Commit
•
1e3005f
1
Parent(s):
5148d5e
Upload cuda_training_transformers_token-classification_microsoft/deberta-v3-base/benchmark.json with huggingface_hub
Browse files
cuda_training_transformers_token-classification_microsoft/deberta-v3-base/benchmark.json
CHANGED
@@ -3,7 +3,7 @@
|
|
3 |
"name": "cuda_training_transformers_token-classification_microsoft/deberta-v3-base",
|
4 |
"backend": {
|
5 |
"name": "pytorch",
|
6 |
-
"version": "2.
|
7 |
"_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
|
8 |
"task": "token-classification",
|
9 |
"library": "transformers",
|
@@ -110,33 +110,33 @@
|
|
110 |
"overall": {
|
111 |
"memory": {
|
112 |
"unit": "MB",
|
113 |
-
"max_ram":
|
114 |
-
"max_global_vram":
|
115 |
-
"max_process_vram":
|
116 |
"max_reserved": 3919.577088,
|
117 |
"max_allocated": 3695.353344
|
118 |
},
|
119 |
"latency": {
|
120 |
"unit": "s",
|
121 |
"count": 5,
|
122 |
-
"total":
|
123 |
-
"mean": 0.
|
124 |
-
"stdev": 0.
|
125 |
-
"p50": 0.
|
126 |
-
"p90": 0.
|
127 |
-
"p95": 0.
|
128 |
-
"p99": 0.
|
129 |
"values": [
|
130 |
-
0.
|
131 |
-
0.
|
132 |
-
0.
|
133 |
-
0.
|
134 |
-
0.
|
135 |
]
|
136 |
},
|
137 |
"throughput": {
|
138 |
"unit": "samples/s",
|
139 |
-
"value":
|
140 |
},
|
141 |
"energy": null,
|
142 |
"efficiency": null
|
@@ -144,30 +144,30 @@
|
|
144 |
"warmup": {
|
145 |
"memory": {
|
146 |
"unit": "MB",
|
147 |
-
"max_ram":
|
148 |
-
"max_global_vram":
|
149 |
-
"max_process_vram":
|
150 |
"max_reserved": 3919.577088,
|
151 |
"max_allocated": 3695.353344
|
152 |
},
|
153 |
"latency": {
|
154 |
"unit": "s",
|
155 |
"count": 2,
|
156 |
-
"total": 0.
|
157 |
-
"mean": 0.
|
158 |
-
"stdev": 0.
|
159 |
-
"p50": 0.
|
160 |
-
"p90": 0.
|
161 |
-
"p95": 0.
|
162 |
-
"p99": 0.
|
163 |
"values": [
|
164 |
-
0.
|
165 |
-
0.
|
166 |
]
|
167 |
},
|
168 |
"throughput": {
|
169 |
"unit": "samples/s",
|
170 |
-
"value":
|
171 |
},
|
172 |
"energy": null,
|
173 |
"efficiency": null
|
@@ -175,31 +175,31 @@
|
|
175 |
"train": {
|
176 |
"memory": {
|
177 |
"unit": "MB",
|
178 |
-
"max_ram":
|
179 |
-
"max_global_vram":
|
180 |
-
"max_process_vram":
|
181 |
"max_reserved": 3919.577088,
|
182 |
"max_allocated": 3695.353344
|
183 |
},
|
184 |
"latency": {
|
185 |
"unit": "s",
|
186 |
"count": 3,
|
187 |
-
"total": 0.
|
188 |
-
"mean": 0.
|
189 |
-
"stdev":
|
190 |
-
"p50": 0.
|
191 |
-
"p90": 0.
|
192 |
-
"p95": 0.
|
193 |
-
"p99": 0.
|
194 |
"values": [
|
195 |
-
0.
|
196 |
-
0.
|
197 |
-
0.
|
198 |
]
|
199 |
},
|
200 |
"throughput": {
|
201 |
"unit": "samples/s",
|
202 |
-
"value":
|
203 |
},
|
204 |
"energy": null,
|
205 |
"efficiency": null
|
|
|
3 |
"name": "cuda_training_transformers_token-classification_microsoft/deberta-v3-base",
|
4 |
"backend": {
|
5 |
"name": "pytorch",
|
6 |
+
"version": "2.2.2+rocm5.7",
|
7 |
"_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
|
8 |
"task": "token-classification",
|
9 |
"library": "transformers",
|
|
|
110 |
"overall": {
|
111 |
"memory": {
|
112 |
"unit": "MB",
|
113 |
+
"max_ram": 1156.796416,
|
114 |
+
"max_global_vram": 4373.233664,
|
115 |
+
"max_process_vram": 472834.469888,
|
116 |
"max_reserved": 3919.577088,
|
117 |
"max_allocated": 3695.353344
|
118 |
},
|
119 |
"latency": {
|
120 |
"unit": "s",
|
121 |
"count": 5,
|
122 |
+
"total": 0.88155233001709,
|
123 |
+
"mean": 0.176310466003418,
|
124 |
+
"stdev": 0.20769205971333674,
|
125 |
+
"p50": 0.07223963165283204,
|
126 |
+
"p90": 0.384371078491211,
|
127 |
+
"p95": 0.48803227996826165,
|
128 |
+
"p99": 0.5709612411499023,
|
129 |
"values": [
|
130 |
+
0.5916934814453125,
|
131 |
+
0.07338747406005859,
|
132 |
+
0.0721372299194336,
|
133 |
+
0.07209451293945313,
|
134 |
+
0.07223963165283204
|
135 |
]
|
136 |
},
|
137 |
"throughput": {
|
138 |
"unit": "samples/s",
|
139 |
+
"value": 56.718130390547195
|
140 |
},
|
141 |
"energy": null,
|
142 |
"efficiency": null
|
|
|
144 |
"warmup": {
|
145 |
"memory": {
|
146 |
"unit": "MB",
|
147 |
+
"max_ram": 1156.796416,
|
148 |
+
"max_global_vram": 4373.233664,
|
149 |
+
"max_process_vram": 472834.469888,
|
150 |
"max_reserved": 3919.577088,
|
151 |
"max_allocated": 3695.353344
|
152 |
},
|
153 |
"latency": {
|
154 |
"unit": "s",
|
155 |
"count": 2,
|
156 |
+
"total": 0.6650809555053712,
|
157 |
+
"mean": 0.3325404777526856,
|
158 |
+
"stdev": 0.25915300369262695,
|
159 |
+
"p50": 0.3325404777526856,
|
160 |
+
"p90": 0.5398628807067871,
|
161 |
+
"p95": 0.5657781810760498,
|
162 |
+
"p99": 0.58651042137146,
|
163 |
"values": [
|
164 |
+
0.5916934814453125,
|
165 |
+
0.07338747406005859
|
166 |
]
|
167 |
},
|
168 |
"throughput": {
|
169 |
"unit": "samples/s",
|
170 |
+
"value": 12.028610853728456
|
171 |
},
|
172 |
"energy": null,
|
173 |
"efficiency": null
|
|
|
175 |
"train": {
|
176 |
"memory": {
|
177 |
"unit": "MB",
|
178 |
+
"max_ram": 1156.796416,
|
179 |
+
"max_global_vram": 4373.233664,
|
180 |
+
"max_process_vram": 472834.469888,
|
181 |
"max_reserved": 3919.577088,
|
182 |
"max_allocated": 3695.353344
|
183 |
},
|
184 |
"latency": {
|
185 |
"unit": "s",
|
186 |
"count": 3,
|
187 |
+
"total": 0.21647137451171877,
|
188 |
+
"mean": 0.07215712483723959,
|
189 |
+
"stdev": 6.089179506341817e-05,
|
190 |
+
"p50": 0.0721372299194336,
|
191 |
+
"p90": 0.07221915130615235,
|
192 |
+
"p95": 0.07222939147949219,
|
193 |
+
"p99": 0.07223758361816407,
|
194 |
"values": [
|
195 |
+
0.0721372299194336,
|
196 |
+
0.07209451293945313,
|
197 |
+
0.07223963165283204
|
198 |
]
|
199 |
},
|
200 |
"throughput": {
|
201 |
"unit": "samples/s",
|
202 |
+
"value": 83.15187188422256
|
203 |
},
|
204 |
"energy": null,
|
205 |
"efficiency": null
|