HelloSun commited on
Commit
c021d6e
·
verified ·
1 Parent(s): a6291a0

Delete results/strict_headline.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. results/strict_headline.json +0 -214
results/strict_headline.json DELETED
@@ -1,214 +0,0 @@
1
- {
2
- "config": {
3
- "model": "/root/models/minicpm-v4.7-35b-a3b-q4_k_m.gguf",
4
- "bins": "['base=/root/app/llama.cpp/build/bin/epq-chat']",
5
- "bin": "",
6
- "rounds": "9",
7
- "strict_mb": "8192",
8
- "strict_data_mb": "0",
9
- "metric": "anon",
10
- "cli": "chat",
11
- "ram_budget_mb": "0",
12
- "ctx": "4096",
13
- "threads": "8",
14
- "ubatch": "128",
15
- "n_predict": "48",
16
- "no_prefetch": "False",
17
- "no_lock": "False",
18
- "extra": "[]",
19
- "label": "headline",
20
- "workdir": "/tmp/bench",
21
- "keep_text": "False"
22
- },
23
- "results": [
24
- {
25
- "arm": "base",
26
- "round": 0,
27
- "wall_s": 12.26,
28
- "returncode": 0,
29
- "strict_ok": true,
30
- "peak_anon_mb": 7781.2,
31
- "peak_swap_mb": 0.0,
32
- "vm_lck_mb": 7642.9,
33
- "swap_samples_over_zero": 0,
34
- "peak_file_mb": 1321.9,
35
- "decode_tokens": 47,
36
- "decode_s": 7.09,
37
- "decode_tok_s": 6.63,
38
- "slots": 3714,
39
- "arena_gib": 6.35,
40
- "hit_rate_pct": 68.6775,
41
- "ssd_mb_per_token": 220.29495239257812,
42
- "prefetch_used_pct": 0,
43
- "stdout_tail": "ixture-of-Experts (MoE) models route tokens by using a lightweight gating network to dynamically select a small subset of specialized \"expert\"在 $100 \\mathrm{~m"
44
- },
45
- {
46
- "arm": "base",
47
- "round": 1,
48
- "wall_s": 13.43,
49
- "returncode": 0,
50
- "strict_ok": true,
51
- "peak_anon_mb": 7799.5,
52
- "peak_swap_mb": 0.0,
53
- "vm_lck_mb": 7650.5,
54
- "swap_samples_over_zero": 0,
55
- "peak_file_mb": 1321.9,
56
- "decode_tokens": 47,
57
- "decode_s": 7.57,
58
- "decode_tok_s": 6.21,
59
- "slots": 3714,
60
- "arena_gib": 6.35,
61
- "hit_rate_pct": 72.1095,
62
- "ssd_mb_per_token": 192.59357452392578,
63
- "prefetch_used_pct": 0,
64
- "stdout_tail": ".gguf\n[sdq] ctx=4096 threads=8 ubatch=128 RAM(RSS)=8568 MB 模型載入 0.7s\nassistant: 1. **CPU** 是電腦的「大腦」,負責執行所有運算與指令,但本身不儲存大量資料。\n2. **RAM** 是「短期工作記憶」,存放 CPU 正在處理"
65
- },
66
- {
67
- "arm": "base",
68
- "round": 2,
69
- "wall_s": 12.62,
70
- "returncode": 0,
71
- "strict_ok": true,
72
- "peak_anon_mb": 7799.5,
73
- "peak_swap_mb": 0.0,
74
- "vm_lck_mb": 7650.5,
75
- "swap_samples_over_zero": 0,
76
- "peak_file_mb": 1322.1,
77
- "decode_tokens": 47,
78
- "decode_s": 7.42,
79
- "decode_tok_s": 6.34,
80
- "slots": 3714,
81
- "arena_gib": 6.35,
82
- "hit_rate_pct": 68.6112,
83
- "ssd_mb_per_token": 214.11705017089844,
84
- "prefetch_used_pct": 0,
85
- "stdout_tail": "*Analyze the Request:**\n * Task: List five daily work tasks that involve using a computer.\n * Constraint: Each task must be explained in one sentence."
86
- },
87
- {
88
- "arm": "base",
89
- "round": 3,
90
- "wall_s": 15.36,
91
- "returncode": 0,
92
- "strict_ok": true,
93
- "peak_anon_mb": 7781.6,
94
- "peak_swap_mb": 0.0,
95
- "vm_lck_mb": 7642.9,
96
- "swap_samples_over_zero": 0,
97
- "peak_file_mb": 1321.8,
98
- "decode_tokens": 47,
99
- "decode_s": 9.05,
100
- "decode_tok_s": 5.19,
101
- "slots": 3714,
102
- "arena_gib": 6.35,
103
- "hit_rate_pct": 68.6775,
104
- "ssd_mb_per_token": 220.29495239257812,
105
- "prefetch_used_pct": 0,
106
- "stdout_tail": "ixture-of-Experts (MoE) models route tokens by using a lightweight gating network to dynamically select a small subset of specialized \"expert\"在 $100 \\mathrm{~m"
107
- },
108
- {
109
- "arm": "base",
110
- "round": 4,
111
- "wall_s": 13.22,
112
- "returncode": 0,
113
- "strict_ok": true,
114
- "peak_anon_mb": 7799.1,
115
- "peak_swap_mb": 0.0,
116
- "vm_lck_mb": 7650.5,
117
- "swap_samples_over_zero": 0,
118
- "peak_file_mb": 1322.1,
119
- "decode_tokens": 47,
120
- "decode_s": 8.13,
121
- "decode_tok_s": 5.78,
122
- "slots": 3714,
123
- "arena_gib": 6.35,
124
- "hit_rate_pct": 72.1152,
125
- "ssd_mb_per_token": 192.5516128540039,
126
- "prefetch_used_pct": 0,
127
- "stdout_tail": ".gguf\n[sdq] ctx=4096 threads=8 ubatch=128 RAM(RSS)=8568 MB 模型載入 0.7s\nassistant: 1. **CPU** 是電腦的「大腦」,負責執行所有運算與指令,但本身不儲存大量資料。\n2. **RAM** 是「短期工作記憶」,存放 CPU 正在處理"
128
- },
129
- {
130
- "arm": "base",
131
- "round": 5,
132
- "wall_s": 14.53,
133
- "returncode": 0,
134
- "strict_ok": true,
135
- "peak_anon_mb": 7799.2,
136
- "peak_swap_mb": 0.0,
137
- "vm_lck_mb": 7650.5,
138
- "swap_samples_over_zero": 0,
139
- "peak_file_mb": 1322.1,
140
- "decode_tokens": 47,
141
- "decode_s": 8.67,
142
- "decode_tok_s": 5.42,
143
- "slots": 3714,
144
- "arena_gib": 6.35,
145
- "hit_rate_pct": 68.617,
146
- "ssd_mb_per_token": 214.07604217529297,
147
- "prefetch_used_pct": 0,
148
- "stdout_tail": "*Analyze the Request:**\n * Task: List five daily work tasks that involve using a computer.\n * Constraint: Each task must be explained in one sentence."
149
- },
150
- {
151
- "arm": "base",
152
- "round": 6,
153
- "wall_s": 15.05,
154
- "returncode": 0,
155
- "strict_ok": true,
156
- "peak_anon_mb": 7781.2,
157
- "peak_swap_mb": 0.0,
158
- "vm_lck_mb": 7642.9,
159
- "swap_samples_over_zero": 0,
160
- "peak_file_mb": 1321.6,
161
- "decode_tokens": 47,
162
- "decode_s": 8.9,
163
- "decode_tok_s": 5.28,
164
- "slots": 3714,
165
- "arena_gib": 6.35,
166
- "hit_rate_pct": 68.6717,
167
- "ssd_mb_per_token": 220.3378677368164,
168
- "prefetch_used_pct": 0,
169
- "stdout_tail": "ixture-of-Experts (MoE) models route tokens by using a lightweight gating network to dynamically select a small subset of specialized \"expert\"在 $100 \\mathrm{~m"
170
- },
171
- {
172
- "arm": "base",
173
- "round": 7,
174
- "wall_s": 13.28,
175
- "returncode": 0,
176
- "strict_ok": true,
177
- "peak_anon_mb": 7799.3,
178
- "peak_swap_mb": 0.0,
179
- "vm_lck_mb": 7650.5,
180
- "swap_samples_over_zero": 0,
181
- "peak_file_mb": 1322.1,
182
- "decode_tokens": 47,
183
- "decode_s": 7.51,
184
- "decode_tok_s": 6.26,
185
- "slots": 3714,
186
- "arena_gib": 6.35,
187
- "hit_rate_pct": 72.1209,
188
- "ssd_mb_per_token": 192.51060485839844,
189
- "prefetch_used_pct": 0,
190
- "stdout_tail": ".gguf\n[sdq] ctx=4096 threads=8 ubatch=128 RAM(RSS)=8568 MB 模型載入 0.8s\nassistant: 1. **CPU** 是電腦的「大腦」,負責執行所有運算與指令,但本身不儲存大量資料。\n2. **RAM** 是「短期工作記憶」,存放 CPU 正在處理"
191
- },
192
- {
193
- "arm": "base",
194
- "round": 8,
195
- "wall_s": 14.18,
196
- "returncode": 0,
197
- "strict_ok": true,
198
- "peak_anon_mb": 7799.1,
199
- "peak_swap_mb": 0.0,
200
- "vm_lck_mb": 7650.5,
201
- "swap_samples_over_zero": 0,
202
- "peak_file_mb": 1322.4,
203
- "decode_tokens": 47,
204
- "decode_s": 8.07,
205
- "decode_tok_s": 5.83,
206
- "slots": 3714,
207
- "arena_gib": 6.35,
208
- "hit_rate_pct": 68.617,
209
- "ssd_mb_per_token": 214.07604217529297,
210
- "prefetch_used_pct": 0,
211
- "stdout_tail": "*Analyze the Request:**\n * Task: List five daily work tasks that involve using a computer.\n * Constraint: Each task must be explained in one sentence."
212
- }
213
- ]
214
- }