IlyasMoutawwakil HF staff commited on
Commit
ec759d1
·
verified ·
1 Parent(s): e66ec5c

Upload cuda_training_transformers_text-classification_FacebookAI/roberta-base/benchmark.json with huggingface_hub

Browse files
cuda_training_transformers_text-classification_FacebookAI/roberta-base/benchmark.json CHANGED
@@ -3,7 +3,7 @@
3
  "name": "cuda_training_transformers_text-classification_FacebookAI/roberta-base",
4
  "backend": {
5
  "name": "pytorch",
6
- "version": "2.4.0+rocm6.1",
7
  "_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
8
  "task": "text-classification",
9
  "library": "transformers",
@@ -71,7 +71,7 @@
71
  "name": "process",
72
  "_target_": "optimum_benchmark.launchers.process.launcher.ProcessLauncher",
73
  "device_isolation": true,
74
- "device_isolation_action": "error",
75
  "numactl": false,
76
  "numactl_kwargs": {},
77
  "start_method": "spawn"
@@ -101,7 +101,7 @@
101
  "optimum_benchmark_commit": null,
102
  "transformers_version": "4.44.2",
103
  "transformers_commit": null,
104
- "accelerate_version": "0.33.0",
105
  "accelerate_commit": null,
106
  "diffusers_version": "0.30.2",
107
  "diffusers_commit": null,
@@ -117,33 +117,33 @@
117
  "overall": {
118
  "memory": {
119
  "unit": "MB",
120
- "max_ram": 1623.77728,
121
- "max_global_vram": 11.48928,
122
- "max_process_vram": 0.0,
123
  "max_reserved": 2707.423232,
124
  "max_allocated": 2497.900032
125
  },
126
  "latency": {
127
  "unit": "s",
128
  "count": 5,
129
- "total": 0.8003283233642579,
130
- "mean": 0.16006566467285158,
131
- "stdev": 0.22566088893550196,
132
- "p50": 0.04699919128417969,
133
- "p90": 0.3860635055541993,
134
- "p95": 0.4987250511169433,
135
- "p99": 0.5888542875671386,
136
  "values": [
137
- 0.6113865966796875,
138
- 0.048078868865966795,
139
- 0.046969593048095705,
140
- 0.04699919128417969,
141
- 0.046894073486328125
142
  ]
143
  },
144
  "throughput": {
145
  "unit": "samples/s",
146
- "value": 62.47436025982454
147
  },
148
  "energy": null,
149
  "efficiency": null
@@ -151,30 +151,30 @@
151
  "warmup": {
152
  "memory": {
153
  "unit": "MB",
154
- "max_ram": 1623.77728,
155
- "max_global_vram": 11.48928,
156
- "max_process_vram": 0.0,
157
  "max_reserved": 2707.423232,
158
  "max_allocated": 2497.900032
159
  },
160
  "latency": {
161
  "unit": "s",
162
  "count": 2,
163
- "total": 0.6594654655456543,
164
- "mean": 0.32973273277282716,
165
- "stdev": 0.28165386390686037,
166
- "p50": 0.32973273277282716,
167
- "p90": 0.5550558238983154,
168
- "p95": 0.5832212102890014,
169
- "p99": 0.6057535194015503,
170
  "values": [
171
- 0.6113865966796875,
172
- 0.048078868865966795
173
  ]
174
  },
175
  "throughput": {
176
  "unit": "samples/s",
177
- "value": 12.131037056475197
178
  },
179
  "energy": null,
180
  "efficiency": null
@@ -182,31 +182,31 @@
182
  "train": {
183
  "memory": {
184
  "unit": "MB",
185
- "max_ram": 1623.77728,
186
- "max_global_vram": 11.48928,
187
- "max_process_vram": 0.0,
188
  "max_reserved": 2707.423232,
189
  "max_allocated": 2497.900032
190
  },
191
  "latency": {
192
  "unit": "s",
193
  "count": 3,
194
- "total": 0.14086285781860353,
195
- "mean": 0.04695428593953451,
196
- "stdev": 4.4258095573343614e-05,
197
- "p50": 0.046969593048095705,
198
- "p90": 0.046993271636962895,
199
- "p95": 0.04699623146057129,
200
- "p99": 0.04699859931945801,
201
  "values": [
202
- 0.046969593048095705,
203
- 0.04699919128417969,
204
- 0.046894073486328125
205
  ]
206
  },
207
  "throughput": {
208
  "unit": "samples/s",
209
- "value": 127.78386211061785
210
  },
211
  "energy": null,
212
  "efficiency": null
 
3
  "name": "cuda_training_transformers_text-classification_FacebookAI/roberta-base",
4
  "backend": {
5
  "name": "pytorch",
6
+ "version": "2.3.1+rocm5.7",
7
  "_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
8
  "task": "text-classification",
9
  "library": "transformers",
 
71
  "name": "process",
72
  "_target_": "optimum_benchmark.launchers.process.launcher.ProcessLauncher",
73
  "device_isolation": true,
74
+ "device_isolation_action": "warn",
75
  "numactl": false,
76
  "numactl_kwargs": {},
77
  "start_method": "spawn"
 
101
  "optimum_benchmark_commit": null,
102
  "transformers_version": "4.44.2",
103
  "transformers_commit": null,
104
+ "accelerate_version": "0.34.0",
105
  "accelerate_commit": null,
106
  "diffusers_version": "0.30.2",
107
  "diffusers_commit": null,
 
117
  "overall": {
118
  "memory": {
119
  "unit": "MB",
120
+ "max_ram": 1271.427072,
121
+ "max_global_vram": 68702.69952,
122
+ "max_process_vram": 248457.261056,
123
  "max_reserved": 2707.423232,
124
  "max_allocated": 2497.900032
125
  },
126
  "latency": {
127
  "unit": "s",
128
  "count": 5,
129
+ "total": 0.7138150825500487,
130
+ "mean": 0.14276301651000972,
131
+ "stdev": 0.20032875869937944,
132
+ "p50": 0.042495990753173826,
133
+ "p90": 0.34335127563476564,
134
+ "p95": 0.4433855987548827,
135
+ "p99": 0.5234130572509765,
136
  "values": [
137
+ 0.543419921875,
138
+ 0.04324830627441406,
139
+ 0.042495990753173826,
140
+ 0.04220063018798828,
141
+ 0.04245023345947266
142
  ]
143
  },
144
  "throughput": {
145
  "unit": "samples/s",
146
+ "value": 70.04615231913972
147
  },
148
  "energy": null,
149
  "efficiency": null
 
151
  "warmup": {
152
  "memory": {
153
  "unit": "MB",
154
+ "max_ram": 1271.427072,
155
+ "max_global_vram": 68702.69952,
156
+ "max_process_vram": 248457.261056,
157
  "max_reserved": 2707.423232,
158
  "max_allocated": 2497.900032
159
  },
160
  "latency": {
161
  "unit": "s",
162
  "count": 2,
163
+ "total": 0.586668228149414,
164
+ "mean": 0.293334114074707,
165
+ "stdev": 0.25008580780029294,
166
+ "p50": 0.293334114074707,
167
+ "p90": 0.4934027603149414,
168
+ "p95": 0.5184113410949707,
169
+ "p99": 0.5384182057189941,
170
  "values": [
171
+ 0.543419921875,
172
+ 0.04324830627441406
173
  ]
174
  },
175
  "throughput": {
176
  "unit": "samples/s",
177
+ "value": 13.636327341664975
178
  },
179
  "energy": null,
180
  "efficiency": null
 
182
  "train": {
183
  "memory": {
184
  "unit": "MB",
185
+ "max_ram": 1271.427072,
186
+ "max_global_vram": 68702.69952,
187
+ "max_process_vram": 248457.261056,
188
  "max_reserved": 2707.423232,
189
  "max_allocated": 2497.900032
190
  },
191
  "latency": {
192
  "unit": "s",
193
  "count": 3,
194
+ "total": 0.12714685440063478,
195
+ "mean": 0.04238228480021159,
196
+ "stdev": 0.00012980043936292865,
197
+ "p50": 0.04245023345947266,
198
+ "p90": 0.042486839294433594,
199
+ "p95": 0.04249141502380371,
200
+ "p99": 0.0424950756072998,
201
  "values": [
202
+ 0.042495990753173826,
203
+ 0.04220063018798828,
204
+ 0.04245023345947266
205
  ]
206
  },
207
  "throughput": {
208
  "unit": "samples/s",
209
+ "value": 141.56858291816408
210
  },
211
  "energy": null,
212
  "efficiency": null