IlyasMoutawwakil HF Staff commited on
Commit
7f98754
·
verified ·
1 Parent(s): 059822f

Upload cuda_training_transformers_text-generation_hf-internal-testing/tiny-random-LlamaForCausalLM/benchmark.json with huggingface_hub

Browse files
cuda_training_transformers_text-generation_hf-internal-testing/tiny-random-LlamaForCausalLM/benchmark.json CHANGED
@@ -3,7 +3,7 @@
3
  "name": "cuda_training_transformers_text-generation_hf-internal-testing/tiny-random-LlamaForCausalLM",
4
  "backend": {
5
  "name": "pytorch",
6
- "version": "2.5.1+cu124",
7
  "_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
8
  "model": "hf-internal-testing/tiny-random-LlamaForCausalLM",
9
  "processor": "hf-internal-testing/tiny-random-LlamaForCausalLM",
@@ -52,7 +52,6 @@
52
  "per_device_train_batch_size": 2,
53
  "gradient_accumulation_steps": 1,
54
  "output_dir": "./trainer_output",
55
- "evaluation_strategy": "no",
56
  "eval_strategy": "no",
57
  "save_strategy": "no",
58
  "do_train": true,
@@ -80,10 +79,10 @@
80
  "environment": {
81
  "cpu": " AMD EPYC 7R32",
82
  "cpu_count": 16,
83
- "cpu_ram_mb": 66697.248768,
84
  "system": "Linux",
85
  "machine": "x86_64",
86
- "platform": "Linux-5.10.234-225.921.amzn2.x86_64-x86_64-with-glibc2.35",
87
  "processor": "x86_64",
88
  "python_version": "3.10.12",
89
  "gpu": [
@@ -93,9 +92,9 @@
93
  "gpu_vram_mb": 24146608128,
94
  "optimum_benchmark_version": "0.6.0.dev0",
95
  "optimum_benchmark_commit": null,
96
- "transformers_version": "4.47.1",
97
  "transformers_commit": null,
98
- "accelerate_version": "1.3.0",
99
  "accelerate_commit": null,
100
  "diffusers_version": "0.33.1",
101
  "diffusers_commit": null,
@@ -103,7 +102,7 @@
103
  "optimum_commit": null,
104
  "timm_version": "1.0.15",
105
  "timm_commit": null,
106
- "peft_version": "0.14.0",
107
  "peft_commit": null
108
  },
109
  "print_report": true,
@@ -113,75 +112,75 @@
113
  "overall": {
114
  "memory": {
115
  "unit": "MB",
116
- "max_ram": 1401.5488,
117
- "max_global_vram": 699.92448,
118
  "max_process_vram": 0.0,
119
- "max_reserved": 44.040192,
120
- "max_allocated": 42.173952
121
  },
122
  "latency": {
123
  "unit": "s",
124
  "values": [
125
- 0.2644462890625,
126
- 0.0090415678024292,
127
- 0.008861120223999024,
128
- 0.00874233627319336,
129
- 0.008732831954956055
130
  ],
131
  "count": 5,
132
- "total": 0.2998241453170777,
133
- "mean": 0.059964829063415534,
134
- "p50": 0.008861120223999024,
135
- "p90": 0.1622844005584717,
136
- "p95": 0.2133653448104858,
137
- "p99": 0.2542301002120972,
138
- "stdev": 0.10224079065942444,
139
- "stdev_": 170.50126258393925
140
  },
141
  "throughput": {
142
  "unit": "samples/s",
143
- "value": 166.76442101460083
144
  },
145
  "energy": {
146
  "unit": "kWh",
147
- "cpu": 6.209379211806029e-06,
148
- "ram": 2.920926838888881e-06,
149
- "gpu": 8.258062162000077e-06,
150
- "total": 1.7388368212694988e-05
151
  },
152
  "efficiency": {
153
  "unit": "samples/kWh",
154
- "value": 575097.0923596586
155
  }
156
  },
157
  "warmup": {
158
  "memory": {
159
  "unit": "MB",
160
- "max_ram": 1401.5488,
161
- "max_global_vram": 699.92448,
162
  "max_process_vram": 0.0,
163
- "max_reserved": 44.040192,
164
- "max_allocated": 42.173952
165
  },
166
  "latency": {
167
  "unit": "s",
168
  "values": [
169
- 0.2644462890625,
170
- 0.0090415678024292
171
  ],
172
  "count": 2,
173
- "total": 0.2734878568649292,
174
- "mean": 0.1367439284324646,
175
- "p50": 0.1367439284324646,
176
- "p90": 0.23890581693649293,
177
- "p95": 0.25167605299949647,
178
- "p99": 0.2618922418498993,
179
- "stdev": 0.1277023606300354,
180
- "stdev_": 93.38795666756445
181
  },
182
  "throughput": {
183
  "unit": "samples/s",
184
- "value": 29.25175578801313
185
  },
186
  "energy": null,
187
  "efficiency": null
@@ -189,32 +188,32 @@
189
  "train": {
190
  "memory": {
191
  "unit": "MB",
192
- "max_ram": 1401.5488,
193
- "max_global_vram": 699.92448,
194
  "max_process_vram": 0.0,
195
- "max_reserved": 44.040192,
196
- "max_allocated": 42.173952
197
  },
198
  "latency": {
199
  "unit": "s",
200
  "values": [
201
- 0.008861120223999024,
202
- 0.00874233627319336,
203
- 0.008732831954956055
204
  ],
205
  "count": 3,
206
- "total": 0.02633628845214844,
207
- "mean": 0.008778762817382813,
208
- "p50": 0.00874233627319336,
209
- "p90": 0.008837363433837892,
210
- "p95": 0.008849241828918457,
211
- "p99": 0.008858744544982911,
212
- "stdev": 5.836460019966407e-05,
213
- "stdev_": 0.6648385588467708
214
  },
215
  "throughput": {
216
  "unit": "samples/s",
217
- "value": 683.4676052665884
218
  },
219
  "energy": null,
220
  "efficiency": null
 
3
  "name": "cuda_training_transformers_text-generation_hf-internal-testing/tiny-random-LlamaForCausalLM",
4
  "backend": {
5
  "name": "pytorch",
6
+ "version": "2.7.0+cu128",
7
  "_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
8
  "model": "hf-internal-testing/tiny-random-LlamaForCausalLM",
9
  "processor": "hf-internal-testing/tiny-random-LlamaForCausalLM",
 
52
  "per_device_train_batch_size": 2,
53
  "gradient_accumulation_steps": 1,
54
  "output_dir": "./trainer_output",
 
55
  "eval_strategy": "no",
56
  "save_strategy": "no",
57
  "do_train": true,
 
79
  "environment": {
80
  "cpu": " AMD EPYC 7R32",
81
  "cpu_count": 16,
82
+ "cpu_ram_mb": 66697.252864,
83
  "system": "Linux",
84
  "machine": "x86_64",
85
+ "platform": "Linux-5.10.236-227.928.amzn2.x86_64-x86_64-with-glibc2.35",
86
  "processor": "x86_64",
87
  "python_version": "3.10.12",
88
  "gpu": [
 
92
  "gpu_vram_mb": 24146608128,
93
  "optimum_benchmark_version": "0.6.0.dev0",
94
  "optimum_benchmark_commit": null,
95
+ "transformers_version": "4.51.3",
96
  "transformers_commit": null,
97
+ "accelerate_version": "1.6.0",
98
  "accelerate_commit": null,
99
  "diffusers_version": "0.33.1",
100
  "diffusers_commit": null,
 
102
  "optimum_commit": null,
103
  "timm_version": "1.0.15",
104
  "timm_commit": null,
105
+ "peft_version": null,
106
  "peft_commit": null
107
  },
108
  "print_report": true,
 
112
  "overall": {
113
  "memory": {
114
  "unit": "MB",
115
+ "max_ram": 1647.321088,
116
+ "max_global_vram": 720.896,
117
  "max_process_vram": 0.0,
118
+ "max_reserved": 65.011712,
119
+ "max_allocated": 42.172928
120
  },
121
  "latency": {
122
  "unit": "s",
123
  "values": [
124
+ 0.3522263793945313,
125
+ 0.009275391578674316,
126
+ 0.00832921600341797,
127
+ 0.008478719711303711,
128
+ 0.008241312026977539
129
  ],
130
  "count": 5,
131
+ "total": 0.38655101871490477,
132
+ "mean": 0.07731020374298095,
133
+ "p50": 0.008478719711303711,
134
+ "p90": 0.21504598426818852,
135
+ "p95": 0.2836361818313598,
136
+ "p99": 0.338508339881897,
137
+ "stdev": 0.13745857628210206,
138
+ "stdev_": 177.80133750401873
139
  },
140
  "throughput": {
141
  "unit": "samples/s",
142
+ "value": 129.34903176875804
143
  },
144
  "energy": {
145
  "unit": "kWh",
146
+ "cpu": 7.863714000694068e-06,
147
+ "ram": 3.6993284888887327e-06,
148
+ "gpu": 1.1788620542000153e-05,
149
+ "total": 2.3351663031582952e-05
150
  },
151
  "efficiency": {
152
  "unit": "samples/kWh",
153
+ "value": 428235.02490914986
154
  }
155
  },
156
  "warmup": {
157
  "memory": {
158
  "unit": "MB",
159
+ "max_ram": 1647.321088,
160
+ "max_global_vram": 720.896,
161
  "max_process_vram": 0.0,
162
+ "max_reserved": 65.011712,
163
+ "max_allocated": 42.172928
164
  },
165
  "latency": {
166
  "unit": "s",
167
  "values": [
168
+ 0.3522263793945313,
169
+ 0.009275391578674316
170
  ],
171
  "count": 2,
172
+ "total": 0.36150177097320557,
173
+ "mean": 0.18075088548660279,
174
+ "p50": 0.18075088548660279,
175
+ "p90": 0.31793128061294557,
176
+ "p95": 0.3350788300037384,
177
+ "p99": 0.3487968695163727,
178
+ "stdev": 0.1714754939079285,
179
+ "stdev_": 94.8684115412747
180
  },
181
  "throughput": {
182
  "unit": "samples/s",
183
+ "value": 22.12990541778828
184
  },
185
  "energy": null,
186
  "efficiency": null
 
188
  "train": {
189
  "memory": {
190
  "unit": "MB",
191
+ "max_ram": 1647.321088,
192
+ "max_global_vram": 720.896,
193
  "max_process_vram": 0.0,
194
+ "max_reserved": 65.011712,
195
+ "max_allocated": 42.172928
196
  },
197
  "latency": {
198
  "unit": "s",
199
  "values": [
200
+ 0.00832921600341797,
201
+ 0.008478719711303711,
202
+ 0.008241312026977539
203
  ],
204
  "count": 3,
205
+ "total": 0.02504924774169922,
206
+ "mean": 0.008349749247233073,
207
+ "p50": 0.00832921600341797,
208
+ "p90": 0.008448818969726563,
209
+ "p95": 0.008463769340515136,
210
+ "p99": 0.008475729637145997,
211
+ "stdev": 9.800276431630811e-05,
212
+ "stdev_": 1.1737210473570103
213
  },
214
  "throughput": {
215
  "unit": "samples/s",
216
+ "value": 718.584453537724
217
  },
218
  "energy": null,
219
  "efficiency": null