens (111.5 ms) cpos=44
874
+ [margins] 0.93 0.11 0.95 0.58 0.51 K=2
875
+ verify 6: accepted 1/1 (K=2) + correction -> +2 tokens (95.3 ms) cpos=46
876
+ [margins] 1.23 0.60 2.10 0.23 1.67 K=2
877
+ verify 7: accepted 0/1 (K=2) + correction -> +1 tokens (100.5 ms) cpos=47
878
+ [margins] 1.98 2.73 0.27 1.36 1.25 K=3
879
+ verify 8: accepted 1/2 (K=3) + correction -> +2 tokens (113.5 ms) cpos=49
880
+ [margins] 8.14 7.45 5.85 2.81 0.95 K=5
881
+ verify 9: accepted 4/4 (K=5) + correction -> +5 tokens (138.0 ms) cpos=54
882
+ [margins] 0.07 0.51 3.32 3.65 2.33 K=2
883
+ verify 10: accepted 1/1 (K=2) + correction -> +2 tokens (101.4 ms) cpos=56
884
+ [margins] 5.32 2.34 1.44 0.50 0.23 K=3
885
+ verify 11: accepted 2/2 (K=3) + correction -> +3 tokens (108.2 ms) cpos=59
886
+ [margins] 3.87 2.80 0.71 0.19 0.17 K=3
887
+ verify 12: accepted 2/2 (K=3) + correction -> +3 tokens (113.4 ms) cpos=62
888
+ [margins] 1.30 0.09 0.12 0.10 0.55 K=2
889
+ verify 13: accepted 1/1 (K=2) + correction -> +2 tokens (102.1 ms) cpos=64
890
+ [margins] 6.59 8.84 2.89 2.77 4.30 K=6
891
+ verify 14: accepted 3/5 (K=6) + correction -> +4 tokens (155.4 ms) cpos=68
892
+ [margins] 5.22 3.18 0.00 0.74 0.71 K=3
893
+ verify 15: accepted 2/2 (K=3) + correction -> +3 tokens (108.9 ms) cpos=71
894
+ [margins] 2.28 2.15 4.05 0.33 1.54 K=4
895
+ verify 16: accepted 3/3 (K=4) + correction -> +4 tokens (124.9 ms) cpos=75
896
+ [margins] 2.78 9.55 1.46 3.24 3.25 K=3
897
+ verify 17: accepted 2/2 (K=3) + correction -> +3 tokens (111.9 ms) cpos=78
898
+ [margins] 8.55 8.33 6.76 7.80 9.39 K=6
899
+ verify 18: accepted 5/5 (K=6) + correction -> +6 tokens (160.2 ms) cpos=84
900
+ [margins] 8.06 9.12 6.77 3.97 0.30 K=5
901
+ verify 19: accepted 4/4 (K=5) + correction -> +5 tokens (141.2 ms) cpos=89
902
+ [margins] 8.30 6.70 4.01 0.88 0.64 K=4
903
+ verify 20: accepted 3/3 (K=4) + correction -> +4 tokens (128.1 ms) cpos=93
904
+ [margins] 4.78 3.63 2.63 1.18 0.38 K=4
905
+ verify 21: accepted 3/3 (K=4) + correction -> +4 tokens (125.8 ms) cpos=97
906
+ [margins] 9.19 13.02 10.49 9.59 4.17 K=6
907
+ verify 22: accepted 5/5 (K=6) + correction -> +6 tokens (159.5 ms) cpos=103
908
+ [margins] 10.06 4.68 2.62 1.37 9.48 K=4
909
+ verify 23: accepted 3/3 (K=4) + correction -> +4 tokens (129.8 ms) cpos=107
910
+ [margins] 4.98 2.65 9.85 0.80 3.91 K=4
911
+ verify 24: accepted 3/3 (K=4) + correction -> +4 tokens (130.7 ms) cpos=111
912
+ [margins] 10.80 6.39 8.33 4.40 7.22 K=6
913
+ verify 25: accepted 5/5 (K=6) + correction -> +6 tokens (167.1 ms) cpos=117
914
+ [margins] 10.69 9.06 3.06 6.98 5.17 K=6
915
+ verify 26: accepted 5/5 (K=6) + correction -> +6 tokens (161.4 ms) cpos=123
916
+ [margins] 14.25 5.24 9.30 8.57 0.13 K=5
917
+ verify 27: accepted 4/4 (K=5) + correction -> +5 tokens (162.2 ms) cpos=128
918
+ [margins] 5.53 5.44 4.11 2.61 2.87 K=6
919
+ verify 28: accepted 3/5 (K=6) + correction -> +4 tokens (160.3 ms) cpos=132
920
+ [margins] 4.96 2.24 3.59 2.65 0.91 K=5
921
+ verify 29: accepted 4/4 (K=5) + correction -> +5 tokens (141.8 ms) cpos=137
922
+ [margins] 2.39 0.36 0.23 0.74 1.09 K=2
923
+ verify 30: accepted 1/1 (K=2) + correction -> +2 tokens (101.8 ms) cpos=139
924
+ [margins] 6.28 0.98 5.60 7.46 5.65 K=2
925
+ verify 31: accepted 1/1 (K=2) + correction -> +2 tokens (104.7 ms) cpos=141
926
+ [margins] 9.96 7.95 5.89 1.56 0.59 K=5
927
+ verify 32: accepted 4/4 (K=5) + correction -> +5 tokens (144.9 ms) cpos=146
928
+ [margins] 7.20 9.85 10.72 12.55 2.23 K=6
929
+ verify 33: accepted 4/5 (K=6) + correction -> +5 tokens (168.0 ms) cpos=151
930
+ [margins] 2.68 1.92 0.54 7.88 0.26 K=3
931
+ verify 34: accepted 1/2 (K=3) + correction -> +2 tokens (122.7 ms) cpos=153
932
+ [margins] 2.43 2.62 0.59 1.61 3.92 K=3
933
+ verify 35: accepted 0/2 (K=3) + correction -> +1 tokens (108.8 ms) cpos=154
934
+ [margins] 6.38 1.89 0.83 3.32 1.64 K=3
935
+ verify 36: accepted 1/2 (K=3) + correction -> +2 tokens (115.7 ms) cpos=156
936
+ [margins] 5.80 3.12 0.25 1.75 1.85 K=3
937
+ verify 37: accepted 1/2 (K=3) + correction -> +2 tokens (109.9 ms) cpos=158
938
+ [margins] 3.76 1.07 2.73 0.34 1.11 K=2
939
+ verify 38: accepted 1/1 (K=2) + correction -> +2 tokens (107.3 ms) cpos=160
940
+ [margins] 4.21 4.35 3.57 0.66 2.46 K=4
941
+ verify 39: accepted 3/3 (K=4) + correction -> +4 tokens (123.6 ms) cpos=164
942
+ [margins] 4.04 9.14 4.30 3.93 4.38 K=6
943
+ verify 40: accepted 5/5 (K=6) + correction -> +6 tokens (149.2 ms) cpos=170
944
+ [margins] 3.42 10.00 3.39 6.18 4.08 K=6
945
+ verify 41: accepted 5/5 (K=6) + correction -> +6 tokens (154.8 ms) cpos=176
946
+ [margins] 3.19 9.88 2.90 7.75 4.63 K=6
947
+ verify 42: accepted 5/5 (K=6) + correction -> +6 tokens (150.2 ms) cpos=182
948
+ [margins] 2.21 9.64 4.14 6.78 5.27 K=6
949
+ verify 43: accepted 5/5 (K=6) + correction -> +6 tokens (155.5 ms) cpos=188
950
+ [margins] 1.39 9.14 4.18 6.29 4.95 K=6
951
+ verify 44: accepted 5/5 (K=6) + correction -> +6 tokens (150.1 ms) cpos=194
952
+ [margins] 0.97 9.17 3.51 6.29 4.72 K=6
953
+ verify 45: accepted 5/5 (K=6) + correction -> +6 tokens (155.4 ms) cpos=200
954
+ [margins] 1.45 9.81 3.84 5.34 4.26 K=6
955
+ verify 46: accepted 5/5 (K=6) + correction -> +6 tokens (150.4 ms) cpos=206
956
+ [margins] 1.91 8.81 3.53 5.06 5.15 K=6
957
+ verify 47: accepted 5/5 (K=6) + correction -> +6 tokens (156.2 ms) cpos=212
958
+ [margins] 2.59 8.61 3.35 4.86 4.77 K=6
959
+ verify 48: accepted 5/5 (K=6) + correction -> +6 tokens (148.9 ms) cpos=218
960
+ [margins] 2.78 8.38 4.41 4.10 4.96 K=6
961
+ verify 49: accepted 5/5 (K=6) + correction -> +6 tokens (156.2 ms) cpos=224
962
+ [margins] 2.94 8.82 3.58 4.41 4.62 K=6
963
+ verify 50: accepted 5/5 (K=6) + correction -> +6 tokens (150.1 ms) cpos=230
964
+ [margins] 2.74 8.39 4.09 2.96 4.43 K=6
965
+ verify 51: accepted 0/5 (K=6) + correction -> +1 tokens (156.3 ms) cpos=231
966
+ [margins] 8.66 0.97 5.34 4.16 2.21 K=2
967
+ verify 52: accepted 0/1 (K=2) + correction -> +1 tokens (104.3 ms) cpos=232
968
+ [margins] 5.92 2.53 7.56 1.29 0.11 K=4
969
+ verify 53: accepted 1/3 (K=4) + correction -> +2 tokens (126.5 ms) cpos=234
970
+ [margins] 2.77 1.24 3.19 1.19 0.19 K=2
971
+ verify 54: accepted 1/1 (K=2) + correction -> +2 tokens (104.1 ms) cpos=236
972
+ [margins] 2.99 0.70 6.03 0.88 0.37 K=2
973
+ verify 55: accepted 0/1 (K=2) + correction -> +1 tokens (100.9 ms) cpos=237
974
+
975
+ [spec] generated 200 tokens over 55 verifies: mean tokens/verify = 3.64 (block=5, max 6)
976
+ [spec] tokens: 343 12888 28 566 344 260 19075 3417 9508 5718 270 2753 294 270 6616 366 1512 1527 27819 13459 16 5715 270 12919 6329 1512 1527 940 223 19 27819 13459 16 455 3452 2753 344 979 7490 201
977
+ [spec] SPEC-DECODE: 35.4 ms/tok = 28.28 tok/s (vs base M=1 68.6 ms/tok = 14.59 tok/s -> 1.94x)
978
+ [decode] PREFILL: 12 positions in 250.5 ms = 47.9 tok/s (20.878 ms/tok)
979
+ [spec] decoding 200 tokens (block=5, draft passes=1, adaptK=1.50, prompt=8 s=13)...
980
+ [margins] 3.14 0.86 0.12 0.69 1.59 K=2
981
+ verify 1: accepted 1/1 (K=2) + correction -> +2 tokens (98.4 ms) cpos=14
982
+ [margins] 0.08 6.14 3.02 0.67 6.14 K=4
983
+ verify 2: accepted 0/3 (K=4) + correction -> +1 tokens (127.3 ms) cpos=15
984
+ [margins] 0.91 0.83 1.56 1.83 0.16 K=2
985
+ verify 3: accepted 1/1 (K=2) + correction -> +2 tokens (101.9 ms) cpos=17
986
+ [margins] 0.82 0.35 0.44 2.75 0.13 K=2
987
+ verify 4: accepted 1/1 (K=2) + correction -> +2 tokens (98.4 ms) cpos=19
988
+ [margins] 2.77 0.40 1.11 0.12 3.15 K=2
989
+ verify 5: accepted 1/1 (K=2) + correction -> +2 tokens (101.7 ms) cpos=21
990
+ [margins] 2.92 0.11 1.70 0.23 3.41 K=2
991
+ verify 6: accepted 1/1 (K=2) + correction -> +2 tokens (97.8 ms) cpos=23
992
+ [margins] 3.26 0.11 0.65 0.24 3.60 K=2
993
+ verify 7: accepted 1/1 (K=2) + correction -> +2 tokens (101.9 ms) cpos=25
994
+ [margins] 4.15 0.28 1.54 0.11 4.26 K=2
995
+ verify 8: accepted 1/1 (K=2) + correction -> +2 tokens (98.2 ms) cpos=27
996
+ [margins] 5.13 0.24 1.91 0.44 5.05 K=2
997
+ verify 9: accepted 1/1 (K=2) + correction -> +2 tokens (102.0 ms) cpos=29
998
+ [margins] 5.82 0.44 2.75 0.43 4.48 K=2
999
+ verify 10: accepted 1/1 (K=2) + correction -> +2 tokens (98.2 ms) cpos=31
1000
+ [margins] 5.75 0.02 4.17 0.89 4.33 K=2
1001
+ verify 11: accepted 1/1 (K=2) + correction -> +2 tokens (102.2 ms) cpos=33
1002
+ [margins] 5.61 0.12 4.29 0.63 3.96 K=2
1003
+ verify 12: accepted 1/1 (K=2) + correction -> +2 tokens (98.5 ms) cpos=35
1004
+ [margins] 6.12 0.08 5.58 0.61 4.70 K=2
1005
+ verify 13: accepted 1/1 (K=2) + correction -> +2 tokens (102.2 ms) cpos=37
1006
+ [margins] 5.75 0.02 6.25 0.29 4.04 K=2
1007
+ verify 14: accepted 1/1 (K=2) + correction -> +2 tokens (98.6 ms) cpos=39
1008
+ [margins] 6.58 0.36 7.45 0.38 4.84 K=2
1009
+ verify 15: accepted 1/1 (K=2) + correction -> +2 tokens (102.4 ms) cpos=41
1010
+ [margins] 6.14 0.37 5.41 0.61 4.06 K=2
1011
+ verify 16: accepted 1/1 (K=2) + correction -> +2 tokens (99.2 ms) cpos=43
1012
+ [margins] 6.96 0.53 6.08 1.09 4.41 K=2
1013
+ verify 17: accepted 1/1 (K=2) + correction -> +2 tokens (102.9 ms) cpos=45
1014
+ [margins] 5.95 1.26 4.55 0.74 3.41 K=2
1015
+ verify 18: accepted 0/1 (K=2) + correction -> +1 tokens (95.9 ms) cpos=46
1016
+ [margins] 5.36 3.16 6.41 0.01 0.50 K=4
1017
+ verify 19: accepted 3/3 (K=4) + correction -> +4 tokens (125.8 ms) cpos=50
1018
+ [margins] 2.74 2.29 0.61 1.90 3.94 K=3
1019
+ verify 20: accepted 2/2 (K=3) + correction -> +3 tokens (113.3 ms) cpos=53
1020
+ [margins] 6.77 8.82 0.32 1.08 1.52 K=3
1021
+ verify 21: accepted 2/2 (K=3) + correction -> +3 tokens (111.6 ms) cpos=56
1022
+ [margins] 3.10 1.80 1.56 2.61 2.99 K=6
1023
+ verify 22: accepted 2/5 (K=6) + correction -> +3 tokens (152.3 ms) cpos=59
1024
+ [margins] 1.77 1.82 3.71 5.28 0.31 K=5
1025
+ verify 23: accepted 0/4 (K=5) + correction -> +1 tokens (139.5 ms) cpos=60
1026
+ [margins] 5.89 4.03 5.48 6.96 2.77 K=6
1027
+ verify 24: accepted 5/5 (K=6) + correction -> +6 tokens (154.5 ms) cpos=66
1028
+ [margins] 7.39 10.81 8.44 1.98 0.79 K=5
1029
+ verify 25: accepted 3/4 (K=5) + correction -> +4 tokens (140.4 ms) cpos=70
1030
+ [margins] 2.23 4.21 0.58 4.72 2.06 K=3
1031
+ verify 26: accepted 0/2 (K=3) + correction -> +1 tokens (115.1 ms) cpos=71
1032
+ [margins] 0.16 3.48 3.56 2.89 1.67 K=6
1033
+ verify 27: accepted 0/5 (K=6) + correction -> +1 tokens (159.4 ms) cpos=72
1034
+ [margins] 2.66 3.28 4.26 6.14 2.85 K=6
1035
+ verify 28: accepted 5/5 (K=6) + correction -> +6 tokens (155.8 ms) cpos=78
1036
+ [margins] 5.85 8.12 6.17 0.03 0.93 K=4
1037
+ verify 29: accepted 3/3 (K=4) + correction -> +4 tokens (128.4 ms) cpos=82
1038
+ [margins] 3.80 6.93 4.12 5.12 2.08 K=6
1039
+ verify 30: accepted 5/5 (K=6) + correction -> +6 tokens (162.2 ms) cpos=88
1040
+ [margins] 2.34 6.57 1.20 7.30 0.76 K=3
1041
+ verify 31: accepted 2/2 (K=3) + correction -> +3 tokens (111.8 ms) cpos=91
1042
+ [margins] 3.94 13.01 0.78 2.76 4.42 K=3
1043
+ verify 32: accepted 2/2 (K=3) + correction -> +3 tokens (115.7 ms) cpos=94
1044
+ [margins] 4.92 8.46 8.46 5.81 3.65 K=6
1045
+ verify 33: accepted 5/5 (K=6) + correction -> +6 tokens (164.8 ms) cpos=100
1046
+ [margins] 5.29 7.88 3.85 7.03 1.45 K=5
1047
+ verify 34: accepted 4/4 (K=5) + correction -> +5 tokens (143.2 ms) cpos=105
1048
+ [margins] 2.32 4.73 3.71 6.06 6.22 K=6
1049
+ verify 35: accepted 5/5 (K=6) + correction -> +6 tokens (158.3 ms) cpos=111
1050
+ [margins] 7.38 4.36 1.80 1.55 6.08 K=6
1051
+ verify 36: accepted 5/5 (K=6) + correction -> +6 tokens (164.2 ms) cpos=117
1052
+ [margins] 3.15 6.52 4.22 8.44 7.61 K=6
1053
+ verify 37: accepted 5/5 (K=6) + correction -> +6 tokens (158.3 ms) cpos=123
1054
+ [margins] 11.24 4.79 10.20 6.33 12.11 K=6
1055
+ verify 38: accepted 5/5 (K=6) + correction -> +6 tokens (179.2 ms) cpos=129
1056
+ [margins] 3.91 5.10 7.09 7.39 4.58 K=6
1057
+ verify 39: accepted 5/5 (K=6) + correction -> +6 tokens (161.6 ms) cpos=135
1058
+ [margins] 10.83 6.33 10.14 5.56 9.80 K=6
1059
+ verify 40: accepted 5/5 (K=6) + correction -> +6 tokens (168.5 ms) cpos=141
1060
+ [margins] 3.67 6.98 3.33 8.54 6.88 K=6
1061
+ verify 41: accepted 5/5 (K=6) + correction -> +6 tokens (161.2 ms) cpos=147
1062
+ [margins] 9.98 2.49 4.75 2.90 7.73 K=6
1063
+ verify 42: accepted 5/5 (K=6) + correction -> +6 tokens (167.1 ms) cpos=153
1064
+ [margins] 2.75 7.37 6.80 6.38 5.22 K=6
1065
+ verify 43: accepted 5/5 (K=6) + correction -> +6 tokens (161.8 ms) cpos=159
1066
+ [margins] 8.03 4.03 3.35 2.89 9.59 K=6
1067
+ verify 44: accepted 5/5 (K=6) + correction -> +6 tokens (165.9 ms) cpos=165
1068
+ [margins] 3.55 5.71 3.57 7.79 7.62 K=6
1069
+ verify 45: accepted 5/5 (K=6) + correction -> +6 tokens (160.1 ms) cpos=171
1070
+ [margins] 10.88 6.37 9.15 5.15 10.63 K=6
1071
+ verify 46: accepted 5/5 (K=6) + correction -> +6 tokens (167.3 ms) cpos=177
1072
+ [margins] 4.23 5.06 4.40 7.36 8.68 K=6
1073
+ verify 47: accepted 5/5 (K=6) + correction -> +6 tokens (159.7 ms) cpos=183
1074
+ [margins] 10.42 6.43 11.57 4.39 13.03 K=6
1075
+ verify 48: accepted 5/5 (K=6) + correction -> +6 tokens (168.2 ms) cpos=189
1076
+ [margins] 4.50 5.78 7.52 8.69 8.14 K=6
1077
+ verify 49: accepted 5/5 (K=6) + correction -> +6 tokens (160.4 ms) cpos=195
1078
+ [margins] 11.28 6.24 11.15 8.14 8.30 K=6
1079
+ verify 50: accepted 5/5 (K=6) + correction -> +6 tokens (167.7 ms) cpos=201
1080
+ [margins] 4.24 5.55 5.98 8.11 5.82 K=6
1081
+ verify 51: accepted 5/5 (K=6) + correction -> +6 tokens (161.5 ms) cpos=207
1082
+ [margins] 11.87 6.54 12.01 9.27 9.55 K=6
1083
+ verify 52: accepted 5/5 (K=6) + correction -> +6 tokens (167.5 ms) cpos=213
1084
+
1085
+ [spec] generated 201 tokens over 52 verifies: mean tokens/verify = 3.87 (block=5, max 6)
1086
+ [spec] tokens: 1 201 1 201 1 201 1 201 1 201 1 201 1 201 1 201 1 201 1 201 1 201 1 201 1 201 1 201 1 201 1 201 3085 344 270 6102 294 8760 33 9361
1087
+ [spec] SPEC-DECODE: 34.6 ms/tok = 28.92 tok/s (vs base M=1 68.6 ms/tok = 14.59 tok/s -> 1.98x)
1088
+
1089
+ [blksweep] BLK passes adaptK->eff prompt | mean tok/verify | ms/tok | tok/s | vs base 14.59 tok/s
1090
+ [blksweep] 5 1 1.50 1.50 0 | 3.30 | 39.2 | 25.49 | 1.75x
1091
+ [blksweep] 5 1 1.50 1.50 1 | 4.25 | 33.1 | 30.20 | 2.07x
1092
+ [blksweep] 5 1 1.50 1.50 2 | 4.00 | 34.9 | 28.62 | 1.96x
1093
+ [blksweep] 5 1 1.50 1.50 3 | 2.52 | 47.1 | 21.24 | 1.46x
1094
+ [blksweep] 5 1 1.50 1.50 4 | 3.04 | 40.8 | 24.53 | 1.68x
1095
+ [blksweep] 5 1 1.50 1.50 5 | 4.34 | 33.4 | 29.94 | 2.05x
1096
+ [blksweep] 5 1 1.50 1.50 6 | 5.08 | 28.7 | 34.89 | 2.39x
1097
+ [blksweep] 5 1 1.50 1.50 7 | 3.64 | 35.4 | 28.28 | 1.94x
1098
+ [blksweep] 5 1 1.50 1.50 8 | 3.87 | 34.6 | 28.92 | 1.98x
1099
+ [decode] mem 106.9/122.8 GiB