IlyasMoutawwakil HF Staff commited on
Commit
cd8595c
·
verified ·
1 Parent(s): 9c77ab9

Upload cuda_training_transformers_fill-mask_hf-internal-testing/tiny-random-BertModel/benchmark.json with huggingface_hub

Browse files
cuda_training_transformers_fill-mask_hf-internal-testing/tiny-random-BertModel/benchmark.json CHANGED
@@ -3,7 +3,7 @@
3
  "name": "cuda_training_transformers_fill-mask_hf-internal-testing/tiny-random-BertModel",
4
  "backend": {
5
  "name": "pytorch",
6
- "version": "2.3.1+rocm5.7",
7
  "_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
8
  "model": "hf-internal-testing/tiny-random-BertModel",
9
  "processor": "hf-internal-testing/tiny-random-BertModel",
@@ -11,7 +11,7 @@
11
  "library": "transformers",
12
  "model_type": "bert",
13
  "device": "cuda",
14
- "device_ids": "4",
15
  "seed": 42,
16
  "inter_op_num_threads": null,
17
  "intra_op_num_threads": null,
@@ -110,7 +110,7 @@
110
  "optimum_commit": null,
111
  "timm_version": "1.0.15",
112
  "timm_commit": null,
113
- "peft_version": "0.15.2",
114
  "peft_commit": null
115
  },
116
  "print_report": true,
@@ -120,34 +120,34 @@
120
  "overall": {
121
  "memory": {
122
  "unit": "MB",
123
- "max_ram": 1283.076096,
124
  "max_global_vram": 68702.69952,
125
- "max_process_vram": 58069.716992,
126
- "max_reserved": 2.097152,
127
- "max_allocated": 1.940992
128
  },
129
  "latency": {
130
  "unit": "s",
131
  "values": [
132
- 0.61989208984375,
133
- 0.01592492961883545,
134
- 0.014997410774230956,
135
- 0.015156611442565918,
136
- 0.015388931274414062
137
  ],
138
  "count": 5,
139
- "total": 0.6813599729537965,
140
- "mean": 0.13627199459075928,
141
- "p50": 0.015388931274414062,
142
- "p90": 0.3783052257537842,
143
- "p95": 0.49909865779876694,
144
- "p99": 0.5957334034347533,
145
- "stdev": 0.241810251350772,
146
- "stdev_": 177.4467689248671
147
  },
148
  "throughput": {
149
  "unit": "samples/s",
150
- "value": 73.38264938464553
151
  },
152
  "energy": null,
153
  "efficiency": null
@@ -155,31 +155,31 @@
155
  "warmup": {
156
  "memory": {
157
  "unit": "MB",
158
- "max_ram": 1283.076096,
159
  "max_global_vram": 68702.69952,
160
- "max_process_vram": 58069.716992,
161
- "max_reserved": 2.097152,
162
- "max_allocated": 1.940992
163
  },
164
  "latency": {
165
  "unit": "s",
166
  "values": [
167
- 0.61989208984375,
168
- 0.01592492961883545
169
  ],
170
  "count": 2,
171
- "total": 0.6358170194625854,
172
- "mean": 0.3179085097312927,
173
- "p50": 0.3179085097312927,
174
- "p90": 0.5594953738212586,
175
- "p95": 0.5896937318325043,
176
- "p99": 0.6138524182415008,
177
- "stdev": 0.30198358011245724,
178
- "stdev_": 94.99071930087818
179
  },
180
  "throughput": {
181
  "unit": "samples/s",
182
- "value": 12.582236327618089
183
  },
184
  "energy": null,
185
  "efficiency": null
@@ -187,32 +187,32 @@
187
  "train": {
188
  "memory": {
189
  "unit": "MB",
190
- "max_ram": 1283.076096,
191
  "max_global_vram": 68702.69952,
192
- "max_process_vram": 58069.716992,
193
- "max_reserved": 2.097152,
194
- "max_allocated": 1.940992
195
  },
196
  "latency": {
197
  "unit": "s",
198
  "values": [
199
- 0.014997410774230956,
200
- 0.015156611442565918,
201
- 0.015388931274414062
202
  ],
203
  "count": 3,
204
- "total": 0.04554295349121094,
205
- "mean": 0.015180984497070312,
206
- "p50": 0.015156611442565918,
207
- "p90": 0.015342467308044434,
208
- "p95": 0.015365699291229247,
209
- "p99": 0.015384284877777099,
210
- "stdev": 0.00016076402967354155,
211
- "stdev_": 1.0589828986688343
212
  },
213
  "throughput": {
214
  "unit": "samples/s",
215
- "value": 395.2312843187413
216
  },
217
  "energy": null,
218
  "efficiency": null
 
3
  "name": "cuda_training_transformers_fill-mask_hf-internal-testing/tiny-random-BertModel",
4
  "backend": {
5
  "name": "pytorch",
6
+ "version": "2.7.0+rocm6.3",
7
  "_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
8
  "model": "hf-internal-testing/tiny-random-BertModel",
9
  "processor": "hf-internal-testing/tiny-random-BertModel",
 
11
  "library": "transformers",
12
  "model_type": "bert",
13
  "device": "cuda",
14
+ "device_ids": "5",
15
  "seed": 42,
16
  "inter_op_num_threads": null,
17
  "intra_op_num_threads": null,
 
110
  "optimum_commit": null,
111
  "timm_version": "1.0.15",
112
  "timm_commit": null,
113
+ "peft_version": null,
114
  "peft_commit": null
115
  },
116
  "print_report": true,
 
120
  "overall": {
121
  "memory": {
122
  "unit": "MB",
123
+ "max_ram": 2315.620352,
124
  "max_global_vram": 68702.69952,
125
+ "max_process_vram": 0.0,
126
+ "max_reserved": 81.788928,
127
+ "max_allocated": 81.408512
128
  },
129
  "latency": {
130
  "unit": "s",
131
  "values": [
132
+ 1.025809814453125,
133
+ 0.01719100570678711,
134
+ 0.016827968597412108,
135
+ 0.016571327209472655,
136
+ 0.01682508850097656
137
  ],
138
  "count": 5,
139
+ "total": 1.0932252044677737,
140
+ "mean": 0.21864504089355474,
141
+ "p50": 0.016827968597412108,
142
+ "p90": 0.62236229095459,
143
+ "p95": 0.8240860527038572,
144
+ "p99": 0.9854650621032713,
145
+ "stdev": 0.40358243509461833,
146
+ "stdev_": 184.58339299408058
147
  },
148
  "throughput": {
149
  "unit": "samples/s",
150
+ "value": 45.73623055493129
151
  },
152
  "energy": null,
153
  "efficiency": null
 
155
  "warmup": {
156
  "memory": {
157
  "unit": "MB",
158
+ "max_ram": 2315.620352,
159
  "max_global_vram": 68702.69952,
160
+ "max_process_vram": 0.0,
161
+ "max_reserved": 81.788928,
162
+ "max_allocated": 81.408512
163
  },
164
  "latency": {
165
  "unit": "s",
166
  "values": [
167
+ 1.025809814453125,
168
+ 0.01719100570678711
169
  ],
170
  "count": 2,
171
+ "total": 1.0430008201599121,
172
+ "mean": 0.5215004100799561,
173
+ "p50": 0.5215004100799561,
174
+ "p90": 0.9249479335784911,
175
+ "p95": 0.975378874015808,
176
+ "p99": 1.0157236263656615,
177
+ "stdev": 0.504309404373169,
178
+ "stdev_": 96.7035489571041
179
  },
180
  "throughput": {
181
  "unit": "samples/s",
182
+ "value": 7.670176135406534
183
  },
184
  "energy": null,
185
  "efficiency": null
 
187
  "train": {
188
  "memory": {
189
  "unit": "MB",
190
+ "max_ram": 2315.620352,
191
  "max_global_vram": 68702.69952,
192
+ "max_process_vram": 0.0,
193
+ "max_reserved": 81.788928,
194
+ "max_allocated": 81.408512
195
  },
196
  "latency": {
197
  "unit": "s",
198
  "values": [
199
+ 0.016827968597412108,
200
+ 0.016571327209472655,
201
+ 0.01682508850097656
202
  ],
203
  "count": 3,
204
+ "total": 0.050224384307861324,
205
+ "mean": 0.016741461435953776,
206
+ "p50": 0.01682508850097656,
207
+ "p90": 0.016827392578125,
208
+ "p95": 0.016827680587768554,
209
+ "p99": 0.0168279109954834,
210
+ "stdev": 0.0001203088109938094,
211
+ "stdev_": 0.718627889530812
212
  },
213
  "throughput": {
214
  "unit": "samples/s",
215
+ "value": 358.3916507500634
216
  },
217
  "energy": null,
218
  "efficiency": null