hf-transformers-bot commited on
Commit
3badfd3
1 Parent(s): 678c93b

Upload folder using huggingface_hub

Browse files
Files changed (32) hide show
  1. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=False/.hydra/config.yaml +91 -0
  2. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=False/.hydra/hydra.yaml +141 -0
  3. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=False/.hydra/overrides.yaml +3 -0
  4. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=False/benchmark.json +443 -0
  5. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=False/benchmark.log +0 -0
  6. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=False/benchmark_config.json +102 -0
  7. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=False/benchmark_report.json +339 -0
  8. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=False/summary.json +13 -0
  9. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=True/.hydra/config.yaml +91 -0
  10. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=True/.hydra/hydra.yaml +141 -0
  11. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=True/.hydra/overrides.yaml +3 -0
  12. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=True/benchmark.log +1 -0
  13. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=True/benchmark_config.json +102 -0
  14. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=False/.hydra/config.yaml +91 -0
  15. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=False/.hydra/hydra.yaml +141 -0
  16. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=False/.hydra/overrides.yaml +3 -0
  17. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=False/benchmark.json +443 -0
  18. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=False/benchmark.log +0 -0
  19. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=False/benchmark_config.json +102 -0
  20. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=False/benchmark_report.json +339 -0
  21. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=False/summary.json +13 -0
  22. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=True/.hydra/config.yaml +91 -0
  23. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=True/.hydra/hydra.yaml +141 -0
  24. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=True/.hydra/overrides.yaml +3 -0
  25. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=True/benchmark.json +443 -0
  26. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=True/benchmark.log +0 -0
  27. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=True/benchmark_config.json +102 -0
  28. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=True/benchmark_report.json +339 -0
  29. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=True/summary.json +13 -0
  30. 2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/multirun.yaml +232 -0
  31. 2024-06-05/summaries.json +21 -21
  32. 2024-06-05/summary.json +21 -21
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=False/.hydra/config.yaml ADDED
@@ -0,0 +1,91 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ name: pytorch_generate
2
+ backend:
3
+ name: pytorch
4
+ version: 2.3.0+cu121
5
+ _target_: optimum_benchmark.backends.pytorch.backend.PyTorchBackend
6
+ task: null
7
+ library: null
8
+ model: google/gemma-2b
9
+ processor: null
10
+ device: cuda
11
+ device_ids: '0'
12
+ seed: 42
13
+ inter_op_num_threads: null
14
+ intra_op_num_threads: null
15
+ model_kwargs: {}
16
+ processor_kwargs: {}
17
+ hub_kwargs: {}
18
+ no_weights: true
19
+ device_map: null
20
+ torch_dtype: float16
21
+ eval_mode: true
22
+ to_bettertransformer: false
23
+ low_cpu_mem_usage: null
24
+ attn_implementation: null
25
+ cache_implementation: null
26
+ autocast_enabled: false
27
+ autocast_dtype: null
28
+ torch_compile: false
29
+ torch_compile_target: forward
30
+ torch_compile_config:
31
+ backend: inductor
32
+ mode: reduce-overhead
33
+ fullgraph: true
34
+ quantization_scheme: null
35
+ quantization_config: {}
36
+ deepspeed_inference: false
37
+ deepspeed_inference_config: {}
38
+ peft_type: null
39
+ peft_config: {}
40
+ scenario:
41
+ name: inference
42
+ _target_: optimum_benchmark.scenarios.inference.scenario.InferenceScenario
43
+ iterations: 2
44
+ duration: 0
45
+ warmup_runs: 10
46
+ input_shapes:
47
+ batch_size: 1
48
+ sequence_length: 7
49
+ new_tokens: null
50
+ latency: true
51
+ memory: true
52
+ energy: false
53
+ forward_kwargs: {}
54
+ generate_kwargs:
55
+ max_new_tokens: 128
56
+ min_new_tokens: 128
57
+ do_sample: false
58
+ call_kwargs: {}
59
+ launcher:
60
+ name: process
61
+ _target_: optimum_benchmark.launchers.process.launcher.ProcessLauncher
62
+ device_isolation: true
63
+ device_isolation_action: warn
64
+ start_method: spawn
65
+ environment:
66
+ cpu: ' AMD EPYC 7R32'
67
+ cpu_count: 16
68
+ cpu_ram_mb: 66697.29792
69
+ system: Linux
70
+ machine: x86_64
71
+ platform: Linux-5.10.217-205.860.amzn2.x86_64-x86_64-with-glibc2.29
72
+ processor: x86_64
73
+ python_version: 3.8.10
74
+ gpu:
75
+ - NVIDIA A10G
76
+ gpu_count: 1
77
+ gpu_vram_mb: 24146608128
78
+ optimum_benchmark_version: 0.2.1
79
+ optimum_benchmark_commit: null
80
+ transformers_version: 4.42.0.dev0
81
+ transformers_commit: bd5091df8db7cea1a9f94f797fc11487f840ade1
82
+ accelerate_version: 0.31.0.dev0
83
+ accelerate_commit: null
84
+ diffusers_version: null
85
+ diffusers_commit: null
86
+ optimum_version: 1.21.0.dev0
87
+ optimum_commit: null
88
+ timm_version: 0.9.16
89
+ timm_commit: null
90
+ peft_version: 0.11.2.dev0
91
+ peft_commit: null
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=False/.hydra/hydra.yaml ADDED
@@ -0,0 +1,141 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ hydra:
2
+ run:
3
+ dir: _benchmark/tmpjivn_1kh/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/${hydra.job.override_dirname}
4
+ sweep:
5
+ dir: _benchmark/tmpjivn_1kh/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1
6
+ subdir: ${hydra.job.override_dirname}
7
+ launcher:
8
+ _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
9
+ sweeper:
10
+ _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
11
+ max_batch_size: null
12
+ params: null
13
+ help:
14
+ app_name: ${hydra.job.name}
15
+ header: '${hydra.help.app_name} is powered by Hydra.
16
+
17
+ '
18
+ footer: 'Powered by Hydra (https://hydra.cc)
19
+
20
+ Use --hydra-help to view Hydra specific help
21
+
22
+ '
23
+ template: '${hydra.help.header}
24
+
25
+ == Configuration groups ==
26
+
27
+ Compose your configuration from those groups (group=option)
28
+
29
+
30
+ $APP_CONFIG_GROUPS
31
+
32
+
33
+ == Config ==
34
+
35
+ Override anything in the config (foo.bar=value)
36
+
37
+
38
+ $CONFIG
39
+
40
+
41
+ ${hydra.help.footer}
42
+
43
+ '
44
+ hydra_help:
45
+ template: 'Hydra (${hydra.runtime.version})
46
+
47
+ See https://hydra.cc for more info.
48
+
49
+
50
+ == Flags ==
51
+
52
+ $FLAGS_HELP
53
+
54
+
55
+ == Configuration groups ==
56
+
57
+ Compose your configuration from those groups (For example, append hydra/job_logging=disabled
58
+ to command line)
59
+
60
+
61
+ $HYDRA_CONFIG_GROUPS
62
+
63
+
64
+ Use ''--cfg hydra'' to Show the Hydra config.
65
+
66
+ '
67
+ hydra_help: ???
68
+ hydra_logging:
69
+ version: 1
70
+ root:
71
+ level: ERROR
72
+ disable_existing_loggers: true
73
+ job_logging:
74
+ version: 1
75
+ root:
76
+ level: ERROR
77
+ disable_existing_loggers: true
78
+ env: {}
79
+ mode: MULTIRUN
80
+ searchpath: []
81
+ callbacks: {}
82
+ output_subdir: .hydra
83
+ overrides:
84
+ hydra:
85
+ - hydra/job_logging=disabled
86
+ - hydra/hydra_logging=disabled
87
+ - hydra.sweep.dir=_benchmark/tmpjivn_1kh/commit\=bd5091df8db7cea1a9f94f797fc11487f840ade1
88
+ - hydra.run.dir=_benchmark/tmpjivn_1kh/commit\=bd5091df8db7cea1a9f94f797fc11487f840ade1/${hydra.job.override_dirname}
89
+ - hydra.mode=MULTIRUN
90
+ task:
91
+ - backend.model=google/gemma-2b
92
+ - backend.cache_implementation=null
93
+ - backend.torch_compile=False
94
+ job:
95
+ name: cli
96
+ chdir: true
97
+ override_dirname: backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=False
98
+ id: '0'
99
+ num: 0
100
+ config_name: generation
101
+ env_set:
102
+ OVERRIDE_BENCHMARKS: '1'
103
+ LOG_LEVEL: WARN
104
+ env_copy: []
105
+ config:
106
+ override_dirname:
107
+ kv_sep: '='
108
+ item_sep: ','
109
+ exclude_keys: []
110
+ runtime:
111
+ version: 1.3.2
112
+ version_base: '1.3'
113
+ cwd: /transformers
114
+ config_sources:
115
+ - path: hydra.conf
116
+ schema: pkg
117
+ provider: hydra
118
+ - path: optimum_benchmark
119
+ schema: pkg
120
+ provider: main
121
+ - path: /transformers/benchmark/config
122
+ schema: file
123
+ provider: command-line
124
+ - path: ''
125
+ schema: structured
126
+ provider: schema
127
+ output_dir: /transformers/_benchmark/tmpjivn_1kh/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=False
128
+ choices:
129
+ backend: pytorch
130
+ launcher: process
131
+ scenario: inference
132
+ hydra/env: default
133
+ hydra/callbacks: null
134
+ hydra/job_logging: disabled
135
+ hydra/hydra_logging: disabled
136
+ hydra/hydra_help: default
137
+ hydra/help: default
138
+ hydra/sweeper: basic
139
+ hydra/launcher: basic
140
+ hydra/output: default
141
+ verbose: false
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=False/.hydra/overrides.yaml ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ - backend.model=google/gemma-2b
2
+ - backend.cache_implementation=null
3
+ - backend.torch_compile=False
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=False/benchmark.json ADDED
@@ -0,0 +1,443 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "config": {
3
+ "name": "pytorch_generate",
4
+ "backend": {
5
+ "name": "pytorch",
6
+ "version": "2.3.0+cu121",
7
+ "_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
8
+ "task": "text-generation",
9
+ "library": "transformers",
10
+ "model": "google/gemma-2b",
11
+ "processor": "google/gemma-2b",
12
+ "device": "cuda",
13
+ "device_ids": "0",
14
+ "seed": 42,
15
+ "inter_op_num_threads": null,
16
+ "intra_op_num_threads": null,
17
+ "model_kwargs": {},
18
+ "processor_kwargs": {},
19
+ "hub_kwargs": {},
20
+ "no_weights": true,
21
+ "device_map": null,
22
+ "torch_dtype": "float16",
23
+ "eval_mode": true,
24
+ "to_bettertransformer": false,
25
+ "low_cpu_mem_usage": null,
26
+ "attn_implementation": null,
27
+ "cache_implementation": null,
28
+ "autocast_enabled": false,
29
+ "autocast_dtype": null,
30
+ "torch_compile": false,
31
+ "torch_compile_target": "forward",
32
+ "torch_compile_config": {
33
+ "backend": "inductor",
34
+ "mode": "reduce-overhead",
35
+ "fullgraph": true
36
+ },
37
+ "quantization_scheme": null,
38
+ "quantization_config": {},
39
+ "deepspeed_inference": false,
40
+ "deepspeed_inference_config": {},
41
+ "peft_type": null,
42
+ "peft_config": {}
43
+ },
44
+ "scenario": {
45
+ "name": "inference",
46
+ "_target_": "optimum_benchmark.scenarios.inference.scenario.InferenceScenario",
47
+ "iterations": 2,
48
+ "duration": 0,
49
+ "warmup_runs": 10,
50
+ "input_shapes": {
51
+ "batch_size": 1,
52
+ "num_choices": 2,
53
+ "sequence_length": 7
54
+ },
55
+ "new_tokens": null,
56
+ "latency": true,
57
+ "memory": true,
58
+ "energy": false,
59
+ "forward_kwargs": {},
60
+ "generate_kwargs": {
61
+ "max_new_tokens": 128,
62
+ "min_new_tokens": 128,
63
+ "do_sample": false
64
+ },
65
+ "call_kwargs": {}
66
+ },
67
+ "launcher": {
68
+ "name": "process",
69
+ "_target_": "optimum_benchmark.launchers.process.launcher.ProcessLauncher",
70
+ "device_isolation": true,
71
+ "device_isolation_action": "warn",
72
+ "start_method": "spawn"
73
+ },
74
+ "environment": {
75
+ "cpu": " AMD EPYC 7R32",
76
+ "cpu_count": 16,
77
+ "cpu_ram_mb": 66697.29792,
78
+ "system": "Linux",
79
+ "machine": "x86_64",
80
+ "platform": "Linux-5.10.217-205.860.amzn2.x86_64-x86_64-with-glibc2.29",
81
+ "processor": "x86_64",
82
+ "python_version": "3.8.10",
83
+ "gpu": [
84
+ "NVIDIA A10G"
85
+ ],
86
+ "gpu_count": 1,
87
+ "gpu_vram_mb": 24146608128,
88
+ "optimum_benchmark_version": "0.2.1",
89
+ "optimum_benchmark_commit": null,
90
+ "transformers_version": "4.42.0.dev0",
91
+ "transformers_commit": "bd5091df8db7cea1a9f94f797fc11487f840ade1",
92
+ "accelerate_version": "0.31.0.dev0",
93
+ "accelerate_commit": null,
94
+ "diffusers_version": null,
95
+ "diffusers_commit": null,
96
+ "optimum_version": "1.21.0.dev0",
97
+ "optimum_commit": null,
98
+ "timm_version": "0.9.16",
99
+ "timm_commit": null,
100
+ "peft_version": "0.11.2.dev0",
101
+ "peft_commit": null
102
+ }
103
+ },
104
+ "report": {
105
+ "prefill": {
106
+ "memory": {
107
+ "unit": "MB",
108
+ "max_ram": 1704.648704,
109
+ "max_global_vram": 6793.199616,
110
+ "max_process_vram": 0.0,
111
+ "max_reserved": 6146.752512,
112
+ "max_allocated": 5034.576896
113
+ },
114
+ "latency": {
115
+ "unit": "s",
116
+ "count": 2,
117
+ "total": 0.040267551422119144,
118
+ "mean": 0.020133775711059572,
119
+ "stdev": 0.0008556003570556626,
120
+ "p50": 0.020133775711059572,
121
+ "p90": 0.020818255996704102,
122
+ "p95": 0.020903816032409667,
123
+ "p99": 0.02097226406097412,
124
+ "values": [
125
+ 0.020989376068115233,
126
+ 0.019278175354003908
127
+ ]
128
+ },
129
+ "throughput": {
130
+ "unit": "tokens/s",
131
+ "value": 347.67447996129556
132
+ },
133
+ "energy": null,
134
+ "efficiency": null
135
+ },
136
+ "decode": {
137
+ "memory": {
138
+ "unit": "MB",
139
+ "max_ram": 1704.726528,
140
+ "max_global_vram": 6795.296768,
141
+ "max_process_vram": 0.0,
142
+ "max_reserved": 6148.849664,
143
+ "max_allocated": 5034.577408
144
+ },
145
+ "latency": {
146
+ "unit": "s",
147
+ "count": 2,
148
+ "total": 4.883418212890625,
149
+ "mean": 2.4417091064453125,
150
+ "stdev": 0.05995861816406256,
151
+ "p50": 2.4417091064453125,
152
+ "p90": 2.4896760009765626,
153
+ "p95": 2.495671862792969,
154
+ "p99": 2.500468552246094,
155
+ "values": [
156
+ 2.501667724609375,
157
+ 2.38175048828125
158
+ ]
159
+ },
160
+ "throughput": {
161
+ "unit": "tokens/s",
162
+ "value": 52.0127478186331
163
+ },
164
+ "energy": null,
165
+ "efficiency": null
166
+ },
167
+ "per_token": {
168
+ "memory": null,
169
+ "latency": {
170
+ "unit": "s",
171
+ "count": 253,
172
+ "total": 4.88245658493042,
173
+ "mean": 0.019298247371266484,
174
+ "stdev": 0.001368086083430472,
175
+ "p50": 0.018860031127929687,
176
+ "p90": 0.02008965148925781,
177
+ "p95": 0.020330905532836913,
178
+ "p99": 0.020966685943603514,
179
+ "values": [
180
+ 0.020875263214111327,
181
+ 0.020847616195678712,
182
+ 0.020368383407592772,
183
+ 0.020330495834350586,
184
+ 0.02045849609375,
185
+ 0.020093952178955078,
186
+ 0.020340736389160157,
187
+ 0.020060159683227538,
188
+ 0.02003763198852539,
189
+ 0.020312063217163084,
190
+ 0.019573759078979493,
191
+ 0.019325952529907226,
192
+ 0.018966527938842775,
193
+ 0.019517440795898438,
194
+ 0.019360767364501954,
195
+ 0.020485120773315428,
196
+ 0.02046771240234375,
197
+ 0.020135936737060548,
198
+ 0.020221952438354493,
199
+ 0.02389708709716797,
200
+ 0.02106572723388672,
201
+ 0.02044825553894043,
202
+ 0.02006220817565918,
203
+ 0.020006912231445313,
204
+ 0.020107263565063475,
205
+ 0.019998720169067383,
206
+ 0.020331520080566406,
207
+ 0.019545087814331053,
208
+ 0.01947340774536133,
209
+ 0.01924505615234375,
210
+ 0.01922969627380371,
211
+ 0.01940787124633789,
212
+ 0.02029363250732422,
213
+ 0.020384767532348632,
214
+ 0.020206592559814454,
215
+ 0.020295679092407228,
216
+ 0.019958784103393554,
217
+ 0.019999744415283204,
218
+ 0.020195327758789062,
219
+ 0.02004582405090332,
220
+ 0.020176895141601564,
221
+ 0.0200130558013916,
222
+ 0.02007244873046875,
223
+ 0.019893247604370116,
224
+ 0.019912704467773438,
225
+ 0.020329471588134765,
226
+ 0.01985536003112793,
227
+ 0.01983795166015625,
228
+ 0.019800064086914062,
229
+ 0.019962879180908204,
230
+ 0.02010316848754883,
231
+ 0.019629056930541993,
232
+ 0.019495935440063478,
233
+ 0.019523584365844726,
234
+ 0.01965260887145996,
235
+ 0.018594816207885743,
236
+ 0.018861055374145508,
237
+ 0.01882931137084961,
238
+ 0.019094528198242186,
239
+ 0.019744768142700195,
240
+ 0.019742719650268553,
241
+ 0.019775487899780272,
242
+ 0.019877887725830077,
243
+ 0.01971609687805176,
244
+ 0.019719167709350584,
245
+ 0.019715072631835938,
246
+ 0.019720191955566405,
247
+ 0.019887104034423828,
248
+ 0.019720191955566405,
249
+ 0.019687423706054686,
250
+ 0.01886720085144043,
251
+ 0.018889728546142577,
252
+ 0.018729984283447267,
253
+ 0.018740224838256835,
254
+ 0.018711551666259766,
255
+ 0.018711551666259766,
256
+ 0.019500032424926757,
257
+ 0.01990656089782715,
258
+ 0.019615743637084963,
259
+ 0.0196945915222168,
260
+ 0.019612672805786133,
261
+ 0.019586048126220702,
262
+ 0.01982361602783203,
263
+ 0.01966796875,
264
+ 0.01966080093383789,
265
+ 0.019713024139404296,
266
+ 0.019801088333129883,
267
+ 0.01962495994567871,
268
+ 0.01965670394897461,
269
+ 0.01967103958129883,
270
+ 0.0196495361328125,
271
+ 0.019693567276000978,
272
+ 0.019680255889892577,
273
+ 0.01966694450378418,
274
+ 0.019987455368041994,
275
+ 0.019655679702758787,
276
+ 0.018702335357666015,
277
+ 0.018683904647827147,
278
+ 0.018675712585449217,
279
+ 0.0187064323425293,
280
+ 0.01882521629333496,
281
+ 0.018860031127929687,
282
+ 0.018714624404907225,
283
+ 0.019544063568115236,
284
+ 0.019704832077026366,
285
+ 0.01966592025756836,
286
+ 0.019917823791503905,
287
+ 0.019812351226806642,
288
+ 0.018713600158691408,
289
+ 0.018699264526367186,
290
+ 0.01948057556152344,
291
+ 0.019749887466430666,
292
+ 0.02001919937133789,
293
+ 0.020006912231445313,
294
+ 0.019703807830810546,
295
+ 0.01966796875,
296
+ 0.018710527420043945,
297
+ 0.01865727996826172,
298
+ 0.018942975997924806,
299
+ 0.01863987159729004,
300
+ 0.018687999725341797,
301
+ 0.019734527587890623,
302
+ 0.01984614372253418,
303
+ 0.019612672805786133,
304
+ 0.019629056930541993,
305
+ 0.01963315200805664,
306
+ 0.03807129669189453,
307
+ 0.01883033561706543,
308
+ 0.01945395278930664,
309
+ 0.01969254493713379,
310
+ 0.019532800674438477,
311
+ 0.019482624053955077,
312
+ 0.01967616081237793,
313
+ 0.019655679702758787,
314
+ 0.018547712326049806,
315
+ 0.018739200592041014,
316
+ 0.018529279708862305,
317
+ 0.019509248733520508,
318
+ 0.01943142318725586,
319
+ 0.01946009635925293,
320
+ 0.019715072631835938,
321
+ 0.0196628475189209,
322
+ 0.01862348747253418,
323
+ 0.018557952880859374,
324
+ 0.018587648391723634,
325
+ 0.018618368148803712,
326
+ 0.018719743728637696,
327
+ 0.018543615341186523,
328
+ 0.018535423278808593,
329
+ 0.01859686470031738,
330
+ 0.018570240020751954,
331
+ 0.018534400939941405,
332
+ 0.018589696884155273,
333
+ 0.018581504821777343,
334
+ 0.018787328720092773,
335
+ 0.018733055114746093,
336
+ 0.01860095977783203,
337
+ 0.01860505676269531,
338
+ 0.018527231216430663,
339
+ 0.018548736572265623,
340
+ 0.018749439239501953,
341
+ 0.01867263984680176,
342
+ 0.018551807403564453,
343
+ 0.018551807403564453,
344
+ 0.018563072204589845,
345
+ 0.01863270378112793,
346
+ 0.018561023712158203,
347
+ 0.018703359603881836,
348
+ 0.018522111892700196,
349
+ 0.018520063400268554,
350
+ 0.018515968322753908,
351
+ 0.018580480575561522,
352
+ 0.01863680076599121,
353
+ 0.018718719482421875,
354
+ 0.018548736572265623,
355
+ 0.01862144088745117,
356
+ 0.018526208877563476,
357
+ 0.018565120697021483,
358
+ 0.018544639587402344,
359
+ 0.018566144943237304,
360
+ 0.018571264266967775,
361
+ 0.018551807403564453,
362
+ 0.018721792221069337,
363
+ 0.018758655548095703,
364
+ 0.018751487731933594,
365
+ 0.018719743728637696,
366
+ 0.01879756736755371,
367
+ 0.018869247436523438,
368
+ 0.018727935791015626,
369
+ 0.018746368408203123,
370
+ 0.01866854476928711,
371
+ 0.01866649627685547,
372
+ 0.0186562557220459,
373
+ 0.018682880401611326,
374
+ 0.018710527420043945,
375
+ 0.018883583068847656,
376
+ 0.018729984283447267,
377
+ 0.018669567108154296,
378
+ 0.01882009506225586,
379
+ 0.018708480834960937,
380
+ 0.018697216033935548,
381
+ 0.01866547203063965,
382
+ 0.018671615600585938,
383
+ 0.01881395149230957,
384
+ 0.018880512237548826,
385
+ 0.01869004821777344,
386
+ 0.01884262466430664,
387
+ 0.01862144088745117,
388
+ 0.018684928894042968,
389
+ 0.01865727996826172,
390
+ 0.018669567108154296,
391
+ 0.018768896102905275,
392
+ 0.018926591873168946,
393
+ 0.018705408096313478,
394
+ 0.01884671974182129,
395
+ 0.01868083190917969,
396
+ 0.01864806365966797,
397
+ 0.018655231475830078,
398
+ 0.01864192008972168,
399
+ 0.01868083190917969,
400
+ 0.018945024490356444,
401
+ 0.018734079360961914,
402
+ 0.018754560470581053,
403
+ 0.018683904647827147,
404
+ 0.018671615600585938,
405
+ 0.01865727996826172,
406
+ 0.01885081672668457,
407
+ 0.018646015167236327,
408
+ 0.018682880401611326,
409
+ 0.018667520523071288,
410
+ 0.01865318489074707,
411
+ 0.018700288772583007,
412
+ 0.018840576171875,
413
+ 0.01861324882507324,
414
+ 0.018717695236206054,
415
+ 0.018700288772583007,
416
+ 0.01864192008972168,
417
+ 0.018646015167236327,
418
+ 0.018663423538208008,
419
+ 0.018882560729980468,
420
+ 0.018701311111450195,
421
+ 0.01868083190917969,
422
+ 0.018593791961669923,
423
+ 0.01863987159729004,
424
+ 0.01861529541015625,
425
+ 0.018962432861328125,
426
+ 0.018684928894042968,
427
+ 0.018655231475830078,
428
+ 0.018683904647827147,
429
+ 0.018637823104858398,
430
+ 0.01879449653625488,
431
+ 0.018815999984741212,
432
+ 0.01868185615539551
433
+ ]
434
+ },
435
+ "throughput": {
436
+ "unit": "tokens/s",
437
+ "value": 51.81817709979811
438
+ },
439
+ "energy": null,
440
+ "efficiency": null
441
+ }
442
+ }
443
+ }
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=False/benchmark.log ADDED
File without changes
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=False/benchmark_config.json ADDED
@@ -0,0 +1,102 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "pytorch_generate",
3
+ "backend": {
4
+ "name": "pytorch",
5
+ "version": "2.3.0+cu121",
6
+ "_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
7
+ "task": "text-generation",
8
+ "library": "transformers",
9
+ "model": "google/gemma-2b",
10
+ "processor": "google/gemma-2b",
11
+ "device": "cuda",
12
+ "device_ids": "0",
13
+ "seed": 42,
14
+ "inter_op_num_threads": null,
15
+ "intra_op_num_threads": null,
16
+ "model_kwargs": {},
17
+ "processor_kwargs": {},
18
+ "hub_kwargs": {},
19
+ "no_weights": true,
20
+ "device_map": null,
21
+ "torch_dtype": "float16",
22
+ "eval_mode": true,
23
+ "to_bettertransformer": false,
24
+ "low_cpu_mem_usage": null,
25
+ "attn_implementation": null,
26
+ "cache_implementation": null,
27
+ "autocast_enabled": false,
28
+ "autocast_dtype": null,
29
+ "torch_compile": false,
30
+ "torch_compile_target": "forward",
31
+ "torch_compile_config": {
32
+ "backend": "inductor",
33
+ "mode": "reduce-overhead",
34
+ "fullgraph": true
35
+ },
36
+ "quantization_scheme": null,
37
+ "quantization_config": {},
38
+ "deepspeed_inference": false,
39
+ "deepspeed_inference_config": {},
40
+ "peft_type": null,
41
+ "peft_config": {}
42
+ },
43
+ "scenario": {
44
+ "name": "inference",
45
+ "_target_": "optimum_benchmark.scenarios.inference.scenario.InferenceScenario",
46
+ "iterations": 2,
47
+ "duration": 0,
48
+ "warmup_runs": 10,
49
+ "input_shapes": {
50
+ "batch_size": 1,
51
+ "num_choices": 2,
52
+ "sequence_length": 7
53
+ },
54
+ "new_tokens": null,
55
+ "latency": true,
56
+ "memory": true,
57
+ "energy": false,
58
+ "forward_kwargs": {},
59
+ "generate_kwargs": {
60
+ "max_new_tokens": 128,
61
+ "min_new_tokens": 128,
62
+ "do_sample": false
63
+ },
64
+ "call_kwargs": {}
65
+ },
66
+ "launcher": {
67
+ "name": "process",
68
+ "_target_": "optimum_benchmark.launchers.process.launcher.ProcessLauncher",
69
+ "device_isolation": true,
70
+ "device_isolation_action": "warn",
71
+ "start_method": "spawn"
72
+ },
73
+ "environment": {
74
+ "cpu": " AMD EPYC 7R32",
75
+ "cpu_count": 16,
76
+ "cpu_ram_mb": 66697.29792,
77
+ "system": "Linux",
78
+ "machine": "x86_64",
79
+ "platform": "Linux-5.10.217-205.860.amzn2.x86_64-x86_64-with-glibc2.29",
80
+ "processor": "x86_64",
81
+ "python_version": "3.8.10",
82
+ "gpu": [
83
+ "NVIDIA A10G"
84
+ ],
85
+ "gpu_count": 1,
86
+ "gpu_vram_mb": 24146608128,
87
+ "optimum_benchmark_version": "0.2.1",
88
+ "optimum_benchmark_commit": null,
89
+ "transformers_version": "4.42.0.dev0",
90
+ "transformers_commit": "bd5091df8db7cea1a9f94f797fc11487f840ade1",
91
+ "accelerate_version": "0.31.0.dev0",
92
+ "accelerate_commit": null,
93
+ "diffusers_version": null,
94
+ "diffusers_commit": null,
95
+ "optimum_version": "1.21.0.dev0",
96
+ "optimum_commit": null,
97
+ "timm_version": "0.9.16",
98
+ "timm_commit": null,
99
+ "peft_version": "0.11.2.dev0",
100
+ "peft_commit": null
101
+ }
102
+ }
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=False/benchmark_report.json ADDED
@@ -0,0 +1,339 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prefill": {
3
+ "memory": {
4
+ "unit": "MB",
5
+ "max_ram": 1704.648704,
6
+ "max_global_vram": 6793.199616,
7
+ "max_process_vram": 0.0,
8
+ "max_reserved": 6146.752512,
9
+ "max_allocated": 5034.576896
10
+ },
11
+ "latency": {
12
+ "unit": "s",
13
+ "count": 2,
14
+ "total": 0.040267551422119144,
15
+ "mean": 0.020133775711059572,
16
+ "stdev": 0.0008556003570556626,
17
+ "p50": 0.020133775711059572,
18
+ "p90": 0.020818255996704102,
19
+ "p95": 0.020903816032409667,
20
+ "p99": 0.02097226406097412,
21
+ "values": [
22
+ 0.020989376068115233,
23
+ 0.019278175354003908
24
+ ]
25
+ },
26
+ "throughput": {
27
+ "unit": "tokens/s",
28
+ "value": 347.67447996129556
29
+ },
30
+ "energy": null,
31
+ "efficiency": null
32
+ },
33
+ "decode": {
34
+ "memory": {
35
+ "unit": "MB",
36
+ "max_ram": 1704.726528,
37
+ "max_global_vram": 6795.296768,
38
+ "max_process_vram": 0.0,
39
+ "max_reserved": 6148.849664,
40
+ "max_allocated": 5034.577408
41
+ },
42
+ "latency": {
43
+ "unit": "s",
44
+ "count": 2,
45
+ "total": 4.883418212890625,
46
+ "mean": 2.4417091064453125,
47
+ "stdev": 0.05995861816406256,
48
+ "p50": 2.4417091064453125,
49
+ "p90": 2.4896760009765626,
50
+ "p95": 2.495671862792969,
51
+ "p99": 2.500468552246094,
52
+ "values": [
53
+ 2.501667724609375,
54
+ 2.38175048828125
55
+ ]
56
+ },
57
+ "throughput": {
58
+ "unit": "tokens/s",
59
+ "value": 52.0127478186331
60
+ },
61
+ "energy": null,
62
+ "efficiency": null
63
+ },
64
+ "per_token": {
65
+ "memory": null,
66
+ "latency": {
67
+ "unit": "s",
68
+ "count": 253,
69
+ "total": 4.88245658493042,
70
+ "mean": 0.019298247371266484,
71
+ "stdev": 0.001368086083430472,
72
+ "p50": 0.018860031127929687,
73
+ "p90": 0.02008965148925781,
74
+ "p95": 0.020330905532836913,
75
+ "p99": 0.020966685943603514,
76
+ "values": [
77
+ 0.020875263214111327,
78
+ 0.020847616195678712,
79
+ 0.020368383407592772,
80
+ 0.020330495834350586,
81
+ 0.02045849609375,
82
+ 0.020093952178955078,
83
+ 0.020340736389160157,
84
+ 0.020060159683227538,
85
+ 0.02003763198852539,
86
+ 0.020312063217163084,
87
+ 0.019573759078979493,
88
+ 0.019325952529907226,
89
+ 0.018966527938842775,
90
+ 0.019517440795898438,
91
+ 0.019360767364501954,
92
+ 0.020485120773315428,
93
+ 0.02046771240234375,
94
+ 0.020135936737060548,
95
+ 0.020221952438354493,
96
+ 0.02389708709716797,
97
+ 0.02106572723388672,
98
+ 0.02044825553894043,
99
+ 0.02006220817565918,
100
+ 0.020006912231445313,
101
+ 0.020107263565063475,
102
+ 0.019998720169067383,
103
+ 0.020331520080566406,
104
+ 0.019545087814331053,
105
+ 0.01947340774536133,
106
+ 0.01924505615234375,
107
+ 0.01922969627380371,
108
+ 0.01940787124633789,
109
+ 0.02029363250732422,
110
+ 0.020384767532348632,
111
+ 0.020206592559814454,
112
+ 0.020295679092407228,
113
+ 0.019958784103393554,
114
+ 0.019999744415283204,
115
+ 0.020195327758789062,
116
+ 0.02004582405090332,
117
+ 0.020176895141601564,
118
+ 0.0200130558013916,
119
+ 0.02007244873046875,
120
+ 0.019893247604370116,
121
+ 0.019912704467773438,
122
+ 0.020329471588134765,
123
+ 0.01985536003112793,
124
+ 0.01983795166015625,
125
+ 0.019800064086914062,
126
+ 0.019962879180908204,
127
+ 0.02010316848754883,
128
+ 0.019629056930541993,
129
+ 0.019495935440063478,
130
+ 0.019523584365844726,
131
+ 0.01965260887145996,
132
+ 0.018594816207885743,
133
+ 0.018861055374145508,
134
+ 0.01882931137084961,
135
+ 0.019094528198242186,
136
+ 0.019744768142700195,
137
+ 0.019742719650268553,
138
+ 0.019775487899780272,
139
+ 0.019877887725830077,
140
+ 0.01971609687805176,
141
+ 0.019719167709350584,
142
+ 0.019715072631835938,
143
+ 0.019720191955566405,
144
+ 0.019887104034423828,
145
+ 0.019720191955566405,
146
+ 0.019687423706054686,
147
+ 0.01886720085144043,
148
+ 0.018889728546142577,
149
+ 0.018729984283447267,
150
+ 0.018740224838256835,
151
+ 0.018711551666259766,
152
+ 0.018711551666259766,
153
+ 0.019500032424926757,
154
+ 0.01990656089782715,
155
+ 0.019615743637084963,
156
+ 0.0196945915222168,
157
+ 0.019612672805786133,
158
+ 0.019586048126220702,
159
+ 0.01982361602783203,
160
+ 0.01966796875,
161
+ 0.01966080093383789,
162
+ 0.019713024139404296,
163
+ 0.019801088333129883,
164
+ 0.01962495994567871,
165
+ 0.01965670394897461,
166
+ 0.01967103958129883,
167
+ 0.0196495361328125,
168
+ 0.019693567276000978,
169
+ 0.019680255889892577,
170
+ 0.01966694450378418,
171
+ 0.019987455368041994,
172
+ 0.019655679702758787,
173
+ 0.018702335357666015,
174
+ 0.018683904647827147,
175
+ 0.018675712585449217,
176
+ 0.0187064323425293,
177
+ 0.01882521629333496,
178
+ 0.018860031127929687,
179
+ 0.018714624404907225,
180
+ 0.019544063568115236,
181
+ 0.019704832077026366,
182
+ 0.01966592025756836,
183
+ 0.019917823791503905,
184
+ 0.019812351226806642,
185
+ 0.018713600158691408,
186
+ 0.018699264526367186,
187
+ 0.01948057556152344,
188
+ 0.019749887466430666,
189
+ 0.02001919937133789,
190
+ 0.020006912231445313,
191
+ 0.019703807830810546,
192
+ 0.01966796875,
193
+ 0.018710527420043945,
194
+ 0.01865727996826172,
195
+ 0.018942975997924806,
196
+ 0.01863987159729004,
197
+ 0.018687999725341797,
198
+ 0.019734527587890623,
199
+ 0.01984614372253418,
200
+ 0.019612672805786133,
201
+ 0.019629056930541993,
202
+ 0.01963315200805664,
203
+ 0.03807129669189453,
204
+ 0.01883033561706543,
205
+ 0.01945395278930664,
206
+ 0.01969254493713379,
207
+ 0.019532800674438477,
208
+ 0.019482624053955077,
209
+ 0.01967616081237793,
210
+ 0.019655679702758787,
211
+ 0.018547712326049806,
212
+ 0.018739200592041014,
213
+ 0.018529279708862305,
214
+ 0.019509248733520508,
215
+ 0.01943142318725586,
216
+ 0.01946009635925293,
217
+ 0.019715072631835938,
218
+ 0.0196628475189209,
219
+ 0.01862348747253418,
220
+ 0.018557952880859374,
221
+ 0.018587648391723634,
222
+ 0.018618368148803712,
223
+ 0.018719743728637696,
224
+ 0.018543615341186523,
225
+ 0.018535423278808593,
226
+ 0.01859686470031738,
227
+ 0.018570240020751954,
228
+ 0.018534400939941405,
229
+ 0.018589696884155273,
230
+ 0.018581504821777343,
231
+ 0.018787328720092773,
232
+ 0.018733055114746093,
233
+ 0.01860095977783203,
234
+ 0.01860505676269531,
235
+ 0.018527231216430663,
236
+ 0.018548736572265623,
237
+ 0.018749439239501953,
238
+ 0.01867263984680176,
239
+ 0.018551807403564453,
240
+ 0.018551807403564453,
241
+ 0.018563072204589845,
242
+ 0.01863270378112793,
243
+ 0.018561023712158203,
244
+ 0.018703359603881836,
245
+ 0.018522111892700196,
246
+ 0.018520063400268554,
247
+ 0.018515968322753908,
248
+ 0.018580480575561522,
249
+ 0.01863680076599121,
250
+ 0.018718719482421875,
251
+ 0.018548736572265623,
252
+ 0.01862144088745117,
253
+ 0.018526208877563476,
254
+ 0.018565120697021483,
255
+ 0.018544639587402344,
256
+ 0.018566144943237304,
257
+ 0.018571264266967775,
258
+ 0.018551807403564453,
259
+ 0.018721792221069337,
260
+ 0.018758655548095703,
261
+ 0.018751487731933594,
262
+ 0.018719743728637696,
263
+ 0.01879756736755371,
264
+ 0.018869247436523438,
265
+ 0.018727935791015626,
266
+ 0.018746368408203123,
267
+ 0.01866854476928711,
268
+ 0.01866649627685547,
269
+ 0.0186562557220459,
270
+ 0.018682880401611326,
271
+ 0.018710527420043945,
272
+ 0.018883583068847656,
273
+ 0.018729984283447267,
274
+ 0.018669567108154296,
275
+ 0.01882009506225586,
276
+ 0.018708480834960937,
277
+ 0.018697216033935548,
278
+ 0.01866547203063965,
279
+ 0.018671615600585938,
280
+ 0.01881395149230957,
281
+ 0.018880512237548826,
282
+ 0.01869004821777344,
283
+ 0.01884262466430664,
284
+ 0.01862144088745117,
285
+ 0.018684928894042968,
286
+ 0.01865727996826172,
287
+ 0.018669567108154296,
288
+ 0.018768896102905275,
289
+ 0.018926591873168946,
290
+ 0.018705408096313478,
291
+ 0.01884671974182129,
292
+ 0.01868083190917969,
293
+ 0.01864806365966797,
294
+ 0.018655231475830078,
295
+ 0.01864192008972168,
296
+ 0.01868083190917969,
297
+ 0.018945024490356444,
298
+ 0.018734079360961914,
299
+ 0.018754560470581053,
300
+ 0.018683904647827147,
301
+ 0.018671615600585938,
302
+ 0.01865727996826172,
303
+ 0.01885081672668457,
304
+ 0.018646015167236327,
305
+ 0.018682880401611326,
306
+ 0.018667520523071288,
307
+ 0.01865318489074707,
308
+ 0.018700288772583007,
309
+ 0.018840576171875,
310
+ 0.01861324882507324,
311
+ 0.018717695236206054,
312
+ 0.018700288772583007,
313
+ 0.01864192008972168,
314
+ 0.018646015167236327,
315
+ 0.018663423538208008,
316
+ 0.018882560729980468,
317
+ 0.018701311111450195,
318
+ 0.01868083190917969,
319
+ 0.018593791961669923,
320
+ 0.01863987159729004,
321
+ 0.01861529541015625,
322
+ 0.018962432861328125,
323
+ 0.018684928894042968,
324
+ 0.018655231475830078,
325
+ 0.018683904647827147,
326
+ 0.018637823104858398,
327
+ 0.01879449653625488,
328
+ 0.018815999984741212,
329
+ 0.01868185615539551
330
+ ]
331
+ },
332
+ "throughput": {
333
+ "unit": "tokens/s",
334
+ "value": 51.81817709979811
335
+ },
336
+ "energy": null,
337
+ "efficiency": null
338
+ }
339
+ }
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=False/summary.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model": "google/gemma-2b",
3
+ "commit": "bd5091df8db7cea1a9f94f797fc11487f840ade1",
4
+ "config": "backend.cache_implementation=null,backend.torch_compile=False",
5
+ "metrics": {
6
+ "prefill.latency.mean": 0.020133775711059572,
7
+ "prefill.throughput.value": 347.67447996129556,
8
+ "decode.latency.mean": 2.4417091064453125,
9
+ "decode.throughput.value": 52.0127478186331,
10
+ "per_token.latency.mean": 0.019298247371266484,
11
+ "per_token.throughput.value": 51.81817709979811
12
+ }
13
+ }
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=True/.hydra/config.yaml ADDED
@@ -0,0 +1,91 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ name: pytorch_generate
2
+ backend:
3
+ name: pytorch
4
+ version: 2.3.0+cu121
5
+ _target_: optimum_benchmark.backends.pytorch.backend.PyTorchBackend
6
+ task: null
7
+ library: null
8
+ model: google/gemma-2b
9
+ processor: null
10
+ device: cuda
11
+ device_ids: '0'
12
+ seed: 42
13
+ inter_op_num_threads: null
14
+ intra_op_num_threads: null
15
+ model_kwargs: {}
16
+ processor_kwargs: {}
17
+ hub_kwargs: {}
18
+ no_weights: true
19
+ device_map: null
20
+ torch_dtype: float16
21
+ eval_mode: true
22
+ to_bettertransformer: false
23
+ low_cpu_mem_usage: null
24
+ attn_implementation: null
25
+ cache_implementation: null
26
+ autocast_enabled: false
27
+ autocast_dtype: null
28
+ torch_compile: true
29
+ torch_compile_target: forward
30
+ torch_compile_config:
31
+ backend: inductor
32
+ mode: reduce-overhead
33
+ fullgraph: true
34
+ quantization_scheme: null
35
+ quantization_config: {}
36
+ deepspeed_inference: false
37
+ deepspeed_inference_config: {}
38
+ peft_type: null
39
+ peft_config: {}
40
+ scenario:
41
+ name: inference
42
+ _target_: optimum_benchmark.scenarios.inference.scenario.InferenceScenario
43
+ iterations: 2
44
+ duration: 0
45
+ warmup_runs: 10
46
+ input_shapes:
47
+ batch_size: 1
48
+ sequence_length: 7
49
+ new_tokens: null
50
+ latency: true
51
+ memory: true
52
+ energy: false
53
+ forward_kwargs: {}
54
+ generate_kwargs:
55
+ max_new_tokens: 128
56
+ min_new_tokens: 128
57
+ do_sample: false
58
+ call_kwargs: {}
59
+ launcher:
60
+ name: process
61
+ _target_: optimum_benchmark.launchers.process.launcher.ProcessLauncher
62
+ device_isolation: true
63
+ device_isolation_action: warn
64
+ start_method: spawn
65
+ environment:
66
+ cpu: ' AMD EPYC 7R32'
67
+ cpu_count: 16
68
+ cpu_ram_mb: 66697.29792
69
+ system: Linux
70
+ machine: x86_64
71
+ platform: Linux-5.10.217-205.860.amzn2.x86_64-x86_64-with-glibc2.29
72
+ processor: x86_64
73
+ python_version: 3.8.10
74
+ gpu:
75
+ - NVIDIA A10G
76
+ gpu_count: 1
77
+ gpu_vram_mb: 24146608128
78
+ optimum_benchmark_version: 0.2.1
79
+ optimum_benchmark_commit: null
80
+ transformers_version: 4.42.0.dev0
81
+ transformers_commit: bd5091df8db7cea1a9f94f797fc11487f840ade1
82
+ accelerate_version: 0.31.0.dev0
83
+ accelerate_commit: null
84
+ diffusers_version: null
85
+ diffusers_commit: null
86
+ optimum_version: 1.21.0.dev0
87
+ optimum_commit: null
88
+ timm_version: 0.9.16
89
+ timm_commit: null
90
+ peft_version: 0.11.2.dev0
91
+ peft_commit: null
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=True/.hydra/hydra.yaml ADDED
@@ -0,0 +1,141 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ hydra:
2
+ run:
3
+ dir: _benchmark/tmpjivn_1kh/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/${hydra.job.override_dirname}
4
+ sweep:
5
+ dir: _benchmark/tmpjivn_1kh/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1
6
+ subdir: ${hydra.job.override_dirname}
7
+ launcher:
8
+ _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
9
+ sweeper:
10
+ _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
11
+ max_batch_size: null
12
+ params: null
13
+ help:
14
+ app_name: ${hydra.job.name}
15
+ header: '${hydra.help.app_name} is powered by Hydra.
16
+
17
+ '
18
+ footer: 'Powered by Hydra (https://hydra.cc)
19
+
20
+ Use --hydra-help to view Hydra specific help
21
+
22
+ '
23
+ template: '${hydra.help.header}
24
+
25
+ == Configuration groups ==
26
+
27
+ Compose your configuration from those groups (group=option)
28
+
29
+
30
+ $APP_CONFIG_GROUPS
31
+
32
+
33
+ == Config ==
34
+
35
+ Override anything in the config (foo.bar=value)
36
+
37
+
38
+ $CONFIG
39
+
40
+
41
+ ${hydra.help.footer}
42
+
43
+ '
44
+ hydra_help:
45
+ template: 'Hydra (${hydra.runtime.version})
46
+
47
+ See https://hydra.cc for more info.
48
+
49
+
50
+ == Flags ==
51
+
52
+ $FLAGS_HELP
53
+
54
+
55
+ == Configuration groups ==
56
+
57
+ Compose your configuration from those groups (For example, append hydra/job_logging=disabled
58
+ to command line)
59
+
60
+
61
+ $HYDRA_CONFIG_GROUPS
62
+
63
+
64
+ Use ''--cfg hydra'' to Show the Hydra config.
65
+
66
+ '
67
+ hydra_help: ???
68
+ hydra_logging:
69
+ version: 1
70
+ root:
71
+ level: ERROR
72
+ disable_existing_loggers: true
73
+ job_logging:
74
+ version: 1
75
+ root:
76
+ level: ERROR
77
+ disable_existing_loggers: true
78
+ env: {}
79
+ mode: MULTIRUN
80
+ searchpath: []
81
+ callbacks: {}
82
+ output_subdir: .hydra
83
+ overrides:
84
+ hydra:
85
+ - hydra/job_logging=disabled
86
+ - hydra/hydra_logging=disabled
87
+ - hydra.sweep.dir=_benchmark/tmpjivn_1kh/commit\=bd5091df8db7cea1a9f94f797fc11487f840ade1
88
+ - hydra.run.dir=_benchmark/tmpjivn_1kh/commit\=bd5091df8db7cea1a9f94f797fc11487f840ade1/${hydra.job.override_dirname}
89
+ - hydra.mode=MULTIRUN
90
+ task:
91
+ - backend.model=google/gemma-2b
92
+ - backend.cache_implementation=null
93
+ - backend.torch_compile=True
94
+ job:
95
+ name: cli
96
+ chdir: true
97
+ override_dirname: backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=True
98
+ id: '1'
99
+ num: 1
100
+ config_name: generation
101
+ env_set:
102
+ OVERRIDE_BENCHMARKS: '1'
103
+ LOG_LEVEL: WARN
104
+ env_copy: []
105
+ config:
106
+ override_dirname:
107
+ kv_sep: '='
108
+ item_sep: ','
109
+ exclude_keys: []
110
+ runtime:
111
+ version: 1.3.2
112
+ version_base: '1.3'
113
+ cwd: /transformers
114
+ config_sources:
115
+ - path: hydra.conf
116
+ schema: pkg
117
+ provider: hydra
118
+ - path: optimum_benchmark
119
+ schema: pkg
120
+ provider: main
121
+ - path: /transformers/benchmark/config
122
+ schema: file
123
+ provider: command-line
124
+ - path: ''
125
+ schema: structured
126
+ provider: schema
127
+ output_dir: /transformers/_benchmark/tmpjivn_1kh/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=True
128
+ choices:
129
+ backend: pytorch
130
+ launcher: process
131
+ scenario: inference
132
+ hydra/env: default
133
+ hydra/callbacks: null
134
+ hydra/job_logging: disabled
135
+ hydra/hydra_logging: disabled
136
+ hydra/hydra_help: default
137
+ hydra/help: default
138
+ hydra/sweeper: basic
139
+ hydra/launcher: basic
140
+ hydra/output: default
141
+ verbose: false
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=True/.hydra/overrides.yaml ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ - backend.model=google/gemma-2b
2
+ - backend.cache_implementation=null
3
+ - backend.torch_compile=True
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=True/benchmark.log ADDED
@@ -0,0 +1 @@
 
 
1
+ [ISOLATED-PROCESS][2024-06-05 15:13:25,492][process][ERROR] - + Sending traceback to main process
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=null,backend.model=google/gemma-2b,backend.torch_compile=True/benchmark_config.json ADDED
@@ -0,0 +1,102 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "pytorch_generate",
3
+ "backend": {
4
+ "name": "pytorch",
5
+ "version": "2.3.0+cu121",
6
+ "_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
7
+ "task": "text-generation",
8
+ "library": "transformers",
9
+ "model": "google/gemma-2b",
10
+ "processor": "google/gemma-2b",
11
+ "device": "cuda",
12
+ "device_ids": "0",
13
+ "seed": 42,
14
+ "inter_op_num_threads": null,
15
+ "intra_op_num_threads": null,
16
+ "model_kwargs": {},
17
+ "processor_kwargs": {},
18
+ "hub_kwargs": {},
19
+ "no_weights": true,
20
+ "device_map": null,
21
+ "torch_dtype": "float16",
22
+ "eval_mode": true,
23
+ "to_bettertransformer": false,
24
+ "low_cpu_mem_usage": null,
25
+ "attn_implementation": null,
26
+ "cache_implementation": null,
27
+ "autocast_enabled": false,
28
+ "autocast_dtype": null,
29
+ "torch_compile": true,
30
+ "torch_compile_target": "forward",
31
+ "torch_compile_config": {
32
+ "backend": "inductor",
33
+ "mode": "reduce-overhead",
34
+ "fullgraph": true
35
+ },
36
+ "quantization_scheme": null,
37
+ "quantization_config": {},
38
+ "deepspeed_inference": false,
39
+ "deepspeed_inference_config": {},
40
+ "peft_type": null,
41
+ "peft_config": {}
42
+ },
43
+ "scenario": {
44
+ "name": "inference",
45
+ "_target_": "optimum_benchmark.scenarios.inference.scenario.InferenceScenario",
46
+ "iterations": 2,
47
+ "duration": 0,
48
+ "warmup_runs": 10,
49
+ "input_shapes": {
50
+ "batch_size": 1,
51
+ "num_choices": 2,
52
+ "sequence_length": 7
53
+ },
54
+ "new_tokens": null,
55
+ "latency": true,
56
+ "memory": true,
57
+ "energy": false,
58
+ "forward_kwargs": {},
59
+ "generate_kwargs": {
60
+ "max_new_tokens": 128,
61
+ "min_new_tokens": 128,
62
+ "do_sample": false
63
+ },
64
+ "call_kwargs": {}
65
+ },
66
+ "launcher": {
67
+ "name": "process",
68
+ "_target_": "optimum_benchmark.launchers.process.launcher.ProcessLauncher",
69
+ "device_isolation": true,
70
+ "device_isolation_action": "warn",
71
+ "start_method": "spawn"
72
+ },
73
+ "environment": {
74
+ "cpu": " AMD EPYC 7R32",
75
+ "cpu_count": 16,
76
+ "cpu_ram_mb": 66697.29792,
77
+ "system": "Linux",
78
+ "machine": "x86_64",
79
+ "platform": "Linux-5.10.217-205.860.amzn2.x86_64-x86_64-with-glibc2.29",
80
+ "processor": "x86_64",
81
+ "python_version": "3.8.10",
82
+ "gpu": [
83
+ "NVIDIA A10G"
84
+ ],
85
+ "gpu_count": 1,
86
+ "gpu_vram_mb": 24146608128,
87
+ "optimum_benchmark_version": "0.2.1",
88
+ "optimum_benchmark_commit": null,
89
+ "transformers_version": "4.42.0.dev0",
90
+ "transformers_commit": "bd5091df8db7cea1a9f94f797fc11487f840ade1",
91
+ "accelerate_version": "0.31.0.dev0",
92
+ "accelerate_commit": null,
93
+ "diffusers_version": null,
94
+ "diffusers_commit": null,
95
+ "optimum_version": "1.21.0.dev0",
96
+ "optimum_commit": null,
97
+ "timm_version": "0.9.16",
98
+ "timm_commit": null,
99
+ "peft_version": "0.11.2.dev0",
100
+ "peft_commit": null
101
+ }
102
+ }
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=False/.hydra/config.yaml ADDED
@@ -0,0 +1,91 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ name: pytorch_generate
2
+ backend:
3
+ name: pytorch
4
+ version: 2.3.0+cu121
5
+ _target_: optimum_benchmark.backends.pytorch.backend.PyTorchBackend
6
+ task: null
7
+ library: null
8
+ model: google/gemma-2b
9
+ processor: null
10
+ device: cuda
11
+ device_ids: '0'
12
+ seed: 42
13
+ inter_op_num_threads: null
14
+ intra_op_num_threads: null
15
+ model_kwargs: {}
16
+ processor_kwargs: {}
17
+ hub_kwargs: {}
18
+ no_weights: true
19
+ device_map: null
20
+ torch_dtype: float16
21
+ eval_mode: true
22
+ to_bettertransformer: false
23
+ low_cpu_mem_usage: null
24
+ attn_implementation: null
25
+ cache_implementation: static
26
+ autocast_enabled: false
27
+ autocast_dtype: null
28
+ torch_compile: false
29
+ torch_compile_target: forward
30
+ torch_compile_config:
31
+ backend: inductor
32
+ mode: reduce-overhead
33
+ fullgraph: true
34
+ quantization_scheme: null
35
+ quantization_config: {}
36
+ deepspeed_inference: false
37
+ deepspeed_inference_config: {}
38
+ peft_type: null
39
+ peft_config: {}
40
+ scenario:
41
+ name: inference
42
+ _target_: optimum_benchmark.scenarios.inference.scenario.InferenceScenario
43
+ iterations: 2
44
+ duration: 0
45
+ warmup_runs: 10
46
+ input_shapes:
47
+ batch_size: 1
48
+ sequence_length: 7
49
+ new_tokens: null
50
+ latency: true
51
+ memory: true
52
+ energy: false
53
+ forward_kwargs: {}
54
+ generate_kwargs:
55
+ max_new_tokens: 128
56
+ min_new_tokens: 128
57
+ do_sample: false
58
+ call_kwargs: {}
59
+ launcher:
60
+ name: process
61
+ _target_: optimum_benchmark.launchers.process.launcher.ProcessLauncher
62
+ device_isolation: true
63
+ device_isolation_action: warn
64
+ start_method: spawn
65
+ environment:
66
+ cpu: ' AMD EPYC 7R32'
67
+ cpu_count: 16
68
+ cpu_ram_mb: 66697.29792
69
+ system: Linux
70
+ machine: x86_64
71
+ platform: Linux-5.10.217-205.860.amzn2.x86_64-x86_64-with-glibc2.29
72
+ processor: x86_64
73
+ python_version: 3.8.10
74
+ gpu:
75
+ - NVIDIA A10G
76
+ gpu_count: 1
77
+ gpu_vram_mb: 24146608128
78
+ optimum_benchmark_version: 0.2.1
79
+ optimum_benchmark_commit: null
80
+ transformers_version: 4.42.0.dev0
81
+ transformers_commit: bd5091df8db7cea1a9f94f797fc11487f840ade1
82
+ accelerate_version: 0.31.0.dev0
83
+ accelerate_commit: null
84
+ diffusers_version: null
85
+ diffusers_commit: null
86
+ optimum_version: 1.21.0.dev0
87
+ optimum_commit: null
88
+ timm_version: 0.9.16
89
+ timm_commit: null
90
+ peft_version: 0.11.2.dev0
91
+ peft_commit: null
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=False/.hydra/hydra.yaml ADDED
@@ -0,0 +1,141 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ hydra:
2
+ run:
3
+ dir: _benchmark/tmpjivn_1kh/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/${hydra.job.override_dirname}
4
+ sweep:
5
+ dir: _benchmark/tmpjivn_1kh/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1
6
+ subdir: ${hydra.job.override_dirname}
7
+ launcher:
8
+ _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
9
+ sweeper:
10
+ _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
11
+ max_batch_size: null
12
+ params: null
13
+ help:
14
+ app_name: ${hydra.job.name}
15
+ header: '${hydra.help.app_name} is powered by Hydra.
16
+
17
+ '
18
+ footer: 'Powered by Hydra (https://hydra.cc)
19
+
20
+ Use --hydra-help to view Hydra specific help
21
+
22
+ '
23
+ template: '${hydra.help.header}
24
+
25
+ == Configuration groups ==
26
+
27
+ Compose your configuration from those groups (group=option)
28
+
29
+
30
+ $APP_CONFIG_GROUPS
31
+
32
+
33
+ == Config ==
34
+
35
+ Override anything in the config (foo.bar=value)
36
+
37
+
38
+ $CONFIG
39
+
40
+
41
+ ${hydra.help.footer}
42
+
43
+ '
44
+ hydra_help:
45
+ template: 'Hydra (${hydra.runtime.version})
46
+
47
+ See https://hydra.cc for more info.
48
+
49
+
50
+ == Flags ==
51
+
52
+ $FLAGS_HELP
53
+
54
+
55
+ == Configuration groups ==
56
+
57
+ Compose your configuration from those groups (For example, append hydra/job_logging=disabled
58
+ to command line)
59
+
60
+
61
+ $HYDRA_CONFIG_GROUPS
62
+
63
+
64
+ Use ''--cfg hydra'' to Show the Hydra config.
65
+
66
+ '
67
+ hydra_help: ???
68
+ hydra_logging:
69
+ version: 1
70
+ root:
71
+ level: ERROR
72
+ disable_existing_loggers: true
73
+ job_logging:
74
+ version: 1
75
+ root:
76
+ level: ERROR
77
+ disable_existing_loggers: true
78
+ env: {}
79
+ mode: MULTIRUN
80
+ searchpath: []
81
+ callbacks: {}
82
+ output_subdir: .hydra
83
+ overrides:
84
+ hydra:
85
+ - hydra/job_logging=disabled
86
+ - hydra/hydra_logging=disabled
87
+ - hydra.sweep.dir=_benchmark/tmpjivn_1kh/commit\=bd5091df8db7cea1a9f94f797fc11487f840ade1
88
+ - hydra.run.dir=_benchmark/tmpjivn_1kh/commit\=bd5091df8db7cea1a9f94f797fc11487f840ade1/${hydra.job.override_dirname}
89
+ - hydra.mode=MULTIRUN
90
+ task:
91
+ - backend.model=google/gemma-2b
92
+ - backend.cache_implementation=static
93
+ - backend.torch_compile=False
94
+ job:
95
+ name: cli
96
+ chdir: true
97
+ override_dirname: backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=False
98
+ id: '2'
99
+ num: 2
100
+ config_name: generation
101
+ env_set:
102
+ OVERRIDE_BENCHMARKS: '1'
103
+ LOG_LEVEL: WARN
104
+ env_copy: []
105
+ config:
106
+ override_dirname:
107
+ kv_sep: '='
108
+ item_sep: ','
109
+ exclude_keys: []
110
+ runtime:
111
+ version: 1.3.2
112
+ version_base: '1.3'
113
+ cwd: /transformers
114
+ config_sources:
115
+ - path: hydra.conf
116
+ schema: pkg
117
+ provider: hydra
118
+ - path: optimum_benchmark
119
+ schema: pkg
120
+ provider: main
121
+ - path: /transformers/benchmark/config
122
+ schema: file
123
+ provider: command-line
124
+ - path: ''
125
+ schema: structured
126
+ provider: schema
127
+ output_dir: /transformers/_benchmark/tmpjivn_1kh/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=False
128
+ choices:
129
+ backend: pytorch
130
+ launcher: process
131
+ scenario: inference
132
+ hydra/env: default
133
+ hydra/callbacks: null
134
+ hydra/job_logging: disabled
135
+ hydra/hydra_logging: disabled
136
+ hydra/hydra_help: default
137
+ hydra/help: default
138
+ hydra/sweeper: basic
139
+ hydra/launcher: basic
140
+ hydra/output: default
141
+ verbose: false
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=False/.hydra/overrides.yaml ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ - backend.model=google/gemma-2b
2
+ - backend.cache_implementation=static
3
+ - backend.torch_compile=False
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=False/benchmark.json ADDED
@@ -0,0 +1,443 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "config": {
3
+ "name": "pytorch_generate",
4
+ "backend": {
5
+ "name": "pytorch",
6
+ "version": "2.3.0+cu121",
7
+ "_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
8
+ "task": "text-generation",
9
+ "library": "transformers",
10
+ "model": "google/gemma-2b",
11
+ "processor": "google/gemma-2b",
12
+ "device": "cuda",
13
+ "device_ids": "0",
14
+ "seed": 42,
15
+ "inter_op_num_threads": null,
16
+ "intra_op_num_threads": null,
17
+ "model_kwargs": {},
18
+ "processor_kwargs": {},
19
+ "hub_kwargs": {},
20
+ "no_weights": true,
21
+ "device_map": null,
22
+ "torch_dtype": "float16",
23
+ "eval_mode": true,
24
+ "to_bettertransformer": false,
25
+ "low_cpu_mem_usage": null,
26
+ "attn_implementation": null,
27
+ "cache_implementation": "static",
28
+ "autocast_enabled": false,
29
+ "autocast_dtype": null,
30
+ "torch_compile": false,
31
+ "torch_compile_target": "forward",
32
+ "torch_compile_config": {
33
+ "backend": "inductor",
34
+ "mode": "reduce-overhead",
35
+ "fullgraph": true
36
+ },
37
+ "quantization_scheme": null,
38
+ "quantization_config": {},
39
+ "deepspeed_inference": false,
40
+ "deepspeed_inference_config": {},
41
+ "peft_type": null,
42
+ "peft_config": {}
43
+ },
44
+ "scenario": {
45
+ "name": "inference",
46
+ "_target_": "optimum_benchmark.scenarios.inference.scenario.InferenceScenario",
47
+ "iterations": 2,
48
+ "duration": 0,
49
+ "warmup_runs": 10,
50
+ "input_shapes": {
51
+ "batch_size": 1,
52
+ "num_choices": 2,
53
+ "sequence_length": 7
54
+ },
55
+ "new_tokens": null,
56
+ "latency": true,
57
+ "memory": true,
58
+ "energy": false,
59
+ "forward_kwargs": {},
60
+ "generate_kwargs": {
61
+ "max_new_tokens": 128,
62
+ "min_new_tokens": 128,
63
+ "do_sample": false
64
+ },
65
+ "call_kwargs": {}
66
+ },
67
+ "launcher": {
68
+ "name": "process",
69
+ "_target_": "optimum_benchmark.launchers.process.launcher.ProcessLauncher",
70
+ "device_isolation": true,
71
+ "device_isolation_action": "warn",
72
+ "start_method": "spawn"
73
+ },
74
+ "environment": {
75
+ "cpu": " AMD EPYC 7R32",
76
+ "cpu_count": 16,
77
+ "cpu_ram_mb": 66697.29792,
78
+ "system": "Linux",
79
+ "machine": "x86_64",
80
+ "platform": "Linux-5.10.217-205.860.amzn2.x86_64-x86_64-with-glibc2.29",
81
+ "processor": "x86_64",
82
+ "python_version": "3.8.10",
83
+ "gpu": [
84
+ "NVIDIA A10G"
85
+ ],
86
+ "gpu_count": 1,
87
+ "gpu_vram_mb": 24146608128,
88
+ "optimum_benchmark_version": "0.2.1",
89
+ "optimum_benchmark_commit": null,
90
+ "transformers_version": "4.42.0.dev0",
91
+ "transformers_commit": "bd5091df8db7cea1a9f94f797fc11487f840ade1",
92
+ "accelerate_version": "0.31.0.dev0",
93
+ "accelerate_commit": null,
94
+ "diffusers_version": null,
95
+ "diffusers_commit": null,
96
+ "optimum_version": "1.21.0.dev0",
97
+ "optimum_commit": null,
98
+ "timm_version": "0.9.16",
99
+ "timm_commit": null,
100
+ "peft_version": "0.11.2.dev0",
101
+ "peft_commit": null
102
+ }
103
+ },
104
+ "report": {
105
+ "prefill": {
106
+ "memory": {
107
+ "unit": "MB",
108
+ "max_ram": 1777.160192,
109
+ "max_global_vram": 6793.199616,
110
+ "max_process_vram": 0.0,
111
+ "max_reserved": 6146.752512,
112
+ "max_allocated": 5036.936192
113
+ },
114
+ "latency": {
115
+ "unit": "s",
116
+ "count": 2,
117
+ "total": 0.04430672073364258,
118
+ "mean": 0.02215336036682129,
119
+ "stdev": 0.00032209587097167813,
120
+ "p50": 0.02215336036682129,
121
+ "p90": 0.02241103706359863,
122
+ "p95": 0.0224432466506958,
123
+ "p99": 0.022469014320373532,
124
+ "values": [
125
+ 0.022475456237792967,
126
+ 0.02183126449584961
127
+ ]
128
+ },
129
+ "throughput": {
130
+ "unit": "tokens/s",
131
+ "value": 315.97915097719357
132
+ },
133
+ "energy": null,
134
+ "efficiency": null
135
+ },
136
+ "decode": {
137
+ "memory": {
138
+ "unit": "MB",
139
+ "max_ram": 1777.242112,
140
+ "max_global_vram": 6793.199616,
141
+ "max_process_vram": 0.0,
142
+ "max_reserved": 6146.752512,
143
+ "max_allocated": 5036.936704
144
+ },
145
+ "latency": {
146
+ "unit": "s",
147
+ "count": 2,
148
+ "total": 5.13177685546875,
149
+ "mean": 2.565888427734375,
150
+ "stdev": 0.0033464355468748597,
151
+ "p50": 2.565888427734375,
152
+ "p90": 2.568565576171875,
153
+ "p95": 2.5689002197265625,
154
+ "p99": 2.5691679345703125,
155
+ "values": [
156
+ 2.56923486328125,
157
+ 2.5625419921875
158
+ ]
159
+ },
160
+ "throughput": {
161
+ "unit": "tokens/s",
162
+ "value": 49.49552701796091
163
+ },
164
+ "energy": null,
165
+ "efficiency": null
166
+ },
167
+ "per_token": {
168
+ "memory": null,
169
+ "latency": {
170
+ "unit": "s",
171
+ "count": 253,
172
+ "total": 5.132801023483274,
173
+ "mean": 0.020287751080961567,
174
+ "stdev": 0.001393544041619272,
175
+ "p50": 0.020134912490844727,
176
+ "p90": 0.020422451019287108,
177
+ "p95": 0.020565196228027344,
178
+ "p99": 0.021351505813598633,
179
+ "values": [
180
+ 0.02208460807800293,
181
+ 0.020320255279541014,
182
+ 0.020093952178955078,
183
+ 0.020084735870361328,
184
+ 0.02020147132873535,
185
+ 0.020163583755493163,
186
+ 0.02011238479614258,
187
+ 0.02004582405090332,
188
+ 0.020247552871704103,
189
+ 0.020156415939331054,
190
+ 0.02044313621520996,
191
+ 0.020108287811279296,
192
+ 0.020092927932739257,
193
+ 0.02005606460571289,
194
+ 0.020098047256469728,
195
+ 0.02002943992614746,
196
+ 0.019914751052856446,
197
+ 0.02127667236328125,
198
+ 0.02065510368347168,
199
+ 0.020178943634033202,
200
+ 0.02022400093078613,
201
+ 0.020082687377929686,
202
+ 0.02026803207397461,
203
+ 0.020188159942626953,
204
+ 0.02007347106933594,
205
+ 0.020090879440307616,
206
+ 0.020247552871704103,
207
+ 0.02004787254333496,
208
+ 0.020208639144897463,
209
+ 0.020011007308959963,
210
+ 0.02021990394592285,
211
+ 0.02004582405090332,
212
+ 0.020148223876953125,
213
+ 0.020282367706298828,
214
+ 0.02017791938781738,
215
+ 0.020230144500732423,
216
+ 0.020215808868408205,
217
+ 0.02001817512512207,
218
+ 0.02002943992614746,
219
+ 0.020006912231445313,
220
+ 0.020114431381225584,
221
+ 0.020328447341918944,
222
+ 0.02007756805419922,
223
+ 0.02001408004760742,
224
+ 0.02009600067138672,
225
+ 0.0200447998046875,
226
+ 0.0202106876373291,
227
+ 0.020132863998413086,
228
+ 0.02045142364501953,
229
+ 0.020527008056640626,
230
+ 0.02075449562072754,
231
+ 0.02032736015319824,
232
+ 0.020100095748901366,
233
+ 0.02008883285522461,
234
+ 0.020238336563110353,
235
+ 0.020023296356201172,
236
+ 0.02011955261230469,
237
+ 0.020082687377929686,
238
+ 0.020118528366088868,
239
+ 0.02026188850402832,
240
+ 0.0202926082611084,
241
+ 0.020195327758789062,
242
+ 0.02024550437927246,
243
+ 0.020101119995117187,
244
+ 0.02012876892089844,
245
+ 0.020552703857421875,
246
+ 0.02035916709899902,
247
+ 0.021029888153076173,
248
+ 0.021415935516357423,
249
+ 0.02057318305969238,
250
+ 0.020413440704345705,
251
+ 0.02029465675354004,
252
+ 0.020117504119873047,
253
+ 0.020101119995117187,
254
+ 0.02011136054992676,
255
+ 0.020124671936035156,
256
+ 0.020182016372680665,
257
+ 0.02005401611328125,
258
+ 0.02009600067138672,
259
+ 0.020463615417480468,
260
+ 0.02049945640563965,
261
+ 0.020676607131958007,
262
+ 0.020747264862060546,
263
+ 0.020136959075927736,
264
+ 0.020125696182250977,
265
+ 0.020189184188842774,
266
+ 0.020157440185546875,
267
+ 0.020246528625488282,
268
+ 0.02007859230041504,
269
+ 0.020248575210571287,
270
+ 0.020113407135009767,
271
+ 0.020158464431762696,
272
+ 0.02008166313171387,
273
+ 0.02007859230041504,
274
+ 0.02025369644165039,
275
+ 0.020102144241333008,
276
+ 0.020044832229614257,
277
+ 0.0200928955078125,
278
+ 0.02020966339111328,
279
+ 0.020185087203979494,
280
+ 0.020126720428466797,
281
+ 0.02004582405090332,
282
+ 0.020067327499389647,
283
+ 0.02021785545349121,
284
+ 0.020125696182250977,
285
+ 0.020230144500732423,
286
+ 0.020068351745605468,
287
+ 0.020135936737060548,
288
+ 0.020102144241333008,
289
+ 0.020121599197387697,
290
+ 0.02026393508911133,
291
+ 0.020306943893432617,
292
+ 0.02016972732543945,
293
+ 0.020184064865112306,
294
+ 0.020135936737060548,
295
+ 0.020093952178955078,
296
+ 0.02022502326965332,
297
+ 0.020212736129760742,
298
+ 0.020089920043945313,
299
+ 0.02016966438293457,
300
+ 0.02012057685852051,
301
+ 0.02008576011657715,
302
+ 0.02023628807067871,
303
+ 0.020170751571655272,
304
+ 0.02004787254333496,
305
+ 0.02006425666809082,
306
+ 0.042106880187988284,
307
+ 0.02042470359802246,
308
+ 0.02011136054992676,
309
+ 0.020142080307006836,
310
+ 0.020065280914306642,
311
+ 0.020092927932739257,
312
+ 0.02023219108581543,
313
+ 0.020230144500732423,
314
+ 0.020076543807983398,
315
+ 0.020099071502685546,
316
+ 0.02003455924987793,
317
+ 0.020039680480957032,
318
+ 0.020361215591430663,
319
+ 0.020092927932739257,
320
+ 0.020093952178955078,
321
+ 0.02008678436279297,
322
+ 0.02000383949279785,
323
+ 0.02024140739440918,
324
+ 0.02020249557495117,
325
+ 0.02003558349609375,
326
+ 0.020098047256469728,
327
+ 0.020109312057495117,
328
+ 0.020063232421875,
329
+ 0.020264959335327147,
330
+ 0.01985740852355957,
331
+ 0.01984000015258789,
332
+ 0.019978239059448243,
333
+ 0.021292032241821288,
334
+ 0.020559871673583984,
335
+ 0.020134912490844727,
336
+ 0.02007142448425293,
337
+ 0.020098047256469728,
338
+ 0.02023321533203125,
339
+ 0.02023321533203125,
340
+ 0.020140031814575195,
341
+ 0.02007756805419922,
342
+ 0.02006630325317383,
343
+ 0.020076608657836913,
344
+ 0.02018911933898926,
345
+ 0.020315135955810547,
346
+ 0.0206059513092041,
347
+ 0.020067327499389647,
348
+ 0.019916799545288084,
349
+ 0.020093952178955078,
350
+ 0.020366336822509764,
351
+ 0.020518911361694335,
352
+ 0.020180992126464844,
353
+ 0.020140031814575195,
354
+ 0.02011238479614258,
355
+ 0.02008883285522461,
356
+ 0.02023526382446289,
357
+ 0.02010419273376465,
358
+ 0.020180992126464844,
359
+ 0.02006937599182129,
360
+ 0.020136959075927736,
361
+ 0.020099071502685546,
362
+ 0.020129791259765627,
363
+ 0.020247552871704103,
364
+ 0.020189184188842774,
365
+ 0.02006118392944336,
366
+ 0.02011238479614258,
367
+ 0.020131839752197265,
368
+ 0.020238336563110353,
369
+ 0.020508672714233397,
370
+ 0.02049228858947754,
371
+ 0.020607999801635742,
372
+ 0.02009600067138672,
373
+ 0.020151296615600587,
374
+ 0.020134912490844727,
375
+ 0.020082687377929686,
376
+ 0.02028441619873047,
377
+ 0.0200949764251709,
378
+ 0.020256767272949217,
379
+ 0.020212736129760742,
380
+ 0.020101119995117187,
381
+ 0.020290559768676757,
382
+ 0.02007040023803711,
383
+ 0.020220928192138672,
384
+ 0.020146175384521483,
385
+ 0.02008166313171387,
386
+ 0.020048896789550782,
387
+ 0.02022707176208496,
388
+ 0.02007040023803711,
389
+ 0.020051967620849608,
390
+ 0.020116479873657226,
391
+ 0.020199424743652345,
392
+ 0.02027827262878418,
393
+ 0.02026393508911133,
394
+ 0.02008780860900879,
395
+ 0.020238336563110353,
396
+ 0.020141056060791016,
397
+ 0.020179967880249023,
398
+ 0.020307968139648438,
399
+ 0.02031001663208008,
400
+ 0.020113407135009767,
401
+ 0.020239360809326173,
402
+ 0.020121599197387697,
403
+ 0.020109312057495117,
404
+ 0.020303871154785155,
405
+ 0.020205568313598633,
406
+ 0.02019430351257324,
407
+ 0.020151296615600587,
408
+ 0.02002739143371582,
409
+ 0.020100095748901366,
410
+ 0.020427776336669923,
411
+ 0.020116479873657226,
412
+ 0.020324352264404297,
413
+ 0.020215808868408205,
414
+ 0.020098047256469728,
415
+ 0.020413440704345705,
416
+ 0.02007961654663086,
417
+ 0.020098047256469728,
418
+ 0.02004582405090332,
419
+ 0.020098047256469728,
420
+ 0.02022400093078613,
421
+ 0.020329471588134765,
422
+ 0.02007859230041504,
423
+ 0.020178943634033202,
424
+ 0.020110336303710938,
425
+ 0.02011955261230469,
426
+ 0.02043289566040039,
427
+ 0.02010419273376465,
428
+ 0.020065280914306642,
429
+ 0.02005606460571289,
430
+ 0.02007347106933594,
431
+ 0.02025881576538086,
432
+ 0.02003865623474121
433
+ ]
434
+ },
435
+ "throughput": {
436
+ "unit": "tokens/s",
437
+ "value": 49.29082558285231
438
+ },
439
+ "energy": null,
440
+ "efficiency": null
441
+ }
442
+ }
443
+ }
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=False/benchmark.log ADDED
File without changes
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=False/benchmark_config.json ADDED
@@ -0,0 +1,102 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "pytorch_generate",
3
+ "backend": {
4
+ "name": "pytorch",
5
+ "version": "2.3.0+cu121",
6
+ "_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
7
+ "task": "text-generation",
8
+ "library": "transformers",
9
+ "model": "google/gemma-2b",
10
+ "processor": "google/gemma-2b",
11
+ "device": "cuda",
12
+ "device_ids": "0",
13
+ "seed": 42,
14
+ "inter_op_num_threads": null,
15
+ "intra_op_num_threads": null,
16
+ "model_kwargs": {},
17
+ "processor_kwargs": {},
18
+ "hub_kwargs": {},
19
+ "no_weights": true,
20
+ "device_map": null,
21
+ "torch_dtype": "float16",
22
+ "eval_mode": true,
23
+ "to_bettertransformer": false,
24
+ "low_cpu_mem_usage": null,
25
+ "attn_implementation": null,
26
+ "cache_implementation": "static",
27
+ "autocast_enabled": false,
28
+ "autocast_dtype": null,
29
+ "torch_compile": false,
30
+ "torch_compile_target": "forward",
31
+ "torch_compile_config": {
32
+ "backend": "inductor",
33
+ "mode": "reduce-overhead",
34
+ "fullgraph": true
35
+ },
36
+ "quantization_scheme": null,
37
+ "quantization_config": {},
38
+ "deepspeed_inference": false,
39
+ "deepspeed_inference_config": {},
40
+ "peft_type": null,
41
+ "peft_config": {}
42
+ },
43
+ "scenario": {
44
+ "name": "inference",
45
+ "_target_": "optimum_benchmark.scenarios.inference.scenario.InferenceScenario",
46
+ "iterations": 2,
47
+ "duration": 0,
48
+ "warmup_runs": 10,
49
+ "input_shapes": {
50
+ "batch_size": 1,
51
+ "num_choices": 2,
52
+ "sequence_length": 7
53
+ },
54
+ "new_tokens": null,
55
+ "latency": true,
56
+ "memory": true,
57
+ "energy": false,
58
+ "forward_kwargs": {},
59
+ "generate_kwargs": {
60
+ "max_new_tokens": 128,
61
+ "min_new_tokens": 128,
62
+ "do_sample": false
63
+ },
64
+ "call_kwargs": {}
65
+ },
66
+ "launcher": {
67
+ "name": "process",
68
+ "_target_": "optimum_benchmark.launchers.process.launcher.ProcessLauncher",
69
+ "device_isolation": true,
70
+ "device_isolation_action": "warn",
71
+ "start_method": "spawn"
72
+ },
73
+ "environment": {
74
+ "cpu": " AMD EPYC 7R32",
75
+ "cpu_count": 16,
76
+ "cpu_ram_mb": 66697.29792,
77
+ "system": "Linux",
78
+ "machine": "x86_64",
79
+ "platform": "Linux-5.10.217-205.860.amzn2.x86_64-x86_64-with-glibc2.29",
80
+ "processor": "x86_64",
81
+ "python_version": "3.8.10",
82
+ "gpu": [
83
+ "NVIDIA A10G"
84
+ ],
85
+ "gpu_count": 1,
86
+ "gpu_vram_mb": 24146608128,
87
+ "optimum_benchmark_version": "0.2.1",
88
+ "optimum_benchmark_commit": null,
89
+ "transformers_version": "4.42.0.dev0",
90
+ "transformers_commit": "bd5091df8db7cea1a9f94f797fc11487f840ade1",
91
+ "accelerate_version": "0.31.0.dev0",
92
+ "accelerate_commit": null,
93
+ "diffusers_version": null,
94
+ "diffusers_commit": null,
95
+ "optimum_version": "1.21.0.dev0",
96
+ "optimum_commit": null,
97
+ "timm_version": "0.9.16",
98
+ "timm_commit": null,
99
+ "peft_version": "0.11.2.dev0",
100
+ "peft_commit": null
101
+ }
102
+ }
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=False/benchmark_report.json ADDED
@@ -0,0 +1,339 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prefill": {
3
+ "memory": {
4
+ "unit": "MB",
5
+ "max_ram": 1777.160192,
6
+ "max_global_vram": 6793.199616,
7
+ "max_process_vram": 0.0,
8
+ "max_reserved": 6146.752512,
9
+ "max_allocated": 5036.936192
10
+ },
11
+ "latency": {
12
+ "unit": "s",
13
+ "count": 2,
14
+ "total": 0.04430672073364258,
15
+ "mean": 0.02215336036682129,
16
+ "stdev": 0.00032209587097167813,
17
+ "p50": 0.02215336036682129,
18
+ "p90": 0.02241103706359863,
19
+ "p95": 0.0224432466506958,
20
+ "p99": 0.022469014320373532,
21
+ "values": [
22
+ 0.022475456237792967,
23
+ 0.02183126449584961
24
+ ]
25
+ },
26
+ "throughput": {
27
+ "unit": "tokens/s",
28
+ "value": 315.97915097719357
29
+ },
30
+ "energy": null,
31
+ "efficiency": null
32
+ },
33
+ "decode": {
34
+ "memory": {
35
+ "unit": "MB",
36
+ "max_ram": 1777.242112,
37
+ "max_global_vram": 6793.199616,
38
+ "max_process_vram": 0.0,
39
+ "max_reserved": 6146.752512,
40
+ "max_allocated": 5036.936704
41
+ },
42
+ "latency": {
43
+ "unit": "s",
44
+ "count": 2,
45
+ "total": 5.13177685546875,
46
+ "mean": 2.565888427734375,
47
+ "stdev": 0.0033464355468748597,
48
+ "p50": 2.565888427734375,
49
+ "p90": 2.568565576171875,
50
+ "p95": 2.5689002197265625,
51
+ "p99": 2.5691679345703125,
52
+ "values": [
53
+ 2.56923486328125,
54
+ 2.5625419921875
55
+ ]
56
+ },
57
+ "throughput": {
58
+ "unit": "tokens/s",
59
+ "value": 49.49552701796091
60
+ },
61
+ "energy": null,
62
+ "efficiency": null
63
+ },
64
+ "per_token": {
65
+ "memory": null,
66
+ "latency": {
67
+ "unit": "s",
68
+ "count": 253,
69
+ "total": 5.132801023483274,
70
+ "mean": 0.020287751080961567,
71
+ "stdev": 0.001393544041619272,
72
+ "p50": 0.020134912490844727,
73
+ "p90": 0.020422451019287108,
74
+ "p95": 0.020565196228027344,
75
+ "p99": 0.021351505813598633,
76
+ "values": [
77
+ 0.02208460807800293,
78
+ 0.020320255279541014,
79
+ 0.020093952178955078,
80
+ 0.020084735870361328,
81
+ 0.02020147132873535,
82
+ 0.020163583755493163,
83
+ 0.02011238479614258,
84
+ 0.02004582405090332,
85
+ 0.020247552871704103,
86
+ 0.020156415939331054,
87
+ 0.02044313621520996,
88
+ 0.020108287811279296,
89
+ 0.020092927932739257,
90
+ 0.02005606460571289,
91
+ 0.020098047256469728,
92
+ 0.02002943992614746,
93
+ 0.019914751052856446,
94
+ 0.02127667236328125,
95
+ 0.02065510368347168,
96
+ 0.020178943634033202,
97
+ 0.02022400093078613,
98
+ 0.020082687377929686,
99
+ 0.02026803207397461,
100
+ 0.020188159942626953,
101
+ 0.02007347106933594,
102
+ 0.020090879440307616,
103
+ 0.020247552871704103,
104
+ 0.02004787254333496,
105
+ 0.020208639144897463,
106
+ 0.020011007308959963,
107
+ 0.02021990394592285,
108
+ 0.02004582405090332,
109
+ 0.020148223876953125,
110
+ 0.020282367706298828,
111
+ 0.02017791938781738,
112
+ 0.020230144500732423,
113
+ 0.020215808868408205,
114
+ 0.02001817512512207,
115
+ 0.02002943992614746,
116
+ 0.020006912231445313,
117
+ 0.020114431381225584,
118
+ 0.020328447341918944,
119
+ 0.02007756805419922,
120
+ 0.02001408004760742,
121
+ 0.02009600067138672,
122
+ 0.0200447998046875,
123
+ 0.0202106876373291,
124
+ 0.020132863998413086,
125
+ 0.02045142364501953,
126
+ 0.020527008056640626,
127
+ 0.02075449562072754,
128
+ 0.02032736015319824,
129
+ 0.020100095748901366,
130
+ 0.02008883285522461,
131
+ 0.020238336563110353,
132
+ 0.020023296356201172,
133
+ 0.02011955261230469,
134
+ 0.020082687377929686,
135
+ 0.020118528366088868,
136
+ 0.02026188850402832,
137
+ 0.0202926082611084,
138
+ 0.020195327758789062,
139
+ 0.02024550437927246,
140
+ 0.020101119995117187,
141
+ 0.02012876892089844,
142
+ 0.020552703857421875,
143
+ 0.02035916709899902,
144
+ 0.021029888153076173,
145
+ 0.021415935516357423,
146
+ 0.02057318305969238,
147
+ 0.020413440704345705,
148
+ 0.02029465675354004,
149
+ 0.020117504119873047,
150
+ 0.020101119995117187,
151
+ 0.02011136054992676,
152
+ 0.020124671936035156,
153
+ 0.020182016372680665,
154
+ 0.02005401611328125,
155
+ 0.02009600067138672,
156
+ 0.020463615417480468,
157
+ 0.02049945640563965,
158
+ 0.020676607131958007,
159
+ 0.020747264862060546,
160
+ 0.020136959075927736,
161
+ 0.020125696182250977,
162
+ 0.020189184188842774,
163
+ 0.020157440185546875,
164
+ 0.020246528625488282,
165
+ 0.02007859230041504,
166
+ 0.020248575210571287,
167
+ 0.020113407135009767,
168
+ 0.020158464431762696,
169
+ 0.02008166313171387,
170
+ 0.02007859230041504,
171
+ 0.02025369644165039,
172
+ 0.020102144241333008,
173
+ 0.020044832229614257,
174
+ 0.0200928955078125,
175
+ 0.02020966339111328,
176
+ 0.020185087203979494,
177
+ 0.020126720428466797,
178
+ 0.02004582405090332,
179
+ 0.020067327499389647,
180
+ 0.02021785545349121,
181
+ 0.020125696182250977,
182
+ 0.020230144500732423,
183
+ 0.020068351745605468,
184
+ 0.020135936737060548,
185
+ 0.020102144241333008,
186
+ 0.020121599197387697,
187
+ 0.02026393508911133,
188
+ 0.020306943893432617,
189
+ 0.02016972732543945,
190
+ 0.020184064865112306,
191
+ 0.020135936737060548,
192
+ 0.020093952178955078,
193
+ 0.02022502326965332,
194
+ 0.020212736129760742,
195
+ 0.020089920043945313,
196
+ 0.02016966438293457,
197
+ 0.02012057685852051,
198
+ 0.02008576011657715,
199
+ 0.02023628807067871,
200
+ 0.020170751571655272,
201
+ 0.02004787254333496,
202
+ 0.02006425666809082,
203
+ 0.042106880187988284,
204
+ 0.02042470359802246,
205
+ 0.02011136054992676,
206
+ 0.020142080307006836,
207
+ 0.020065280914306642,
208
+ 0.020092927932739257,
209
+ 0.02023219108581543,
210
+ 0.020230144500732423,
211
+ 0.020076543807983398,
212
+ 0.020099071502685546,
213
+ 0.02003455924987793,
214
+ 0.020039680480957032,
215
+ 0.020361215591430663,
216
+ 0.020092927932739257,
217
+ 0.020093952178955078,
218
+ 0.02008678436279297,
219
+ 0.02000383949279785,
220
+ 0.02024140739440918,
221
+ 0.02020249557495117,
222
+ 0.02003558349609375,
223
+ 0.020098047256469728,
224
+ 0.020109312057495117,
225
+ 0.020063232421875,
226
+ 0.020264959335327147,
227
+ 0.01985740852355957,
228
+ 0.01984000015258789,
229
+ 0.019978239059448243,
230
+ 0.021292032241821288,
231
+ 0.020559871673583984,
232
+ 0.020134912490844727,
233
+ 0.02007142448425293,
234
+ 0.020098047256469728,
235
+ 0.02023321533203125,
236
+ 0.02023321533203125,
237
+ 0.020140031814575195,
238
+ 0.02007756805419922,
239
+ 0.02006630325317383,
240
+ 0.020076608657836913,
241
+ 0.02018911933898926,
242
+ 0.020315135955810547,
243
+ 0.0206059513092041,
244
+ 0.020067327499389647,
245
+ 0.019916799545288084,
246
+ 0.020093952178955078,
247
+ 0.020366336822509764,
248
+ 0.020518911361694335,
249
+ 0.020180992126464844,
250
+ 0.020140031814575195,
251
+ 0.02011238479614258,
252
+ 0.02008883285522461,
253
+ 0.02023526382446289,
254
+ 0.02010419273376465,
255
+ 0.020180992126464844,
256
+ 0.02006937599182129,
257
+ 0.020136959075927736,
258
+ 0.020099071502685546,
259
+ 0.020129791259765627,
260
+ 0.020247552871704103,
261
+ 0.020189184188842774,
262
+ 0.02006118392944336,
263
+ 0.02011238479614258,
264
+ 0.020131839752197265,
265
+ 0.020238336563110353,
266
+ 0.020508672714233397,
267
+ 0.02049228858947754,
268
+ 0.020607999801635742,
269
+ 0.02009600067138672,
270
+ 0.020151296615600587,
271
+ 0.020134912490844727,
272
+ 0.020082687377929686,
273
+ 0.02028441619873047,
274
+ 0.0200949764251709,
275
+ 0.020256767272949217,
276
+ 0.020212736129760742,
277
+ 0.020101119995117187,
278
+ 0.020290559768676757,
279
+ 0.02007040023803711,
280
+ 0.020220928192138672,
281
+ 0.020146175384521483,
282
+ 0.02008166313171387,
283
+ 0.020048896789550782,
284
+ 0.02022707176208496,
285
+ 0.02007040023803711,
286
+ 0.020051967620849608,
287
+ 0.020116479873657226,
288
+ 0.020199424743652345,
289
+ 0.02027827262878418,
290
+ 0.02026393508911133,
291
+ 0.02008780860900879,
292
+ 0.020238336563110353,
293
+ 0.020141056060791016,
294
+ 0.020179967880249023,
295
+ 0.020307968139648438,
296
+ 0.02031001663208008,
297
+ 0.020113407135009767,
298
+ 0.020239360809326173,
299
+ 0.020121599197387697,
300
+ 0.020109312057495117,
301
+ 0.020303871154785155,
302
+ 0.020205568313598633,
303
+ 0.02019430351257324,
304
+ 0.020151296615600587,
305
+ 0.02002739143371582,
306
+ 0.020100095748901366,
307
+ 0.020427776336669923,
308
+ 0.020116479873657226,
309
+ 0.020324352264404297,
310
+ 0.020215808868408205,
311
+ 0.020098047256469728,
312
+ 0.020413440704345705,
313
+ 0.02007961654663086,
314
+ 0.020098047256469728,
315
+ 0.02004582405090332,
316
+ 0.020098047256469728,
317
+ 0.02022400093078613,
318
+ 0.020329471588134765,
319
+ 0.02007859230041504,
320
+ 0.020178943634033202,
321
+ 0.020110336303710938,
322
+ 0.02011955261230469,
323
+ 0.02043289566040039,
324
+ 0.02010419273376465,
325
+ 0.020065280914306642,
326
+ 0.02005606460571289,
327
+ 0.02007347106933594,
328
+ 0.02025881576538086,
329
+ 0.02003865623474121
330
+ ]
331
+ },
332
+ "throughput": {
333
+ "unit": "tokens/s",
334
+ "value": 49.29082558285231
335
+ },
336
+ "energy": null,
337
+ "efficiency": null
338
+ }
339
+ }
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=False/summary.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model": "google/gemma-2b",
3
+ "commit": "bd5091df8db7cea1a9f94f797fc11487f840ade1",
4
+ "config": "backend.cache_implementation=static,backend.torch_compile=False",
5
+ "metrics": {
6
+ "prefill.latency.mean": 0.02215336036682129,
7
+ "prefill.throughput.value": 315.97915097719357,
8
+ "decode.latency.mean": 2.565888427734375,
9
+ "decode.throughput.value": 49.49552701796091,
10
+ "per_token.latency.mean": 0.020287751080961567,
11
+ "per_token.throughput.value": 49.29082558285231
12
+ }
13
+ }
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=True/.hydra/config.yaml ADDED
@@ -0,0 +1,91 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ name: pytorch_generate
2
+ backend:
3
+ name: pytorch
4
+ version: 2.3.0+cu121
5
+ _target_: optimum_benchmark.backends.pytorch.backend.PyTorchBackend
6
+ task: null
7
+ library: null
8
+ model: google/gemma-2b
9
+ processor: null
10
+ device: cuda
11
+ device_ids: '0'
12
+ seed: 42
13
+ inter_op_num_threads: null
14
+ intra_op_num_threads: null
15
+ model_kwargs: {}
16
+ processor_kwargs: {}
17
+ hub_kwargs: {}
18
+ no_weights: true
19
+ device_map: null
20
+ torch_dtype: float16
21
+ eval_mode: true
22
+ to_bettertransformer: false
23
+ low_cpu_mem_usage: null
24
+ attn_implementation: null
25
+ cache_implementation: static
26
+ autocast_enabled: false
27
+ autocast_dtype: null
28
+ torch_compile: true
29
+ torch_compile_target: forward
30
+ torch_compile_config:
31
+ backend: inductor
32
+ mode: reduce-overhead
33
+ fullgraph: true
34
+ quantization_scheme: null
35
+ quantization_config: {}
36
+ deepspeed_inference: false
37
+ deepspeed_inference_config: {}
38
+ peft_type: null
39
+ peft_config: {}
40
+ scenario:
41
+ name: inference
42
+ _target_: optimum_benchmark.scenarios.inference.scenario.InferenceScenario
43
+ iterations: 2
44
+ duration: 0
45
+ warmup_runs: 10
46
+ input_shapes:
47
+ batch_size: 1
48
+ sequence_length: 7
49
+ new_tokens: null
50
+ latency: true
51
+ memory: true
52
+ energy: false
53
+ forward_kwargs: {}
54
+ generate_kwargs:
55
+ max_new_tokens: 128
56
+ min_new_tokens: 128
57
+ do_sample: false
58
+ call_kwargs: {}
59
+ launcher:
60
+ name: process
61
+ _target_: optimum_benchmark.launchers.process.launcher.ProcessLauncher
62
+ device_isolation: true
63
+ device_isolation_action: warn
64
+ start_method: spawn
65
+ environment:
66
+ cpu: ' AMD EPYC 7R32'
67
+ cpu_count: 16
68
+ cpu_ram_mb: 66697.29792
69
+ system: Linux
70
+ machine: x86_64
71
+ platform: Linux-5.10.217-205.860.amzn2.x86_64-x86_64-with-glibc2.29
72
+ processor: x86_64
73
+ python_version: 3.8.10
74
+ gpu:
75
+ - NVIDIA A10G
76
+ gpu_count: 1
77
+ gpu_vram_mb: 24146608128
78
+ optimum_benchmark_version: 0.2.1
79
+ optimum_benchmark_commit: null
80
+ transformers_version: 4.42.0.dev0
81
+ transformers_commit: bd5091df8db7cea1a9f94f797fc11487f840ade1
82
+ accelerate_version: 0.31.0.dev0
83
+ accelerate_commit: null
84
+ diffusers_version: null
85
+ diffusers_commit: null
86
+ optimum_version: 1.21.0.dev0
87
+ optimum_commit: null
88
+ timm_version: 0.9.16
89
+ timm_commit: null
90
+ peft_version: 0.11.2.dev0
91
+ peft_commit: null
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=True/.hydra/hydra.yaml ADDED
@@ -0,0 +1,141 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ hydra:
2
+ run:
3
+ dir: _benchmark/tmpjivn_1kh/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/${hydra.job.override_dirname}
4
+ sweep:
5
+ dir: _benchmark/tmpjivn_1kh/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1
6
+ subdir: ${hydra.job.override_dirname}
7
+ launcher:
8
+ _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
9
+ sweeper:
10
+ _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
11
+ max_batch_size: null
12
+ params: null
13
+ help:
14
+ app_name: ${hydra.job.name}
15
+ header: '${hydra.help.app_name} is powered by Hydra.
16
+
17
+ '
18
+ footer: 'Powered by Hydra (https://hydra.cc)
19
+
20
+ Use --hydra-help to view Hydra specific help
21
+
22
+ '
23
+ template: '${hydra.help.header}
24
+
25
+ == Configuration groups ==
26
+
27
+ Compose your configuration from those groups (group=option)
28
+
29
+
30
+ $APP_CONFIG_GROUPS
31
+
32
+
33
+ == Config ==
34
+
35
+ Override anything in the config (foo.bar=value)
36
+
37
+
38
+ $CONFIG
39
+
40
+
41
+ ${hydra.help.footer}
42
+
43
+ '
44
+ hydra_help:
45
+ template: 'Hydra (${hydra.runtime.version})
46
+
47
+ See https://hydra.cc for more info.
48
+
49
+
50
+ == Flags ==
51
+
52
+ $FLAGS_HELP
53
+
54
+
55
+ == Configuration groups ==
56
+
57
+ Compose your configuration from those groups (For example, append hydra/job_logging=disabled
58
+ to command line)
59
+
60
+
61
+ $HYDRA_CONFIG_GROUPS
62
+
63
+
64
+ Use ''--cfg hydra'' to Show the Hydra config.
65
+
66
+ '
67
+ hydra_help: ???
68
+ hydra_logging:
69
+ version: 1
70
+ root:
71
+ level: ERROR
72
+ disable_existing_loggers: true
73
+ job_logging:
74
+ version: 1
75
+ root:
76
+ level: ERROR
77
+ disable_existing_loggers: true
78
+ env: {}
79
+ mode: MULTIRUN
80
+ searchpath: []
81
+ callbacks: {}
82
+ output_subdir: .hydra
83
+ overrides:
84
+ hydra:
85
+ - hydra/job_logging=disabled
86
+ - hydra/hydra_logging=disabled
87
+ - hydra.sweep.dir=_benchmark/tmpjivn_1kh/commit\=bd5091df8db7cea1a9f94f797fc11487f840ade1
88
+ - hydra.run.dir=_benchmark/tmpjivn_1kh/commit\=bd5091df8db7cea1a9f94f797fc11487f840ade1/${hydra.job.override_dirname}
89
+ - hydra.mode=MULTIRUN
90
+ task:
91
+ - backend.model=google/gemma-2b
92
+ - backend.cache_implementation=static
93
+ - backend.torch_compile=True
94
+ job:
95
+ name: cli
96
+ chdir: true
97
+ override_dirname: backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=True
98
+ id: '3'
99
+ num: 3
100
+ config_name: generation
101
+ env_set:
102
+ OVERRIDE_BENCHMARKS: '1'
103
+ LOG_LEVEL: WARN
104
+ env_copy: []
105
+ config:
106
+ override_dirname:
107
+ kv_sep: '='
108
+ item_sep: ','
109
+ exclude_keys: []
110
+ runtime:
111
+ version: 1.3.2
112
+ version_base: '1.3'
113
+ cwd: /transformers
114
+ config_sources:
115
+ - path: hydra.conf
116
+ schema: pkg
117
+ provider: hydra
118
+ - path: optimum_benchmark
119
+ schema: pkg
120
+ provider: main
121
+ - path: /transformers/benchmark/config
122
+ schema: file
123
+ provider: command-line
124
+ - path: ''
125
+ schema: structured
126
+ provider: schema
127
+ output_dir: /transformers/_benchmark/tmpjivn_1kh/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=True
128
+ choices:
129
+ backend: pytorch
130
+ launcher: process
131
+ scenario: inference
132
+ hydra/env: default
133
+ hydra/callbacks: null
134
+ hydra/job_logging: disabled
135
+ hydra/hydra_logging: disabled
136
+ hydra/hydra_help: default
137
+ hydra/help: default
138
+ hydra/sweeper: basic
139
+ hydra/launcher: basic
140
+ hydra/output: default
141
+ verbose: false
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=True/.hydra/overrides.yaml ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ - backend.model=google/gemma-2b
2
+ - backend.cache_implementation=static
3
+ - backend.torch_compile=True
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=True/benchmark.json ADDED
@@ -0,0 +1,443 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "config": {
3
+ "name": "pytorch_generate",
4
+ "backend": {
5
+ "name": "pytorch",
6
+ "version": "2.3.0+cu121",
7
+ "_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
8
+ "task": "text-generation",
9
+ "library": "transformers",
10
+ "model": "google/gemma-2b",
11
+ "processor": "google/gemma-2b",
12
+ "device": "cuda",
13
+ "device_ids": "0",
14
+ "seed": 42,
15
+ "inter_op_num_threads": null,
16
+ "intra_op_num_threads": null,
17
+ "model_kwargs": {},
18
+ "processor_kwargs": {},
19
+ "hub_kwargs": {},
20
+ "no_weights": true,
21
+ "device_map": null,
22
+ "torch_dtype": "float16",
23
+ "eval_mode": true,
24
+ "to_bettertransformer": false,
25
+ "low_cpu_mem_usage": null,
26
+ "attn_implementation": null,
27
+ "cache_implementation": "static",
28
+ "autocast_enabled": false,
29
+ "autocast_dtype": null,
30
+ "torch_compile": true,
31
+ "torch_compile_target": "forward",
32
+ "torch_compile_config": {
33
+ "backend": "inductor",
34
+ "mode": "reduce-overhead",
35
+ "fullgraph": true
36
+ },
37
+ "quantization_scheme": null,
38
+ "quantization_config": {},
39
+ "deepspeed_inference": false,
40
+ "deepspeed_inference_config": {},
41
+ "peft_type": null,
42
+ "peft_config": {}
43
+ },
44
+ "scenario": {
45
+ "name": "inference",
46
+ "_target_": "optimum_benchmark.scenarios.inference.scenario.InferenceScenario",
47
+ "iterations": 2,
48
+ "duration": 0,
49
+ "warmup_runs": 10,
50
+ "input_shapes": {
51
+ "batch_size": 1,
52
+ "num_choices": 2,
53
+ "sequence_length": 7
54
+ },
55
+ "new_tokens": null,
56
+ "latency": true,
57
+ "memory": true,
58
+ "energy": false,
59
+ "forward_kwargs": {},
60
+ "generate_kwargs": {
61
+ "max_new_tokens": 128,
62
+ "min_new_tokens": 128,
63
+ "do_sample": false
64
+ },
65
+ "call_kwargs": {}
66
+ },
67
+ "launcher": {
68
+ "name": "process",
69
+ "_target_": "optimum_benchmark.launchers.process.launcher.ProcessLauncher",
70
+ "device_isolation": true,
71
+ "device_isolation_action": "warn",
72
+ "start_method": "spawn"
73
+ },
74
+ "environment": {
75
+ "cpu": " AMD EPYC 7R32",
76
+ "cpu_count": 16,
77
+ "cpu_ram_mb": 66697.29792,
78
+ "system": "Linux",
79
+ "machine": "x86_64",
80
+ "platform": "Linux-5.10.217-205.860.amzn2.x86_64-x86_64-with-glibc2.29",
81
+ "processor": "x86_64",
82
+ "python_version": "3.8.10",
83
+ "gpu": [
84
+ "NVIDIA A10G"
85
+ ],
86
+ "gpu_count": 1,
87
+ "gpu_vram_mb": 24146608128,
88
+ "optimum_benchmark_version": "0.2.1",
89
+ "optimum_benchmark_commit": null,
90
+ "transformers_version": "4.42.0.dev0",
91
+ "transformers_commit": "bd5091df8db7cea1a9f94f797fc11487f840ade1",
92
+ "accelerate_version": "0.31.0.dev0",
93
+ "accelerate_commit": null,
94
+ "diffusers_version": null,
95
+ "diffusers_commit": null,
96
+ "optimum_version": "1.21.0.dev0",
97
+ "optimum_commit": null,
98
+ "timm_version": "0.9.16",
99
+ "timm_commit": null,
100
+ "peft_version": "0.11.2.dev0",
101
+ "peft_commit": null
102
+ }
103
+ },
104
+ "report": {
105
+ "prefill": {
106
+ "memory": {
107
+ "unit": "MB",
108
+ "max_ram": 1929.637888,
109
+ "max_global_vram": 6094.848,
110
+ "max_process_vram": 0.0,
111
+ "max_reserved": 5379.19488,
112
+ "max_allocated": 5021.251584
113
+ },
114
+ "latency": {
115
+ "unit": "s",
116
+ "count": 2,
117
+ "total": 0.028288032531738283,
118
+ "mean": 0.014144016265869142,
119
+ "stdev": 0.0002579679489135749,
120
+ "p50": 0.014144016265869142,
121
+ "p90": 0.014350390625,
122
+ "p95": 0.014376187419891358,
123
+ "p99": 0.014396824855804444,
124
+ "values": [
125
+ 0.014401984214782716,
126
+ 0.013886048316955566
127
+ ]
128
+ },
129
+ "throughput": {
130
+ "unit": "tokens/s",
131
+ "value": 494.9089331077529
132
+ },
133
+ "energy": null,
134
+ "efficiency": null
135
+ },
136
+ "decode": {
137
+ "memory": {
138
+ "unit": "MB",
139
+ "max_ram": 2273.52576,
140
+ "max_global_vram": 8907.128832,
141
+ "max_process_vram": 0.0,
142
+ "max_reserved": 5379.19488,
143
+ "max_allocated": 5027.089408
144
+ },
145
+ "latency": {
146
+ "unit": "s",
147
+ "count": 2,
148
+ "total": 3.147330322265625,
149
+ "mean": 1.5736651611328125,
150
+ "stdev": 0.015655151367187425,
151
+ "p50": 1.5736651611328125,
152
+ "p90": 1.5861892822265624,
153
+ "p95": 1.5877547973632813,
154
+ "p99": 1.5890072094726562,
155
+ "values": [
156
+ 1.5893203125,
157
+ 1.558010009765625
158
+ ]
159
+ },
160
+ "throughput": {
161
+ "unit": "tokens/s",
162
+ "value": 80.70331804802635
163
+ },
164
+ "energy": null,
165
+ "efficiency": null
166
+ },
167
+ "per_token": {
168
+ "memory": null,
169
+ "latency": {
170
+ "unit": "s",
171
+ "count": 253,
172
+ "total": 3.1482787828445424,
173
+ "mean": 0.012443789655512029,
174
+ "stdev": 0.0008989893698007158,
175
+ "p50": 0.012278783798217773,
176
+ "p90": 0.012717465209960937,
177
+ "p95": 0.012741613006591797,
178
+ "p99": 0.012834898147583007,
179
+ "values": [
180
+ 0.012732416152954102,
181
+ 0.012757023811340333,
182
+ 0.012741600036621094,
183
+ 0.012751872062683106,
184
+ 0.012681216239929198,
185
+ 0.012776448249816894,
186
+ 0.012670975685119629,
187
+ 0.01276518440246582,
188
+ 0.012812288284301757,
189
+ 0.012741632461547851,
190
+ 0.012660736083984376,
191
+ 0.012731391906738282,
192
+ 0.012659711837768555,
193
+ 0.012710911750793457,
194
+ 0.012644351959228516,
195
+ 0.012732416152954102,
196
+ 0.012710911750793457,
197
+ 0.012876799583435058,
198
+ 0.012690431594848632,
199
+ 0.012859392166137695,
200
+ 0.012667903900146485,
201
+ 0.012705792427062988,
202
+ 0.012703743934631348,
203
+ 0.012758015632629394,
204
+ 0.012675071716308594,
205
+ 0.012733440399169921,
206
+ 0.01266483211517334,
207
+ 0.012747776031494141,
208
+ 0.01266585636138916,
209
+ 0.012715007781982422,
210
+ 0.01266380786895752,
211
+ 0.012718079566955566,
212
+ 0.012651519775390625,
213
+ 0.012741632461547851,
214
+ 0.01263923168182373,
215
+ 0.012666879653930664,
216
+ 0.012670975685119629,
217
+ 0.012721152305603027,
218
+ 0.012645376205444337,
219
+ 0.012708864212036132,
220
+ 0.012627967834472656,
221
+ 0.012724224090576173,
222
+ 0.012677120208740235,
223
+ 0.01273855972290039,
224
+ 0.012798975944519043,
225
+ 0.012737536430358886,
226
+ 0.012675071716308594,
227
+ 0.012719103813171387,
228
+ 0.012675071716308594,
229
+ 0.012706815719604492,
230
+ 0.012681216239929198,
231
+ 0.012702719688415527,
232
+ 0.012676095962524414,
233
+ 0.012731391906738282,
234
+ 0.012732416152954102,
235
+ 0.012702719688415527,
236
+ 0.012670975685119629,
237
+ 0.012692480087280274,
238
+ 0.012662783622741699,
239
+ 0.012708864212036132,
240
+ 0.012659711837768555,
241
+ 0.012698623657226562,
242
+ 0.012698623657226562,
243
+ 0.012686335563659668,
244
+ 0.01253171157836914,
245
+ 0.012560383796691894,
246
+ 0.01265664005279541,
247
+ 0.012442624092102051,
248
+ 0.012438528060913086,
249
+ 0.012468223571777343,
250
+ 0.012436479568481444,
251
+ 0.012686335563659668,
252
+ 0.012496895790100097,
253
+ 0.012310527801513671,
254
+ 0.012236800193786621,
255
+ 0.012264448165893555,
256
+ 0.01224396800994873,
257
+ 0.012279808044433594,
258
+ 0.01225113582611084,
259
+ 0.012270591735839843,
260
+ 0.012256256103515625,
261
+ 0.012316672325134278,
262
+ 0.012239871978759765,
263
+ 0.012272640228271485,
264
+ 0.01223475170135498,
265
+ 0.012286975860595703,
266
+ 0.012229632377624512,
267
+ 0.012279840469360351,
268
+ 0.012232671737670899,
269
+ 0.012278783798217773,
270
+ 0.012236800193786621,
271
+ 0.012274687767028808,
272
+ 0.01225830364227295,
273
+ 0.01229312038421631,
274
+ 0.012265472412109376,
275
+ 0.012281855583190919,
276
+ 0.012356608390808106,
277
+ 0.012272640228271485,
278
+ 0.012225536346435547,
279
+ 0.012272640228271485,
280
+ 0.012233728408813477,
281
+ 0.012285951614379884,
282
+ 0.012230655670166016,
283
+ 0.012296192169189453,
284
+ 0.012323840141296387,
285
+ 0.012272640228271485,
286
+ 0.012237824440002442,
287
+ 0.012315648078918457,
288
+ 0.012229632377624512,
289
+ 0.01226854419708252,
290
+ 0.012236800193786621,
291
+ 0.01229312038421631,
292
+ 0.012231712341308593,
293
+ 0.012270560264587402,
294
+ 0.012229632377624512,
295
+ 0.012291071891784668,
296
+ 0.0122357759475708,
297
+ 0.012266495704650878,
298
+ 0.012312576293945313,
299
+ 0.012302335739135742,
300
+ 0.012305407524108887,
301
+ 0.012307456016540527,
302
+ 0.012257280349731446,
303
+ 0.01227673625946045,
304
+ 0.012248064041137695,
305
+ 0.012303359985351562,
306
+ 0.026368000030517577,
307
+ 0.012238847732543945,
308
+ 0.012271615982055664,
309
+ 0.01222758388519287,
310
+ 0.012264448165893555,
311
+ 0.012244992256164551,
312
+ 0.012303359985351562,
313
+ 0.012361727714538574,
314
+ 0.012281855583190919,
315
+ 0.012233728408813477,
316
+ 0.012288000106811523,
317
+ 0.012244992256164551,
318
+ 0.012290047645568849,
319
+ 0.012237824440002442,
320
+ 0.012270591735839843,
321
+ 0.01222758388519287,
322
+ 0.012374015808105468,
323
+ 0.01224294376373291,
324
+ 0.012278783798217773,
325
+ 0.012229632377624512,
326
+ 0.012264448165893555,
327
+ 0.012253184318542481,
328
+ 0.012265472412109376,
329
+ 0.012232704162597656,
330
+ 0.01226854419708252,
331
+ 0.012241920471191407,
332
+ 0.012282879829406738,
333
+ 0.012229632377624512,
334
+ 0.012286975860595703,
335
+ 0.012239871978759765,
336
+ 0.012274687767028808,
337
+ 0.012228608131408691,
338
+ 0.01226956844329834,
339
+ 0.012249088287353516,
340
+ 0.012279808044433594,
341
+ 0.01224396800994873,
342
+ 0.012320768356323243,
343
+ 0.012247039794921874,
344
+ 0.01226956844329834,
345
+ 0.012225536346435547,
346
+ 0.012280832290649414,
347
+ 0.012247039794921874,
348
+ 0.012280832290649414,
349
+ 0.012228608131408691,
350
+ 0.012274687767028808,
351
+ 0.012231679916381836,
352
+ 0.012275712013244629,
353
+ 0.012230655670166016,
354
+ 0.012277759552001954,
355
+ 0.01232588768005371,
356
+ 0.012405759811401367,
357
+ 0.012254207611083985,
358
+ 0.012272671699523926,
359
+ 0.012240863800048829,
360
+ 0.012275744438171386,
361
+ 0.012239839553833008,
362
+ 0.012270591735839843,
363
+ 0.012247039794921874,
364
+ 0.012266495704650878,
365
+ 0.012338175773620605,
366
+ 0.012270591735839843,
367
+ 0.012236800193786621,
368
+ 0.012270591735839843,
369
+ 0.012246015548706055,
370
+ 0.012275712013244629,
371
+ 0.012233728408813477,
372
+ 0.012311552047729492,
373
+ 0.012336128234863282,
374
+ 0.012281855583190919,
375
+ 0.012224512100219726,
376
+ 0.01228492832183838,
377
+ 0.012246015548706055,
378
+ 0.012271615982055664,
379
+ 0.012238847732543945,
380
+ 0.012272640228271485,
381
+ 0.012229632377624512,
382
+ 0.012336128234863282,
383
+ 0.01237299156188965,
384
+ 0.012272640228271485,
385
+ 0.012230655670166016,
386
+ 0.012290047645568849,
387
+ 0.012247039794921874,
388
+ 0.012275712013244629,
389
+ 0.012232704162597656,
390
+ 0.012267519950866699,
391
+ 0.012236800193786621,
392
+ 0.012279808044433594,
393
+ 0.012236800193786621,
394
+ 0.012337183952331543,
395
+ 0.01222550392150879,
396
+ 0.012263423919677734,
397
+ 0.012240896224975586,
398
+ 0.012277759552001954,
399
+ 0.012232704162597656,
400
+ 0.012277759552001954,
401
+ 0.01221939182281494,
402
+ 0.012285951614379884,
403
+ 0.012320768356323243,
404
+ 0.012465151786804199,
405
+ 0.012257280349731446,
406
+ 0.012280832290649414,
407
+ 0.012223487854003906,
408
+ 0.01226854419708252,
409
+ 0.012222463607788087,
410
+ 0.012264448165893555,
411
+ 0.012221440315246582,
412
+ 0.012289024353027344,
413
+ 0.012254207611083985,
414
+ 0.012270591735839843,
415
+ 0.012237824440002442,
416
+ 0.012299263954162597,
417
+ 0.012254207611083985,
418
+ 0.012267519950866699,
419
+ 0.012224512100219726,
420
+ 0.012266495704650878,
421
+ 0.012237824440002442,
422
+ 0.012277759552001954,
423
+ 0.012228608131408691,
424
+ 0.012264448165893555,
425
+ 0.012228608131408691,
426
+ 0.012272640228271485,
427
+ 0.012225536346435547,
428
+ 0.012279808044433594,
429
+ 0.01222758388519287,
430
+ 0.012266495704650878,
431
+ 0.012237824440002442,
432
+ 0.012291071891784668
433
+ ]
434
+ },
435
+ "throughput": {
436
+ "unit": "tokens/s",
437
+ "value": 80.36137122882384
438
+ },
439
+ "energy": null,
440
+ "efficiency": null
441
+ }
442
+ }
443
+ }
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=True/benchmark.log ADDED
File without changes
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=True/benchmark_config.json ADDED
@@ -0,0 +1,102 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "pytorch_generate",
3
+ "backend": {
4
+ "name": "pytorch",
5
+ "version": "2.3.0+cu121",
6
+ "_target_": "optimum_benchmark.backends.pytorch.backend.PyTorchBackend",
7
+ "task": "text-generation",
8
+ "library": "transformers",
9
+ "model": "google/gemma-2b",
10
+ "processor": "google/gemma-2b",
11
+ "device": "cuda",
12
+ "device_ids": "0",
13
+ "seed": 42,
14
+ "inter_op_num_threads": null,
15
+ "intra_op_num_threads": null,
16
+ "model_kwargs": {},
17
+ "processor_kwargs": {},
18
+ "hub_kwargs": {},
19
+ "no_weights": true,
20
+ "device_map": null,
21
+ "torch_dtype": "float16",
22
+ "eval_mode": true,
23
+ "to_bettertransformer": false,
24
+ "low_cpu_mem_usage": null,
25
+ "attn_implementation": null,
26
+ "cache_implementation": "static",
27
+ "autocast_enabled": false,
28
+ "autocast_dtype": null,
29
+ "torch_compile": true,
30
+ "torch_compile_target": "forward",
31
+ "torch_compile_config": {
32
+ "backend": "inductor",
33
+ "mode": "reduce-overhead",
34
+ "fullgraph": true
35
+ },
36
+ "quantization_scheme": null,
37
+ "quantization_config": {},
38
+ "deepspeed_inference": false,
39
+ "deepspeed_inference_config": {},
40
+ "peft_type": null,
41
+ "peft_config": {}
42
+ },
43
+ "scenario": {
44
+ "name": "inference",
45
+ "_target_": "optimum_benchmark.scenarios.inference.scenario.InferenceScenario",
46
+ "iterations": 2,
47
+ "duration": 0,
48
+ "warmup_runs": 10,
49
+ "input_shapes": {
50
+ "batch_size": 1,
51
+ "num_choices": 2,
52
+ "sequence_length": 7
53
+ },
54
+ "new_tokens": null,
55
+ "latency": true,
56
+ "memory": true,
57
+ "energy": false,
58
+ "forward_kwargs": {},
59
+ "generate_kwargs": {
60
+ "max_new_tokens": 128,
61
+ "min_new_tokens": 128,
62
+ "do_sample": false
63
+ },
64
+ "call_kwargs": {}
65
+ },
66
+ "launcher": {
67
+ "name": "process",
68
+ "_target_": "optimum_benchmark.launchers.process.launcher.ProcessLauncher",
69
+ "device_isolation": true,
70
+ "device_isolation_action": "warn",
71
+ "start_method": "spawn"
72
+ },
73
+ "environment": {
74
+ "cpu": " AMD EPYC 7R32",
75
+ "cpu_count": 16,
76
+ "cpu_ram_mb": 66697.29792,
77
+ "system": "Linux",
78
+ "machine": "x86_64",
79
+ "platform": "Linux-5.10.217-205.860.amzn2.x86_64-x86_64-with-glibc2.29",
80
+ "processor": "x86_64",
81
+ "python_version": "3.8.10",
82
+ "gpu": [
83
+ "NVIDIA A10G"
84
+ ],
85
+ "gpu_count": 1,
86
+ "gpu_vram_mb": 24146608128,
87
+ "optimum_benchmark_version": "0.2.1",
88
+ "optimum_benchmark_commit": null,
89
+ "transformers_version": "4.42.0.dev0",
90
+ "transformers_commit": "bd5091df8db7cea1a9f94f797fc11487f840ade1",
91
+ "accelerate_version": "0.31.0.dev0",
92
+ "accelerate_commit": null,
93
+ "diffusers_version": null,
94
+ "diffusers_commit": null,
95
+ "optimum_version": "1.21.0.dev0",
96
+ "optimum_commit": null,
97
+ "timm_version": "0.9.16",
98
+ "timm_commit": null,
99
+ "peft_version": "0.11.2.dev0",
100
+ "peft_commit": null
101
+ }
102
+ }
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=True/benchmark_report.json ADDED
@@ -0,0 +1,339 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "prefill": {
3
+ "memory": {
4
+ "unit": "MB",
5
+ "max_ram": 1929.637888,
6
+ "max_global_vram": 6094.848,
7
+ "max_process_vram": 0.0,
8
+ "max_reserved": 5379.19488,
9
+ "max_allocated": 5021.251584
10
+ },
11
+ "latency": {
12
+ "unit": "s",
13
+ "count": 2,
14
+ "total": 0.028288032531738283,
15
+ "mean": 0.014144016265869142,
16
+ "stdev": 0.0002579679489135749,
17
+ "p50": 0.014144016265869142,
18
+ "p90": 0.014350390625,
19
+ "p95": 0.014376187419891358,
20
+ "p99": 0.014396824855804444,
21
+ "values": [
22
+ 0.014401984214782716,
23
+ 0.013886048316955566
24
+ ]
25
+ },
26
+ "throughput": {
27
+ "unit": "tokens/s",
28
+ "value": 494.9089331077529
29
+ },
30
+ "energy": null,
31
+ "efficiency": null
32
+ },
33
+ "decode": {
34
+ "memory": {
35
+ "unit": "MB",
36
+ "max_ram": 2273.52576,
37
+ "max_global_vram": 8907.128832,
38
+ "max_process_vram": 0.0,
39
+ "max_reserved": 5379.19488,
40
+ "max_allocated": 5027.089408
41
+ },
42
+ "latency": {
43
+ "unit": "s",
44
+ "count": 2,
45
+ "total": 3.147330322265625,
46
+ "mean": 1.5736651611328125,
47
+ "stdev": 0.015655151367187425,
48
+ "p50": 1.5736651611328125,
49
+ "p90": 1.5861892822265624,
50
+ "p95": 1.5877547973632813,
51
+ "p99": 1.5890072094726562,
52
+ "values": [
53
+ 1.5893203125,
54
+ 1.558010009765625
55
+ ]
56
+ },
57
+ "throughput": {
58
+ "unit": "tokens/s",
59
+ "value": 80.70331804802635
60
+ },
61
+ "energy": null,
62
+ "efficiency": null
63
+ },
64
+ "per_token": {
65
+ "memory": null,
66
+ "latency": {
67
+ "unit": "s",
68
+ "count": 253,
69
+ "total": 3.1482787828445424,
70
+ "mean": 0.012443789655512029,
71
+ "stdev": 0.0008989893698007158,
72
+ "p50": 0.012278783798217773,
73
+ "p90": 0.012717465209960937,
74
+ "p95": 0.012741613006591797,
75
+ "p99": 0.012834898147583007,
76
+ "values": [
77
+ 0.012732416152954102,
78
+ 0.012757023811340333,
79
+ 0.012741600036621094,
80
+ 0.012751872062683106,
81
+ 0.012681216239929198,
82
+ 0.012776448249816894,
83
+ 0.012670975685119629,
84
+ 0.01276518440246582,
85
+ 0.012812288284301757,
86
+ 0.012741632461547851,
87
+ 0.012660736083984376,
88
+ 0.012731391906738282,
89
+ 0.012659711837768555,
90
+ 0.012710911750793457,
91
+ 0.012644351959228516,
92
+ 0.012732416152954102,
93
+ 0.012710911750793457,
94
+ 0.012876799583435058,
95
+ 0.012690431594848632,
96
+ 0.012859392166137695,
97
+ 0.012667903900146485,
98
+ 0.012705792427062988,
99
+ 0.012703743934631348,
100
+ 0.012758015632629394,
101
+ 0.012675071716308594,
102
+ 0.012733440399169921,
103
+ 0.01266483211517334,
104
+ 0.012747776031494141,
105
+ 0.01266585636138916,
106
+ 0.012715007781982422,
107
+ 0.01266380786895752,
108
+ 0.012718079566955566,
109
+ 0.012651519775390625,
110
+ 0.012741632461547851,
111
+ 0.01263923168182373,
112
+ 0.012666879653930664,
113
+ 0.012670975685119629,
114
+ 0.012721152305603027,
115
+ 0.012645376205444337,
116
+ 0.012708864212036132,
117
+ 0.012627967834472656,
118
+ 0.012724224090576173,
119
+ 0.012677120208740235,
120
+ 0.01273855972290039,
121
+ 0.012798975944519043,
122
+ 0.012737536430358886,
123
+ 0.012675071716308594,
124
+ 0.012719103813171387,
125
+ 0.012675071716308594,
126
+ 0.012706815719604492,
127
+ 0.012681216239929198,
128
+ 0.012702719688415527,
129
+ 0.012676095962524414,
130
+ 0.012731391906738282,
131
+ 0.012732416152954102,
132
+ 0.012702719688415527,
133
+ 0.012670975685119629,
134
+ 0.012692480087280274,
135
+ 0.012662783622741699,
136
+ 0.012708864212036132,
137
+ 0.012659711837768555,
138
+ 0.012698623657226562,
139
+ 0.012698623657226562,
140
+ 0.012686335563659668,
141
+ 0.01253171157836914,
142
+ 0.012560383796691894,
143
+ 0.01265664005279541,
144
+ 0.012442624092102051,
145
+ 0.012438528060913086,
146
+ 0.012468223571777343,
147
+ 0.012436479568481444,
148
+ 0.012686335563659668,
149
+ 0.012496895790100097,
150
+ 0.012310527801513671,
151
+ 0.012236800193786621,
152
+ 0.012264448165893555,
153
+ 0.01224396800994873,
154
+ 0.012279808044433594,
155
+ 0.01225113582611084,
156
+ 0.012270591735839843,
157
+ 0.012256256103515625,
158
+ 0.012316672325134278,
159
+ 0.012239871978759765,
160
+ 0.012272640228271485,
161
+ 0.01223475170135498,
162
+ 0.012286975860595703,
163
+ 0.012229632377624512,
164
+ 0.012279840469360351,
165
+ 0.012232671737670899,
166
+ 0.012278783798217773,
167
+ 0.012236800193786621,
168
+ 0.012274687767028808,
169
+ 0.01225830364227295,
170
+ 0.01229312038421631,
171
+ 0.012265472412109376,
172
+ 0.012281855583190919,
173
+ 0.012356608390808106,
174
+ 0.012272640228271485,
175
+ 0.012225536346435547,
176
+ 0.012272640228271485,
177
+ 0.012233728408813477,
178
+ 0.012285951614379884,
179
+ 0.012230655670166016,
180
+ 0.012296192169189453,
181
+ 0.012323840141296387,
182
+ 0.012272640228271485,
183
+ 0.012237824440002442,
184
+ 0.012315648078918457,
185
+ 0.012229632377624512,
186
+ 0.01226854419708252,
187
+ 0.012236800193786621,
188
+ 0.01229312038421631,
189
+ 0.012231712341308593,
190
+ 0.012270560264587402,
191
+ 0.012229632377624512,
192
+ 0.012291071891784668,
193
+ 0.0122357759475708,
194
+ 0.012266495704650878,
195
+ 0.012312576293945313,
196
+ 0.012302335739135742,
197
+ 0.012305407524108887,
198
+ 0.012307456016540527,
199
+ 0.012257280349731446,
200
+ 0.01227673625946045,
201
+ 0.012248064041137695,
202
+ 0.012303359985351562,
203
+ 0.026368000030517577,
204
+ 0.012238847732543945,
205
+ 0.012271615982055664,
206
+ 0.01222758388519287,
207
+ 0.012264448165893555,
208
+ 0.012244992256164551,
209
+ 0.012303359985351562,
210
+ 0.012361727714538574,
211
+ 0.012281855583190919,
212
+ 0.012233728408813477,
213
+ 0.012288000106811523,
214
+ 0.012244992256164551,
215
+ 0.012290047645568849,
216
+ 0.012237824440002442,
217
+ 0.012270591735839843,
218
+ 0.01222758388519287,
219
+ 0.012374015808105468,
220
+ 0.01224294376373291,
221
+ 0.012278783798217773,
222
+ 0.012229632377624512,
223
+ 0.012264448165893555,
224
+ 0.012253184318542481,
225
+ 0.012265472412109376,
226
+ 0.012232704162597656,
227
+ 0.01226854419708252,
228
+ 0.012241920471191407,
229
+ 0.012282879829406738,
230
+ 0.012229632377624512,
231
+ 0.012286975860595703,
232
+ 0.012239871978759765,
233
+ 0.012274687767028808,
234
+ 0.012228608131408691,
235
+ 0.01226956844329834,
236
+ 0.012249088287353516,
237
+ 0.012279808044433594,
238
+ 0.01224396800994873,
239
+ 0.012320768356323243,
240
+ 0.012247039794921874,
241
+ 0.01226956844329834,
242
+ 0.012225536346435547,
243
+ 0.012280832290649414,
244
+ 0.012247039794921874,
245
+ 0.012280832290649414,
246
+ 0.012228608131408691,
247
+ 0.012274687767028808,
248
+ 0.012231679916381836,
249
+ 0.012275712013244629,
250
+ 0.012230655670166016,
251
+ 0.012277759552001954,
252
+ 0.01232588768005371,
253
+ 0.012405759811401367,
254
+ 0.012254207611083985,
255
+ 0.012272671699523926,
256
+ 0.012240863800048829,
257
+ 0.012275744438171386,
258
+ 0.012239839553833008,
259
+ 0.012270591735839843,
260
+ 0.012247039794921874,
261
+ 0.012266495704650878,
262
+ 0.012338175773620605,
263
+ 0.012270591735839843,
264
+ 0.012236800193786621,
265
+ 0.012270591735839843,
266
+ 0.012246015548706055,
267
+ 0.012275712013244629,
268
+ 0.012233728408813477,
269
+ 0.012311552047729492,
270
+ 0.012336128234863282,
271
+ 0.012281855583190919,
272
+ 0.012224512100219726,
273
+ 0.01228492832183838,
274
+ 0.012246015548706055,
275
+ 0.012271615982055664,
276
+ 0.012238847732543945,
277
+ 0.012272640228271485,
278
+ 0.012229632377624512,
279
+ 0.012336128234863282,
280
+ 0.01237299156188965,
281
+ 0.012272640228271485,
282
+ 0.012230655670166016,
283
+ 0.012290047645568849,
284
+ 0.012247039794921874,
285
+ 0.012275712013244629,
286
+ 0.012232704162597656,
287
+ 0.012267519950866699,
288
+ 0.012236800193786621,
289
+ 0.012279808044433594,
290
+ 0.012236800193786621,
291
+ 0.012337183952331543,
292
+ 0.01222550392150879,
293
+ 0.012263423919677734,
294
+ 0.012240896224975586,
295
+ 0.012277759552001954,
296
+ 0.012232704162597656,
297
+ 0.012277759552001954,
298
+ 0.01221939182281494,
299
+ 0.012285951614379884,
300
+ 0.012320768356323243,
301
+ 0.012465151786804199,
302
+ 0.012257280349731446,
303
+ 0.012280832290649414,
304
+ 0.012223487854003906,
305
+ 0.01226854419708252,
306
+ 0.012222463607788087,
307
+ 0.012264448165893555,
308
+ 0.012221440315246582,
309
+ 0.012289024353027344,
310
+ 0.012254207611083985,
311
+ 0.012270591735839843,
312
+ 0.012237824440002442,
313
+ 0.012299263954162597,
314
+ 0.012254207611083985,
315
+ 0.012267519950866699,
316
+ 0.012224512100219726,
317
+ 0.012266495704650878,
318
+ 0.012237824440002442,
319
+ 0.012277759552001954,
320
+ 0.012228608131408691,
321
+ 0.012264448165893555,
322
+ 0.012228608131408691,
323
+ 0.012272640228271485,
324
+ 0.012225536346435547,
325
+ 0.012279808044433594,
326
+ 0.01222758388519287,
327
+ 0.012266495704650878,
328
+ 0.012237824440002442,
329
+ 0.012291071891784668
330
+ ]
331
+ },
332
+ "throughput": {
333
+ "unit": "tokens/s",
334
+ "value": 80.36137122882384
335
+ },
336
+ "energy": null,
337
+ "efficiency": null
338
+ }
339
+ }
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/backend.cache_implementation=static,backend.model=google/gemma-2b,backend.torch_compile=True/summary.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model": "google/gemma-2b",
3
+ "commit": "bd5091df8db7cea1a9f94f797fc11487f840ade1",
4
+ "config": "backend.cache_implementation=static,backend.torch_compile=True",
5
+ "metrics": {
6
+ "prefill.latency.mean": 0.014144016265869142,
7
+ "prefill.throughput.value": 494.9089331077529,
8
+ "decode.latency.mean": 1.5736651611328125,
9
+ "decode.throughput.value": 80.70331804802635,
10
+ "per_token.latency.mean": 0.012443789655512029,
11
+ "per_token.throughput.value": 80.36137122882384
12
+ }
13
+ }
2024-06-05/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/multirun.yaml ADDED
@@ -0,0 +1,232 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ hydra:
2
+ run:
3
+ dir: _benchmark/tmpjivn_1kh/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1/${hydra.job.override_dirname}
4
+ sweep:
5
+ dir: _benchmark/tmpjivn_1kh/commit=bd5091df8db7cea1a9f94f797fc11487f840ade1
6
+ subdir: ${hydra.job.override_dirname}
7
+ launcher:
8
+ _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
9
+ sweeper:
10
+ _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
11
+ max_batch_size: null
12
+ params: null
13
+ help:
14
+ app_name: ${hydra.job.name}
15
+ header: '${hydra.help.app_name} is powered by Hydra.
16
+
17
+ '
18
+ footer: 'Powered by Hydra (https://hydra.cc)
19
+
20
+ Use --hydra-help to view Hydra specific help
21
+
22
+ '
23
+ template: '${hydra.help.header}
24
+
25
+ == Configuration groups ==
26
+
27
+ Compose your configuration from those groups (group=option)
28
+
29
+
30
+ $APP_CONFIG_GROUPS
31
+
32
+
33
+ == Config ==
34
+
35
+ Override anything in the config (foo.bar=value)
36
+
37
+
38
+ $CONFIG
39
+
40
+
41
+ ${hydra.help.footer}
42
+
43
+ '
44
+ hydra_help:
45
+ template: 'Hydra (${hydra.runtime.version})
46
+
47
+ See https://hydra.cc for more info.
48
+
49
+
50
+ == Flags ==
51
+
52
+ $FLAGS_HELP
53
+
54
+
55
+ == Configuration groups ==
56
+
57
+ Compose your configuration from those groups (For example, append hydra/job_logging=disabled
58
+ to command line)
59
+
60
+
61
+ $HYDRA_CONFIG_GROUPS
62
+
63
+
64
+ Use ''--cfg hydra'' to Show the Hydra config.
65
+
66
+ '
67
+ hydra_help: ???
68
+ hydra_logging:
69
+ version: 1
70
+ root:
71
+ level: ERROR
72
+ disable_existing_loggers: true
73
+ job_logging:
74
+ version: 1
75
+ root:
76
+ level: ERROR
77
+ disable_existing_loggers: true
78
+ env: {}
79
+ mode: MULTIRUN
80
+ searchpath: []
81
+ callbacks: {}
82
+ output_subdir: .hydra
83
+ overrides:
84
+ hydra:
85
+ - hydra/job_logging=disabled
86
+ - hydra/hydra_logging=disabled
87
+ - hydra.sweep.dir=_benchmark/tmpjivn_1kh/commit\=bd5091df8db7cea1a9f94f797fc11487f840ade1
88
+ - hydra.run.dir=_benchmark/tmpjivn_1kh/commit\=bd5091df8db7cea1a9f94f797fc11487f840ade1/${hydra.job.override_dirname}
89
+ - hydra.mode=MULTIRUN
90
+ task:
91
+ - backend.model=google/gemma-2b
92
+ - backend.cache_implementation=null,static
93
+ - backend.torch_compile=false,true
94
+ job:
95
+ name: cli
96
+ chdir: true
97
+ override_dirname: backend.cache_implementation=null,static,backend.model=google/gemma-2b,backend.torch_compile=false,true
98
+ id: ???
99
+ num: ???
100
+ config_name: generation
101
+ env_set:
102
+ OVERRIDE_BENCHMARKS: '1'
103
+ LOG_LEVEL: WARN
104
+ env_copy: []
105
+ config:
106
+ override_dirname:
107
+ kv_sep: '='
108
+ item_sep: ','
109
+ exclude_keys: []
110
+ runtime:
111
+ version: 1.3.2
112
+ version_base: '1.3'
113
+ cwd: /transformers
114
+ config_sources:
115
+ - path: hydra.conf
116
+ schema: pkg
117
+ provider: hydra
118
+ - path: optimum_benchmark
119
+ schema: pkg
120
+ provider: main
121
+ - path: /transformers/benchmark/config
122
+ schema: file
123
+ provider: command-line
124
+ - path: ''
125
+ schema: structured
126
+ provider: schema
127
+ output_dir: ???
128
+ choices:
129
+ backend: pytorch
130
+ launcher: process
131
+ scenario: inference
132
+ hydra/env: default
133
+ hydra/callbacks: null
134
+ hydra/job_logging: disabled
135
+ hydra/hydra_logging: disabled
136
+ hydra/hydra_help: default
137
+ hydra/help: default
138
+ hydra/sweeper: basic
139
+ hydra/launcher: basic
140
+ hydra/output: default
141
+ verbose: false
142
+ name: pytorch_generate
143
+ backend:
144
+ name: pytorch
145
+ version: 2.3.0+cu121
146
+ _target_: optimum_benchmark.backends.pytorch.backend.PyTorchBackend
147
+ task: null
148
+ library: null
149
+ model: google/gemma-2b
150
+ processor: null
151
+ device: cuda
152
+ device_ids: '0'
153
+ seed: 42
154
+ inter_op_num_threads: null
155
+ intra_op_num_threads: null
156
+ model_kwargs: {}
157
+ processor_kwargs: {}
158
+ hub_kwargs: {}
159
+ no_weights: true
160
+ device_map: null
161
+ torch_dtype: float16
162
+ eval_mode: true
163
+ to_bettertransformer: false
164
+ low_cpu_mem_usage: null
165
+ attn_implementation: null
166
+ cache_implementation: static
167
+ autocast_enabled: false
168
+ autocast_dtype: null
169
+ torch_compile: true
170
+ torch_compile_target: forward
171
+ torch_compile_config:
172
+ backend: inductor
173
+ mode: reduce-overhead
174
+ fullgraph: true
175
+ quantization_scheme: null
176
+ quantization_config: {}
177
+ deepspeed_inference: false
178
+ deepspeed_inference_config: {}
179
+ peft_type: null
180
+ peft_config: {}
181
+ scenario:
182
+ name: inference
183
+ _target_: optimum_benchmark.scenarios.inference.scenario.InferenceScenario
184
+ iterations: 2
185
+ duration: 0
186
+ warmup_runs: 10
187
+ input_shapes:
188
+ batch_size: 1
189
+ sequence_length: 7
190
+ new_tokens: null
191
+ latency: true
192
+ memory: true
193
+ energy: false
194
+ forward_kwargs: {}
195
+ generate_kwargs:
196
+ max_new_tokens: 128
197
+ min_new_tokens: 128
198
+ do_sample: false
199
+ call_kwargs: {}
200
+ launcher:
201
+ name: process
202
+ _target_: optimum_benchmark.launchers.process.launcher.ProcessLauncher
203
+ device_isolation: true
204
+ device_isolation_action: warn
205
+ start_method: spawn
206
+ environment:
207
+ cpu: ' AMD EPYC 7R32'
208
+ cpu_count: 16
209
+ cpu_ram_mb: 66697.29792
210
+ system: Linux
211
+ machine: x86_64
212
+ platform: Linux-5.10.217-205.860.amzn2.x86_64-x86_64-with-glibc2.29
213
+ processor: x86_64
214
+ python_version: 3.8.10
215
+ gpu:
216
+ - NVIDIA A10G
217
+ gpu_count: 1
218
+ gpu_vram_mb: 24146608128
219
+ optimum_benchmark_version: 0.2.1
220
+ optimum_benchmark_commit: null
221
+ transformers_version: 4.42.0.dev0
222
+ transformers_commit: bd5091df8db7cea1a9f94f797fc11487f840ade1
223
+ accelerate_version: 0.31.0.dev0
224
+ accelerate_commit: null
225
+ diffusers_version: null
226
+ diffusers_commit: null
227
+ optimum_version: 1.21.0.dev0
228
+ optimum_commit: null
229
+ timm_version: 0.9.16
230
+ timm_commit: null
231
+ peft_version: 0.11.2.dev0
232
+ peft_commit: null
2024-06-05/summaries.json CHANGED
@@ -1,41 +1,41 @@
1
  [
2
  {
3
  "model": "google/gemma-2b",
4
- "commit": "f14d06e0bcaecb02c9a0075ce0c7379345d718bc",
5
  "config": "backend.cache_implementation=null,backend.torch_compile=False",
6
  "metrics": {
7
- "prefill.latency.mean": 0.022776783943176268,
8
- "prefill.throughput.value": 307.3304825415065,
9
- "decode.latency.mean": 2.443668212890625,
10
- "decode.throughput.value": 51.97104882326525,
11
- "per_token.latency.mean": 0.01931815664570322,
12
- "per_token.throughput.value": 51.764773334231236
13
  }
14
  },
15
  {
16
  "model": "google/gemma-2b",
17
- "commit": "f14d06e0bcaecb02c9a0075ce0c7379345d718bc",
18
  "config": "backend.cache_implementation=static,backend.torch_compile=False",
19
  "metrics": {
20
- "prefill.latency.mean": 0.02206830406188965,
21
- "prefill.throughput.value": 317.1970070907483,
22
- "decode.latency.mean": 2.59590380859375,
23
- "decode.throughput.value": 48.92323035220565,
24
- "per_token.latency.mean": 0.02052478912413827,
25
- "per_token.throughput.value": 48.72157243379156
26
  }
27
  },
28
  {
29
  "model": "google/gemma-2b",
30
- "commit": "f14d06e0bcaecb02c9a0075ce0c7379345d718bc",
31
  "config": "backend.cache_implementation=static,backend.torch_compile=True",
32
  "metrics": {
33
- "prefill.latency.mean": 0.013985567569732666,
34
- "prefill.throughput.value": 500.51597585136864,
35
- "decode.latency.mean": 1.5596350097656249,
36
- "decode.throughput.value": 81.42930827071201,
37
- "per_token.latency.mean": 0.012332881980262724,
38
- "per_token.throughput.value": 81.08404844872256
39
  }
40
  }
41
  ]
 
1
  [
2
  {
3
  "model": "google/gemma-2b",
4
+ "commit": "bd5091df8db7cea1a9f94f797fc11487f840ade1",
5
  "config": "backend.cache_implementation=null,backend.torch_compile=False",
6
  "metrics": {
7
+ "prefill.latency.mean": 0.020133775711059572,
8
+ "prefill.throughput.value": 347.67447996129556,
9
+ "decode.latency.mean": 2.4417091064453125,
10
+ "decode.throughput.value": 52.0127478186331,
11
+ "per_token.latency.mean": 0.019298247371266484,
12
+ "per_token.throughput.value": 51.81817709979811
13
  }
14
  },
15
  {
16
  "model": "google/gemma-2b",
17
+ "commit": "bd5091df8db7cea1a9f94f797fc11487f840ade1",
18
  "config": "backend.cache_implementation=static,backend.torch_compile=False",
19
  "metrics": {
20
+ "prefill.latency.mean": 0.02215336036682129,
21
+ "prefill.throughput.value": 315.97915097719357,
22
+ "decode.latency.mean": 2.565888427734375,
23
+ "decode.throughput.value": 49.49552701796091,
24
+ "per_token.latency.mean": 0.020287751080961567,
25
+ "per_token.throughput.value": 49.29082558285231
26
  }
27
  },
28
  {
29
  "model": "google/gemma-2b",
30
+ "commit": "bd5091df8db7cea1a9f94f797fc11487f840ade1",
31
  "config": "backend.cache_implementation=static,backend.torch_compile=True",
32
  "metrics": {
33
+ "prefill.latency.mean": 0.014144016265869142,
34
+ "prefill.throughput.value": 494.9089331077529,
35
+ "decode.latency.mean": 1.5736651611328125,
36
+ "decode.throughput.value": 80.70331804802635,
37
+ "per_token.latency.mean": 0.012443789655512029,
38
+ "per_token.throughput.value": 80.36137122882384
39
  }
40
  }
41
  ]
2024-06-05/summary.json CHANGED
@@ -1,38 +1,38 @@
1
  {
2
  "google/gemma-2b": {
3
  "backend.cache_implementation=null,backend.torch_compile=False": {
4
- "f14d06e0bcaecb02c9a0075ce0c7379345d718bc": {
5
  "metrics": {
6
- "prefill.latency.mean": 0.022776783943176268,
7
- "prefill.throughput.value": 307.3304825415065,
8
- "decode.latency.mean": 2.443668212890625,
9
- "decode.throughput.value": 51.97104882326525,
10
- "per_token.latency.mean": 0.01931815664570322,
11
- "per_token.throughput.value": 51.764773334231236
12
  }
13
  }
14
  },
15
  "backend.cache_implementation=static,backend.torch_compile=False": {
16
- "f14d06e0bcaecb02c9a0075ce0c7379345d718bc": {
17
  "metrics": {
18
- "prefill.latency.mean": 0.02206830406188965,
19
- "prefill.throughput.value": 317.1970070907483,
20
- "decode.latency.mean": 2.59590380859375,
21
- "decode.throughput.value": 48.92323035220565,
22
- "per_token.latency.mean": 0.02052478912413827,
23
- "per_token.throughput.value": 48.72157243379156
24
  }
25
  }
26
  },
27
  "backend.cache_implementation=static,backend.torch_compile=True": {
28
- "f14d06e0bcaecb02c9a0075ce0c7379345d718bc": {
29
  "metrics": {
30
- "prefill.latency.mean": 0.013985567569732666,
31
- "prefill.throughput.value": 500.51597585136864,
32
- "decode.latency.mean": 1.5596350097656249,
33
- "decode.throughput.value": 81.42930827071201,
34
- "per_token.latency.mean": 0.012332881980262724,
35
- "per_token.throughput.value": 81.08404844872256
36
  }
37
  }
38
  }
 
1
  {
2
  "google/gemma-2b": {
3
  "backend.cache_implementation=null,backend.torch_compile=False": {
4
+ "bd5091df8db7cea1a9f94f797fc11487f840ade1": {
5
  "metrics": {
6
+ "prefill.latency.mean": 0.020133775711059572,
7
+ "prefill.throughput.value": 347.67447996129556,
8
+ "decode.latency.mean": 2.4417091064453125,
9
+ "decode.throughput.value": 52.0127478186331,
10
+ "per_token.latency.mean": 0.019298247371266484,
11
+ "per_token.throughput.value": 51.81817709979811
12
  }
13
  }
14
  },
15
  "backend.cache_implementation=static,backend.torch_compile=False": {
16
+ "bd5091df8db7cea1a9f94f797fc11487f840ade1": {
17
  "metrics": {
18
+ "prefill.latency.mean": 0.02215336036682129,
19
+ "prefill.throughput.value": 315.97915097719357,
20
+ "decode.latency.mean": 2.565888427734375,
21
+ "decode.throughput.value": 49.49552701796091,
22
+ "per_token.latency.mean": 0.020287751080961567,
23
+ "per_token.throughput.value": 49.29082558285231
24
  }
25
  }
26
  },
27
  "backend.cache_implementation=static,backend.torch_compile=True": {
28
+ "bd5091df8db7cea1a9f94f797fc11487f840ade1": {
29
  "metrics": {
30
+ "prefill.latency.mean": 0.014144016265869142,
31
+ "prefill.throughput.value": 494.9089331077529,
32
+ "decode.latency.mean": 1.5736651611328125,
33
+ "decode.throughput.value": 80.70331804802635,
34
+ "per_token.latency.mean": 0.012443789655512029,
35
+ "per_token.throughput.value": 80.36137122882384
36
  }
37
  }
38
  }