Commit
•
9a41c6d
1
Parent(s):
3e28d79
Upload cpu_inference_transformers_text-generation_openai-community/gpt2/benchmark.json with huggingface_hub
Browse files
cpu_inference_transformers_text-generation_openai-community/gpt2/benchmark.json
ADDED
@@ -0,0 +1,236 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"config": {
|
3 |
+
"name": "cpu_inference_transformers_text-generation_openai-community/gpt2",
|
4 |
+
"backend": {
|
5 |
+
"name": "pytorch",
|
6 |
+
"version": "2.3.0+cpu",
|
7 |
+
"_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
|
8 |
+
"task": "text-generation",
|
9 |
+
"model": "openai-community/gpt2",
|
10 |
+
"library": "transformers",
|
11 |
+
"device": "cpu",
|
12 |
+
"device_ids": null,
|
13 |
+
"seed": 42,
|
14 |
+
"inter_op_num_threads": null,
|
15 |
+
"intra_op_num_threads": null,
|
16 |
+
"hub_kwargs": {
|
17 |
+
"revision": "main",
|
18 |
+
"force_download": false,
|
19 |
+
"local_files_only": false,
|
20 |
+
"trust_remote_code": false
|
21 |
+
},
|
22 |
+
"no_weights": true,
|
23 |
+
"device_map": null,
|
24 |
+
"torch_dtype": null,
|
25 |
+
"eval_mode": true,
|
26 |
+
"to_bettertransformer": false,
|
27 |
+
"low_cpu_mem_usage": null,
|
28 |
+
"attn_implementation": null,
|
29 |
+
"cache_implementation": null,
|
30 |
+
"autocast_enabled": false,
|
31 |
+
"autocast_dtype": null,
|
32 |
+
"torch_compile": false,
|
33 |
+
"torch_compile_target": "forward",
|
34 |
+
"torch_compile_config": {},
|
35 |
+
"quantization_scheme": null,
|
36 |
+
"quantization_config": {},
|
37 |
+
"deepspeed_inference": false,
|
38 |
+
"deepspeed_inference_config": {},
|
39 |
+
"peft_type": null,
|
40 |
+
"peft_config": {}
|
41 |
+
},
|
42 |
+
"scenario": {
|
43 |
+
"name": "inference",
|
44 |
+
"_target_": "optimum_benchmark.scenarios.inference.scenario.InferenceScenario",
|
45 |
+
"iterations": 1,
|
46 |
+
"duration": 1,
|
47 |
+
"warmup_runs": 1,
|
48 |
+
"input_shapes": {
|
49 |
+
"batch_size": 1,
|
50 |
+
"num_choices": 2,
|
51 |
+
"sequence_length": 2
|
52 |
+
},
|
53 |
+
"new_tokens": null,
|
54 |
+
"latency": true,
|
55 |
+
"memory": true,
|
56 |
+
"energy": true,
|
57 |
+
"forward_kwargs": {},
|
58 |
+
"generate_kwargs": {
|
59 |
+
"max_new_tokens": 2,
|
60 |
+
"min_new_tokens": 2
|
61 |
+
},
|
62 |
+
"call_kwargs": {
|
63 |
+
"num_inference_steps": 2
|
64 |
+
}
|
65 |
+
},
|
66 |
+
"launcher": {
|
67 |
+
"name": "process",
|
68 |
+
"_target_": "optimum_benchmark.launchers.process.launcher.ProcessLauncher",
|
69 |
+
"device_isolation": false,
|
70 |
+
"device_isolation_action": "error",
|
71 |
+
"start_method": "spawn"
|
72 |
+
},
|
73 |
+
"environment": {
|
74 |
+
"cpu": " AMD EPYC 7763 64-Core Processor",
|
75 |
+
"cpu_count": 4,
|
76 |
+
"cpu_ram_mb": 16757.342208,
|
77 |
+
"system": "Linux",
|
78 |
+
"machine": "x86_64",
|
79 |
+
"platform": "Linux-6.5.0-1018-azure-x86_64-with-glibc2.35",
|
80 |
+
"processor": "x86_64",
|
81 |
+
"python_version": "3.10.14",
|
82 |
+
"optimum_benchmark_version": "0.2.0",
|
83 |
+
"optimum_benchmark_commit": "62a285c19141f06dfefa3e3372c4d089bc6a2a7b",
|
84 |
+
"transformers_version": "4.40.2",
|
85 |
+
"transformers_commit": null,
|
86 |
+
"accelerate_version": "0.30.0",
|
87 |
+
"accelerate_commit": null,
|
88 |
+
"diffusers_version": "0.27.2",
|
89 |
+
"diffusers_commit": null,
|
90 |
+
"optimum_version": null,
|
91 |
+
"optimum_commit": null,
|
92 |
+
"timm_version": "0.9.16",
|
93 |
+
"timm_commit": null,
|
94 |
+
"peft_version": null,
|
95 |
+
"peft_commit": null
|
96 |
+
}
|
97 |
+
},
|
98 |
+
"report": {
|
99 |
+
"prefill": {
|
100 |
+
"memory": {
|
101 |
+
"unit": "MB",
|
102 |
+
"max_ram": 958.943232,
|
103 |
+
"max_global_vram": null,
|
104 |
+
"max_process_vram": null,
|
105 |
+
"max_reserved": null,
|
106 |
+
"max_allocated": null
|
107 |
+
},
|
108 |
+
"latency": {
|
109 |
+
"unit": "s",
|
110 |
+
"count": 13,
|
111 |
+
"total": 0.6204911969999785,
|
112 |
+
"mean": 0.04773009207692142,
|
113 |
+
"stdev": 0.0009133680248022937,
|
114 |
+
"p50": 0.0475830690000123,
|
115 |
+
"p90": 0.04884250680000264,
|
116 |
+
"p95": 0.04896724660000586,
|
117 |
+
"p99": 0.04901646292000919,
|
118 |
+
"values": [
|
119 |
+
0.048926233000003094,
|
120 |
+
0.04840851700001281,
|
121 |
+
0.047442687000000205,
|
122 |
+
0.048507602000000816,
|
123 |
+
0.04719415299999241,
|
124 |
+
0.046810287999988986,
|
125 |
+
0.0475830690000123,
|
126 |
+
0.047273870999987366,
|
127 |
+
0.048422614999992675,
|
128 |
+
0.04715296599999874,
|
129 |
+
0.04902876700001002,
|
130 |
+
0.048079754999974966,
|
131 |
+
0.045660674000004065
|
132 |
+
]
|
133 |
+
},
|
134 |
+
"throughput": {
|
135 |
+
"unit": "tokens/s",
|
136 |
+
"value": 41.9022866492027
|
137 |
+
},
|
138 |
+
"energy": {
|
139 |
+
"unit": "kWh",
|
140 |
+
"cpu": 1.8244317685714877e-06,
|
141 |
+
"ram": 7.624539440536698e-08,
|
142 |
+
"gpu": 0.0,
|
143 |
+
"total": 1.9006771629768547e-06
|
144 |
+
},
|
145 |
+
"efficiency": {
|
146 |
+
"unit": "tokens/kWh",
|
147 |
+
"value": 1052256.5530632173
|
148 |
+
}
|
149 |
+
},
|
150 |
+
"decode": {
|
151 |
+
"memory": {
|
152 |
+
"unit": "MB",
|
153 |
+
"max_ram": 958.943232,
|
154 |
+
"max_global_vram": null,
|
155 |
+
"max_process_vram": null,
|
156 |
+
"max_reserved": null,
|
157 |
+
"max_allocated": null
|
158 |
+
},
|
159 |
+
"latency": {
|
160 |
+
"unit": "s",
|
161 |
+
"count": 13,
|
162 |
+
"total": 0.3882265669999754,
|
163 |
+
"mean": 0.029863582076921184,
|
164 |
+
"stdev": 0.0007283825675714093,
|
165 |
+
"p50": 0.029798438999989685,
|
166 |
+
"p90": 0.030556406799996695,
|
167 |
+
"p95": 0.031099695599999675,
|
168 |
+
"p99": 0.03164168952000523,
|
169 |
+
"values": [
|
170 |
+
0.029798438999989685,
|
171 |
+
0.029936547000005476,
|
172 |
+
0.02971778899998867,
|
173 |
+
0.03002830799999856,
|
174 |
+
0.029664007999997466,
|
175 |
+
0.02983449600000654,
|
176 |
+
0.0294206250000002,
|
177 |
+
0.02939688099999671,
|
178 |
+
0.030189898000003268,
|
179 |
+
0.029199772999987772,
|
180 |
+
0.03064803399999505,
|
181 |
+
0.03177718800000662,
|
182 |
+
0.02861458099999936
|
183 |
+
]
|
184 |
+
},
|
185 |
+
"throughput": {
|
186 |
+
"unit": "tokens/s",
|
187 |
+
"value": 33.48560120565068
|
188 |
+
},
|
189 |
+
"energy": {
|
190 |
+
"unit": "kWh",
|
191 |
+
"cpu": 1.0716519572518082e-06,
|
192 |
+
"ram": 4.478705194024805e-08,
|
193 |
+
"gpu": 0.0,
|
194 |
+
"total": 1.1164390091920568e-06
|
195 |
+
},
|
196 |
+
"efficiency": {
|
197 |
+
"unit": "tokens/kWh",
|
198 |
+
"value": 895704.9975561842
|
199 |
+
}
|
200 |
+
},
|
201 |
+
"per_token": {
|
202 |
+
"memory": null,
|
203 |
+
"latency": {
|
204 |
+
"unit": "s",
|
205 |
+
"count": 12,
|
206 |
+
"total": 0.9303966850000052,
|
207 |
+
"mean": 0.07753305708333376,
|
208 |
+
"stdev": 0.001510353011715292,
|
209 |
+
"p50": 0.07709178250000548,
|
210 |
+
"p90": 0.07956558189999896,
|
211 |
+
"p95": 0.0797690607999968,
|
212 |
+
"p99": 0.07986273855999429,
|
213 |
+
"values": [
|
214 |
+
0.07855636799999388,
|
215 |
+
0.0771555060000253,
|
216 |
+
0.07857870899999853,
|
217 |
+
0.07688182599997617,
|
218 |
+
0.07662799300001666,
|
219 |
+
0.07702805899998566,
|
220 |
+
0.07669302400000788,
|
221 |
+
0.07859653299999536,
|
222 |
+
0.07640625900000941,
|
223 |
+
0.07967325399999936,
|
224 |
+
0.07988615799999366,
|
225 |
+
0.0743129960000033
|
226 |
+
]
|
227 |
+
},
|
228 |
+
"throughput": {
|
229 |
+
"unit": "tokens/s",
|
230 |
+
"value": 12.897724372265937
|
231 |
+
},
|
232 |
+
"energy": null,
|
233 |
+
"efficiency": null
|
234 |
+
}
|
235 |
+
}
|
236 |
+
}
|