imb 1.0.0__tar.gz → 1.0.1__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (20) hide show
  1. imb-1.0.1/PKG-INFO +105 -0
  2. imb-1.0.1/README.md +70 -0
  3. imb-1.0.1/imb/__init__.py +0 -0
  4. imb-1.0.1/imb.egg-info/PKG-INFO +105 -0
  5. {imb-1.0.0 → imb-1.0.1}/imb.egg-info/SOURCES.txt +4 -5
  6. imb-1.0.1/imb.egg-info/requires.txt +15 -0
  7. {imb-1.0.0 → imb-1.0.1}/setup.py +8 -1
  8. imb-1.0.0/PKG-INFO +0 -30
  9. imb-1.0.0/README.md +0 -4
  10. imb-1.0.0/imb/__init__.py +0 -1
  11. imb-1.0.0/imb/inference_clients/__init__.py +0 -2
  12. imb-1.0.0/imb.egg-info/PKG-INFO +0 -30
  13. imb-1.0.0/imb.egg-info/requires.txt +0 -3
  14. {imb-1.0.0 → imb-1.0.1}/LICENSE +0 -0
  15. {imb-1.0.0/imb/inference_clients → imb-1.0.1/imb}/base.py +0 -0
  16. {imb-1.0.0/imb/inference_clients → imb-1.0.1/imb}/onnx.py +0 -0
  17. {imb-1.0.0/imb/inference_clients → imb-1.0.1/imb}/triton.py +0 -0
  18. {imb-1.0.0 → imb-1.0.1}/imb.egg-info/dependency_links.txt +0 -0
  19. {imb-1.0.0 → imb-1.0.1}/imb.egg-info/top_level.txt +0 -0
  20. {imb-1.0.0 → imb-1.0.1}/setup.cfg +0 -0
imb-1.0.1/PKG-INFO ADDED
@@ -0,0 +1,105 @@
1
+ Metadata-Version: 2.2
2
+ Name: imb
3
+ Version: 1.0.1
4
+ Summary: Python library for run inference of deep learning models in different backends
5
+ Home-page: https://github.com/TheConstant3/InferenceMultiBackend
6
+ Author: p-constant
7
+ Author-email: nikshorop@gmail.com
8
+ Classifier: Programming Language :: Python :: 3.8
9
+ Classifier: License :: OSI Approved :: MIT License
10
+ Classifier: Operating System :: OS Independent
11
+ Requires-Python: >=3.8
12
+ Description-Content-Type: text/markdown
13
+ License-File: LICENSE
14
+ Requires-Dist: numpy
15
+ Provides-Extra: triton
16
+ Requires-Dist: tritonclient[all]>=2.38.0; extra == "triton"
17
+ Provides-Extra: onnxcpu
18
+ Requires-Dist: onnxruntime>=1.16.0; extra == "onnxcpu"
19
+ Provides-Extra: onnxgpu
20
+ Requires-Dist: onnxruntime-gpu>=1.16.0; extra == "onnxgpu"
21
+ Provides-Extra: all
22
+ Requires-Dist: tritonclient[all]>=2.38.0; extra == "all"
23
+ Requires-Dist: onnxruntime>=1.16.0; extra == "all"
24
+ Requires-Dist: onnxruntime-gpu>=1.16.0; extra == "all"
25
+ Dynamic: author
26
+ Dynamic: author-email
27
+ Dynamic: classifier
28
+ Dynamic: description
29
+ Dynamic: description-content-type
30
+ Dynamic: home-page
31
+ Dynamic: provides-extra
32
+ Dynamic: requires-dist
33
+ Dynamic: requires-python
34
+ Dynamic: summary
35
+
36
+ # InferenceMultiBackend
37
+
38
+ Python library for run inference of deep learning models in different backends
39
+
40
+ ## Installation
41
+
42
+ For use triton inference client:
43
+ ```pip install imb[triton]```
44
+
45
+ For use onnxruntime-gpu client:
46
+ ```pip install imb[onnxgpu]```
47
+
48
+ For use onnxruntime client:
49
+ ```pip install imb[onnxcpu]```
50
+
51
+ For support all implemented clients:
52
+ ```pip install imb[all]```
53
+
54
+ ## Usage
55
+
56
+ OnnxClient usage example
57
+ ```
58
+ onnx_client = OnnxClient(
59
+ model_path='model.onnx',
60
+ model_name='any name',
61
+ providers=['CUDAExecutionProvider', 'CPUExecutionProvider'],
62
+ max_batch_size=16,
63
+ return_dict=True,
64
+ fixed_batch=True,
65
+ warmup=True
66
+ )
67
+ # if model has fixed input size (except batch size) then sample_inputs will be created
68
+ sample_inputs = onnx_client.sample_inputs
69
+ print('inputs shapes', [o.shape for o in sample_inputs])
70
+ outputs = onnx_client(*sample_inputs)
71
+ print('outputs shapes', [(o_name, o_value.shape) for o_name, o_value in outputs.items()])
72
+ ```
73
+
74
+ TritonClient usage example
75
+ ```
76
+ triton_client = TritonClient(
77
+ url='localhost:8000',
78
+ model_name='arcface',
79
+ max_batch_size=16,
80
+ timeout=10,
81
+ resend_count=10,
82
+ fixed_batch=True,
83
+ is_async=False,
84
+ cuda_shm=False,
85
+ max_shm_regions=2,
86
+ scheme='http',
87
+ return_dict=True,
88
+ warmup=False
89
+ )
90
+ # if model has fixed input size (except batch size) then sample_inputs will be created
91
+ sample_inputs = triton_client.sample_inputs
92
+ print('inputs shapes', [o.shape for o in sample_inputs])
93
+ outputs = triton_client(*sample_inputs)
94
+ print('outputs shapes', [(o_name, o_value.shape) for o_name, o_value in outputs.items()])
95
+ ```
96
+
97
+ ## Notes
98
+
99
+ max_batch_size - maximum batch size for inference. If input data larger that max_batch_size, then input data will be splitted to several batches.
100
+
101
+ fixed_batch - if fixed batch is True, then each batch will have fixed size (padding the smallest batch to max_batch_size).
102
+
103
+ warmup - if True, model will run several calls on sample_inputs while initialization.
104
+
105
+ return_dict - if True, __call__ return dict {'output_name1': output_value1, ...}, else [output_value1, ...]
imb-1.0.1/README.md ADDED
@@ -0,0 +1,70 @@
1
+ # InferenceMultiBackend
2
+
3
+ Python library for run inference of deep learning models in different backends
4
+
5
+ ## Installation
6
+
7
+ For use triton inference client:
8
+ ```pip install imb[triton]```
9
+
10
+ For use onnxruntime-gpu client:
11
+ ```pip install imb[onnxgpu]```
12
+
13
+ For use onnxruntime client:
14
+ ```pip install imb[onnxcpu]```
15
+
16
+ For support all implemented clients:
17
+ ```pip install imb[all]```
18
+
19
+ ## Usage
20
+
21
+ OnnxClient usage example
22
+ ```
23
+ onnx_client = OnnxClient(
24
+ model_path='model.onnx',
25
+ model_name='any name',
26
+ providers=['CUDAExecutionProvider', 'CPUExecutionProvider'],
27
+ max_batch_size=16,
28
+ return_dict=True,
29
+ fixed_batch=True,
30
+ warmup=True
31
+ )
32
+ # if model has fixed input size (except batch size) then sample_inputs will be created
33
+ sample_inputs = onnx_client.sample_inputs
34
+ print('inputs shapes', [o.shape for o in sample_inputs])
35
+ outputs = onnx_client(*sample_inputs)
36
+ print('outputs shapes', [(o_name, o_value.shape) for o_name, o_value in outputs.items()])
37
+ ```
38
+
39
+ TritonClient usage example
40
+ ```
41
+ triton_client = TritonClient(
42
+ url='localhost:8000',
43
+ model_name='arcface',
44
+ max_batch_size=16,
45
+ timeout=10,
46
+ resend_count=10,
47
+ fixed_batch=True,
48
+ is_async=False,
49
+ cuda_shm=False,
50
+ max_shm_regions=2,
51
+ scheme='http',
52
+ return_dict=True,
53
+ warmup=False
54
+ )
55
+ # if model has fixed input size (except batch size) then sample_inputs will be created
56
+ sample_inputs = triton_client.sample_inputs
57
+ print('inputs shapes', [o.shape for o in sample_inputs])
58
+ outputs = triton_client(*sample_inputs)
59
+ print('outputs shapes', [(o_name, o_value.shape) for o_name, o_value in outputs.items()])
60
+ ```
61
+
62
+ ## Notes
63
+
64
+ max_batch_size - maximum batch size for inference. If input data larger that max_batch_size, then input data will be splitted to several batches.
65
+
66
+ fixed_batch - if fixed batch is True, then each batch will have fixed size (padding the smallest batch to max_batch_size).
67
+
68
+ warmup - if True, model will run several calls on sample_inputs while initialization.
69
+
70
+ return_dict - if True, __call__ return dict {'output_name1': output_value1, ...}, else [output_value1, ...]
imb-1.0.1/imb/__init__.py ADDED
File without changes
imb-1.0.1/imb.egg-info/PKG-INFO ADDED
@@ -0,0 +1,105 @@
1
+ Metadata-Version: 2.2
2
+ Name: imb
3
+ Version: 1.0.1
4
+ Summary: Python library for run inference of deep learning models in different backends
5
+ Home-page: https://github.com/TheConstant3/InferenceMultiBackend
6
+ Author: p-constant
7
+ Author-email: nikshorop@gmail.com
8
+ Classifier: Programming Language :: Python :: 3.8
9
+ Classifier: License :: OSI Approved :: MIT License
10
+ Classifier: Operating System :: OS Independent
11
+ Requires-Python: >=3.8
12
+ Description-Content-Type: text/markdown
13
+ License-File: LICENSE
14
+ Requires-Dist: numpy
15
+ Provides-Extra: triton
16
+ Requires-Dist: tritonclient[all]>=2.38.0; extra == "triton"
17
+ Provides-Extra: onnxcpu
18
+ Requires-Dist: onnxruntime>=1.16.0; extra == "onnxcpu"
19
+ Provides-Extra: onnxgpu
20
+ Requires-Dist: onnxruntime-gpu>=1.16.0; extra == "onnxgpu"
21
+ Provides-Extra: all
22
+ Requires-Dist: tritonclient[all]>=2.38.0; extra == "all"
23
+ Requires-Dist: onnxruntime>=1.16.0; extra == "all"
24
+ Requires-Dist: onnxruntime-gpu>=1.16.0; extra == "all"
25
+ Dynamic: author
26
+ Dynamic: author-email
27
+ Dynamic: classifier
28
+ Dynamic: description
29
+ Dynamic: description-content-type
30
+ Dynamic: home-page
31
+ Dynamic: provides-extra
32
+ Dynamic: requires-dist
33
+ Dynamic: requires-python
34
+ Dynamic: summary
35
+
36
+ # InferenceMultiBackend
37
+
38
+ Python library for run inference of deep learning models in different backends
39
+
40
+ ## Installation
41
+
42
+ For use triton inference client:
43
+ ```pip install imb[triton]```
44
+
45
+ For use onnxruntime-gpu client:
46
+ ```pip install imb[onnxgpu]```
47
+
48
+ For use onnxruntime client:
49
+ ```pip install imb[onnxcpu]```
50
+
51
+ For support all implemented clients:
52
+ ```pip install imb[all]```
53
+
54
+ ## Usage
55
+
56
+ OnnxClient usage example
57
+ ```
58
+ onnx_client = OnnxClient(
59
+ model_path='model.onnx',
60
+ model_name='any name',
61
+ providers=['CUDAExecutionProvider', 'CPUExecutionProvider'],
62
+ max_batch_size=16,
63
+ return_dict=True,
64
+ fixed_batch=True,
65
+ warmup=True
66
+ )
67
+ # if model has fixed input size (except batch size) then sample_inputs will be created
68
+ sample_inputs = onnx_client.sample_inputs
69
+ print('inputs shapes', [o.shape for o in sample_inputs])
70
+ outputs = onnx_client(*sample_inputs)
71
+ print('outputs shapes', [(o_name, o_value.shape) for o_name, o_value in outputs.items()])
72
+ ```
73
+
74
+ TritonClient usage example
75
+ ```
76
+ triton_client = TritonClient(
77
+ url='localhost:8000',
78
+ model_name='arcface',
79
+ max_batch_size=16,
80
+ timeout=10,
81
+ resend_count=10,
82
+ fixed_batch=True,
83
+ is_async=False,
84
+ cuda_shm=False,
85
+ max_shm_regions=2,
86
+ scheme='http',
87
+ return_dict=True,
88
+ warmup=False
89
+ )
90
+ # if model has fixed input size (except batch size) then sample_inputs will be created
91
+ sample_inputs = triton_client.sample_inputs
92
+ print('inputs shapes', [o.shape for o in sample_inputs])
93
+ outputs = triton_client(*sample_inputs)
94
+ print('outputs shapes', [(o_name, o_value.shape) for o_name, o_value in outputs.items()])
95
+ ```
96
+
97
+ ## Notes
98
+
99
+ max_batch_size - maximum batch size for inference. If input data larger that max_batch_size, then input data will be splitted to several batches.
100
+
101
+ fixed_batch - if fixed batch is True, then each batch will have fixed size (padding the smallest batch to max_batch_size).
102
+
103
+ warmup - if True, model will run several calls on sample_inputs while initialization.
104
+
105
+ return_dict - if True, __call__ return dict {'output_name1': output_value1, ...}, else [output_value1, ...]
{imb-1.0.0 → imb-1.0.1}/imb.egg-info/SOURCES.txt RENAMED
@@ -3,12 +3,11 @@ README.md
3
3
  setup.cfg
4
4
  setup.py
5
5
  imb/__init__.py
6
+ imb/base.py
7
+ imb/onnx.py
8
+ imb/triton.py
6
9
  imb.egg-info/PKG-INFO
7
10
  imb.egg-info/SOURCES.txt
8
11
  imb.egg-info/dependency_links.txt
9
12
  imb.egg-info/requires.txt
10
- imb.egg-info/top_level.txt
11
- imb/inference_clients/__init__.py
12
- imb/inference_clients/base.py
13
- imb/inference_clients/onnx.py
14
- imb/inference_clients/triton.py
13
+ imb.egg-info/top_level.txt
imb-1.0.1/imb.egg-info/requires.txt ADDED
@@ -0,0 +1,15 @@
1
+ numpy
2
+
3
+ [all]
4
+ tritonclient[all]>=2.38.0
5
+ onnxruntime>=1.16.0
6
+ onnxruntime-gpu>=1.16.0
7
+
8
+ [onnxcpu]
9
+ onnxruntime>=1.16.0
10
+
11
+ [onnxgpu]
12
+ onnxruntime-gpu>=1.16.0
13
+
14
+ [triton]
15
+ tritonclient[all]>=2.38.0
{imb-1.0.0 → imb-1.0.1}/setup.py RENAMED
@@ -1,4 +1,5 @@
1
1
  from setuptools import setup, find_packages
2
+ from itertools import chain
2
3
  import os
3
4
 
4
5
 
@@ -11,9 +12,14 @@ def readme():
11
12
  with open('README.md', 'r') as f:
12
13
  return f.read()
13
14
 
15
+ extras = ['triton', 'onnxcpu', 'onnxgpu']
16
+ extras_require = {extra: req_file(f"requirements_{extra}.txt") for extra in extras}
17
+ extras_require["all"] = list(chain(extras_require.values()))
18
+
19
+
14
20
  setup(
15
21
  name='imb',
16
- version='1.0.0',
22
+ version='1.0.1',
17
23
  author='p-constant',
18
24
  author_email='nikshorop@gmail.com',
19
25
  description='Python library for run inference of deep learning models in different backends',
@@ -22,6 +28,7 @@ setup(
22
28
  url='https://github.com/TheConstant3/InferenceMultiBackend',
23
29
  packages=find_packages(),
24
30
  install_requires=req_file(),
31
+ extras_require=extras_require,
25
32
  classifiers=[
26
33
  "Programming Language :: Python :: 3.8",
27
34
  "License :: OSI Approved :: MIT License",
imb-1.0.0/PKG-INFO DELETED
@@ -1,30 +0,0 @@
1
- Metadata-Version: 2.2
2
- Name: imb
3
- Version: 1.0.0
4
- Summary: Python library for run inference of deep learning models in different backends
5
- Home-page: https://github.com/TheConstant3/InferenceMultiBackend
6
- Author: p-constant
7
- Author-email: nikshorop@gmail.com
8
- Classifier: Programming Language :: Python :: 3.8
9
- Classifier: License :: OSI Approved :: MIT License
10
- Classifier: Operating System :: OS Independent
11
- Requires-Python: >=3.8
12
- Description-Content-Type: text/markdown
13
- License-File: LICENSE
14
- Requires-Dist: onnxruntime-gpu>=1.16.0
15
- Requires-Dist: tritonclient[all]>=2.38.0
16
- Requires-Dist: numpy>=1.19.4
17
- Dynamic: author
18
- Dynamic: author-email
19
- Dynamic: classifier
20
- Dynamic: description
21
- Dynamic: description-content-type
22
- Dynamic: home-page
23
- Dynamic: requires-dist
24
- Dynamic: requires-python
25
- Dynamic: summary
26
-
27
- # InferenceMultiBackend
28
-
29
- Python library for run inference of deep learning models in different backends
30
-
imb-1.0.0/README.md DELETED
@@ -1,4 +0,0 @@
1
- # InferenceMultiBackend
2
-
3
- Python library for run inference of deep learning models in different backends
4
-
imb-1.0.0/imb/__init__.py DELETED
@@ -1 +0,0 @@
1
- from .inference_clients import OnnxClient, TritonClient
imb-1.0.0/imb/inference_clients/__init__.py DELETED
@@ -1,2 +0,0 @@
1
- from .onnx import OnnxClient
2
- from .triton import TritonClient
imb-1.0.0/imb.egg-info/PKG-INFO DELETED
@@ -1,30 +0,0 @@
1
- Metadata-Version: 2.2
2
- Name: imb
3
- Version: 1.0.0
4
- Summary: Python library for run inference of deep learning models in different backends
5
- Home-page: https://github.com/TheConstant3/InferenceMultiBackend
6
- Author: p-constant
7
- Author-email: nikshorop@gmail.com
8
- Classifier: Programming Language :: Python :: 3.8
9
- Classifier: License :: OSI Approved :: MIT License
10
- Classifier: Operating System :: OS Independent
11
- Requires-Python: >=3.8
12
- Description-Content-Type: text/markdown
13
- License-File: LICENSE
14
- Requires-Dist: onnxruntime-gpu>=1.16.0
15
- Requires-Dist: tritonclient[all]>=2.38.0
16
- Requires-Dist: numpy>=1.19.4
17
- Dynamic: author
18
- Dynamic: author-email
19
- Dynamic: classifier
20
- Dynamic: description
21
- Dynamic: description-content-type
22
- Dynamic: home-page
23
- Dynamic: requires-dist
24
- Dynamic: requires-python
25
- Dynamic: summary
26
-
27
- # InferenceMultiBackend
28
-
29
- Python library for run inference of deep learning models in different backends
30
-
imb-1.0.0/imb.egg-info/requires.txt DELETED
@@ -1,3 +0,0 @@
1
- onnxruntime-gpu>=1.16.0
2
- tritonclient[all]>=2.38.0
3
- numpy>=1.19.4
{imb-1.0.0 → imb-1.0.1}/LICENSE RENAMED
File without changes
{imb-1.0.0 → imb-1.0.1}/imb.egg-info/top_level.txt RENAMED
File without changes
{imb-1.0.0 → imb-1.0.1}/setup.cfg RENAMED
File without changes