Skip to content
This repository was archived by the owner on Jun 3, 2025. It is now read-only.

Commit adceb91

Browse files
Beth-Kosisrobertgshaw2-redhatjeanniefinks
authored
B kosis products (#134)
* Update deepsparse.mdx * Update community.mdx * Update enterprise.mdx * Update community.mdx * Update community.mdx * Update community.mdx * Update enterprise.mdx * Update sparseml.mdx * Update sparsezoo.mdx * Update enterprise.mdx Updated to a better link * Update src/content/products/deepsparse/community.mdx Co-authored-by: Robert Shaw <[email protected]> * Update src/content/products/deepsparse/community.mdx Co-authored-by: Robert Shaw <[email protected]> * Update src/content/products/deepsparse/community.mdx Co-authored-by: Robert Shaw <[email protected]> * Update src/content/products/deepsparse/enterprise.mdx Co-authored-by: Robert Shaw <[email protected]> * Update src/content/products/deepsparse/enterprise.mdx Co-authored-by: Robert Shaw <[email protected]> * Update src/content/products/deepsparse/enterprise.mdx Co-authored-by: Robert Shaw <[email protected]> * Update src/content/products/deepsparse/enterprise.mdx Co-authored-by: Robert Shaw <[email protected]> * Update src/content/products/deepsparse/enterprise.mdx Co-authored-by: Robert Shaw <[email protected]> * Update src/content/products/deepsparse/enterprise.mdx Co-authored-by: Robert Shaw <[email protected]> * Update src/content/products/sparsezoo.mdx Co-authored-by: Jeannie Finks <[email protected]> * Update src/content/products/deepsparse/community.mdx * Update src/content/products/deepsparse/community.mdx * Update src/content/products/deepsparse/enterprise.mdx * Update src/content/products/deepsparse/enterprise.mdx Co-authored-by: Robert Shaw <[email protected]> Co-authored-by: Jeannie Finks <[email protected]>
1 parent 999a946 commit adceb91

File tree

5 files changed

+159
-192
lines changed

5 files changed

+159
-192
lines changed

src/content/products/deepsparse.mdx

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -42,7 +42,7 @@ index: 1000
4242
</div>
4343
</div>
4444

45-
A CPU runtime that takes advantage of sparsity within neural networks to reduce compute. Read [more about sparsification](https://docs.neuralmagic.com/user-guide/sparsification).
45+
DeepSparse is a CPU runtime that takes advantage of sparsity within neural networks to reduce compute. Read [more about sparsification](https://docs.neuralmagic.com/user-guide/sparsification).
4646

4747
Neural Magic's DeepSparse Engine is able to integrate into popular deep learning libraries (e.g., Hugging Face, Ultralytics) allowing you to leverage DeepSparse for loading and deploying sparse models with ONNX.
4848
ONNX gives the flexibility to serve your model in a framework-agnostic environment.

src/content/products/deepsparse/community.mdx

Lines changed: 45 additions & 62 deletions
Original file line numberDiff line numberDiff line change
@@ -42,47 +42,34 @@ index: 1000
4242
</div>
4343
</div>
4444

45-
A CPU runtime that takes advantage of sparsity within neural networks to reduce compute. Read more about sparsification Read more about sparsification [here](https://docs.neuralmagic.com/user-guide/sparsification).
45+
<p><br></p>
46+
DeepSparse Community Edition is open-source and free for evaluation, research, and non-production use with our [Engine Community License](https://neuralmagic.com/legal/engine-license-agreement/). (Alternatively, the [Enterprise Edition](https://docs.neuralmagic.com/products/deepsparse-ent) requires a Trial License or can be fully licensed for production, commercial applications.)
4647

47-
Neural Magic's DeepSparse Engine is able to integrate into popular deep learning libraries (e.g., Hugging Face, Ultralytics) allowing you to leverage DeepSparse for loading and deploying sparse models with ONNX.
48-
ONNX gives the flexibility to serve your model in a framework-agnostic environment.
49-
Support includes [PyTorch,](https://pytorch.org/docs/stable/onnx.html) [TensorFlow,](https://github.com/onnx/tensorflow-onnx) [Keras,](https://github.com/onnx/keras-onnx) and [many other frameworks](https://github.com/onnx/onnxmltools).
50-
51-
The DeepSparse Engine is available in two editions:
52-
1. [**The Community Edition**](#installation) is open-source and free for evaluation, research, and non-production use with our [Engine Community License](https://neuralmagic.com/legal/engine-license-agreement/).
53-
2. [**The Enterprise Edition**](https://docs.neuralmagic.com/products/deepsparse-ent) requires a Trial License or [can be fully licensed](https://neuralmagic.com/legal/master-software-license-and-service-agreement/) for production, commercial applications.
54-
55-
## Features
56-
57-
- 🔌 [DeepSparse Server](https://github.com/neuralmagic/deepsparse/tree/main/src/deepsparse/server)
58-
- 📜 [DeepSparse Benchmark](https://github.com/neuralmagic/deepsparse/tree/main/src/deepsparse/benchmark)
59-
- 👩‍💻 [NLP and Computer Vision Tasks Supported](https://github.com/neuralmagic/deepsparse/tree/main/examples)
60-
61-
## 🧰 Hardware Support and System Requirements
48+
## Hardware Support and System Requirements
6249

6350
Review [CPU Hardware Support for Various Architectures](https://docs.neuralmagic.com/deepsparse/source/hardware.html) to understand system requirements.
64-
The DeepSparse Engine works natively on Linux; Mac and Windows require running Linux in a Docker or virtual machine; it will not run natively on those operating systems.
51+
The DeepSparse Engine works natively on Linux. Mac and Windows require running Linux in a Docker or virtual machine; it will not run natively on those operating systems.
6552

66-
The DeepSparse Engine is tested on Python 3.7-3.10, ONNX 1.5.0-1.12.0, ONNX opset version 11+, and manylinux compliant.
53+
The DeepSparse Engine is tested on Python 3.7-3.10, ONNX 1.5.0-1.12.0, and ONNX opset version 11+. It is manylinux compliant.
6754
Using a [virtual environment](https://docs.python.org/3/library/venv.html) is highly recommended.
6855

6956
## Installation
7057

71-
Install the DeepSparse Community Edition as follows:
58+
Install the DeepSparse Community Edition with `pip`:
7259

7360
```bash
7461
pip install deepsparse
7562
```
7663

77-
See the [DeepSparse Community Installation Page](https://docs.neuralmagic.com/get-started/install/deepsparse) for further installation options.
64+
See the [DeepSparse Community Installation page](https://docs.neuralmagic.com/get-started/install/deepsparse) for further installation options.
7865

79-
To trial or inquire about licensing for DeepSparse Enterprise Edition, see the [DeepSparse Enterprise documentation](https://docs.neuralmagic.com/products/deepsparse-enterprise).
66+
## DeepSparse Community Edition Features
8067

81-
## Features
68+
### DeepSparse Server
8269

83-
### 🔌 DeepSparse Server
70+
The DeepSparse Server allows you to serve models and pipelines from the terminal. The server runs on top of the popular FastAPI web framework and Uvicorn web server.
8471

85-
The DeepSparse Server allows you to serve models and pipelines from the terminal. The server runs on top of the popular FastAPI web framework and Uvicorn web server. Install the server using the following command:
72+
Install the server with `pip`:
8673

8774
```bash
8875
pip install deepsparse[server]
@@ -98,10 +85,10 @@ deepsparse.server \
9885
--model_path "zoo:nlp/question_answering/bert-base/pytorch/huggingface/squad/12layer_pruned80_quant-none-vnni"
9986
```
10087

101-
To look up arguments run: `deepsparse.server --help`.
88+
To look up arguments, run `deepsparse.server --help`.
10289

10390
#### Multiple Models
104-
To serve multiple models in your deployment you can easily build a `config.yaml`. In the example below, we define two BERT models in our configuration for the question answering task:
91+
To serve multiple models in your deployment, you can easily build a `config.yaml`. In the example below, we define two BERT models in our configuration for the question answering task:
10592

10693
```yaml
10794
num_cores: 1
@@ -117,16 +104,16 @@ endpoints:
117104
batch_size: 1
118105
```
119106
120-
Finally, after your `config.yaml` file is built, run the server with the config file path as an argument:
107+
Finally, after your `config.yaml` file is built, run the server with the configuration file path as an argument:
121108
```bash
122109
deepsparse.server config config.yaml
123110
```
124111

125-
[Getting Started with the DeepSparse Server](https://github.com/neuralmagic/deepsparse/tree/main/src/deepsparse/server) for more info.
112+
See [Getting Started with the DeepSparse Server](https://github.com/neuralmagic/deepsparse/tree/main/src/deepsparse/server) for more information.
126113

127-
### 📜 DeepSparse Benchmark
114+
### DeepSparse Benchmark
128115

129-
The benchmark tool is available on your CLI to run expressive model benchmarks on the DeepSparse Engine with minimal parameters.
116+
The benchmark tool is available on your CLI to run expressive model benchmarks on the DeepSparse Engine with minimal parameters.
130117

131118
Run `deepsparse.benchmark -h` to look up arguments:
132119

@@ -144,7 +131,9 @@ deepsparse.benchmark [-h] [-b BATCH_SIZE] [-shapes INPUT_SHAPES]
144131
- Asynchronous (Multi-stream) Scenario
145132

146133

147-
### 👩‍💻 NLP Inference Example
134+
### NLP and Computer Vision Tasks Supported
135+
136+
An NLP inference example is:
148137

149138
```python
150139
from deepsparse import Pipeline
@@ -160,30 +149,37 @@ qa_pipeline = Pipeline.create(
160149
my_name = qa_pipeline(question="What's my name?", context="My name is Snorlax")
161150
```
162151

163-
NLP Tutorials:
164-
- [Getting Started with Hugging Face Transformers 🤗](https://github.com/neuralmagic/deepsparse/tree/main/examples/huggingface-transformers)
152+
Refer also to [Using Pipelines](https://github.com/neuralmagic/deepsparse/blob/main/src/deepsparse/PIPELINES.md).
153+
154+
- For Image Classification tutorials, see [Image Classification Inference Pipelines](https://github.com/neuralmagic/deepsparse/tree/main/src/deepsparse/image_classification).
165155

166-
Tasks Supported:
156+
- For Object Detection tutorials, see [YOLOv5 Inference Pipelines](https://github.com/neuralmagic/deepsparse/tree/main/src/deepsparse/yolo).
157+
158+
- For Segmentation tutorials, see [YOLACT Inference Pipelines](https://github.com/neuralmagic/deepsparse/tree/main/src/deepsparse/yolact).
159+
160+
- For NLP tutorials, see [Getting Started with Hugging Face Transformers](https://github.com/neuralmagic/deepsparse/tree/main/examples/huggingface-transformers).
161+
162+
Supported NLP tasks include:
167163
- [Token Classification: Named Entity Recognition](https://neuralmagic.com/use-cases/sparse-named-entity-recognition/)
168164
- [Text Classification: Multi-Class](https://neuralmagic.com/use-cases/sparse-multi-class-text-classification/)
169165
- [Text Classification: Binary](https://neuralmagic.com/use-cases/sparse-binary-text-classification/)
170166
- [Text Classification: Sentiment Analysis](https://neuralmagic.com/use-cases/sparse-sentiment-analysis/)
171167
- [Question Answering](https://neuralmagic.com/use-cases/sparse-question-answering/)
172168

173-
### 🦉 SparseZoo ONNX vs. Custom ONNX Models
169+
### SparseZoo ONNX vs. Custom ONNX Models
174170

175171
DeepSparse can accept ONNX models from two sources:
176172

177-
- **SparseZoo ONNX**: our open-source collection of sparse models available for download. [SparseZoo](https://github.com/neuralmagic/sparsezoo) hosts inference-optimized models, trained on repeatable sparsification recipes using state-of-the-art techniques from [SparseML](https://github.com/neuralmagic/sparseml).
173+
- **SparseZoo ONNX**: [SparseZoo](https://github.com/neuralmagic/sparsezoo) hosts open-source inference-optimized models, trained on repeatable sparsification recipes using state-of-the-art techniques from [SparseML](https://github.com/neuralmagic/sparseml). The ONNX representation of each model is available for download.
178174

179-
- **Custom ONNX**: your own ONNX model, can be dense or sparse. Plug in your model to compare performance with other solutions.
175+
- **Custom ONNX**: DeepSparse allows you to use your own model in ONNX format. It can be dense or sparse. Plug in your model to compare performance with other solutions.
180176

181177
```bash
182178
> wget https://github.com/onnx/models/raw/main/vision/classification/mobilenet/model/mobilenetv2-7.onnx
183179
Saving to: ‘mobilenetv2-7.onnx’
184180
```
185181

186-
Custom ONNX Benchmark example:
182+
Here is a custom ONNX benchmark example:
187183
```python
188184
from deepsparse import compile_model
189185
from deepsparse.utils import generate_random_inputs
@@ -202,46 +198,35 @@ The [GitHub repository](https://github.com/neuralmagic/deepsparse) includes pack
202198

203199
### Scheduling Single-Stream, Multi-Stream, and Elastic Inference
204200

205-
The DeepSparse Engine offers up to three types of inferences based on your use case. Read more details here: [Inference Types](https://github.com/neuralmagic/deepsparse/blob/main/docs/source/scheduler.md).
206-
207-
1 ⚡ Single-stream scheduling: the latency/synchronous scenario, requests execute serially. [`default`]
208-
209-
<img src="https://raw.githubusercontent.com/neuralmagic/deepsparse/main/docs/source/single-stream.png" alt="single stream diagram" />
201+
The DeepSparse Engine offers up to three types of inferences based on your use case. Refer also to [Inference Types](https://github.com/neuralmagic/deepsparse/blob/main/docs/source/scheduler.md).
210202

211-
Use Case: It's highly optimized for minimum per-request latency, using all of the system's resources provided to it on every request it gets.
203+
1. Single-stream scheduling (the default) is the latency/synchronous scenario. Requests execute serially.
204+
<br><img src="https://raw.githubusercontent.com/neuralmagic/deepsparse/main/docs/source/single-stream.png" alt="single stream diagram" />
205+
<br>Use Case: It's highly optimized for minimum per-request latency, using all of the system's resources provided to it on every request it gets.
212206

213-
2 ⚡ Multi-stream scheduling: the throughput/asynchronous scenario, requests execute in parallel.
207+
2. Multi-stream scheduling is the throughput/asynchronous scenario. Requests execute in parallel.
208+
<br><img src="https://raw.githubusercontent.com/neuralmagic/deepsparse/main/docs/source/multi-stream.png" alt="multi stream diagram" />
209+
<br>Use Case: The most common use cases for the multi-stream scheduler are those in which parallelism is low with respect to core count, and requests need to be made asynchronously without time to batch them.
214210

215-
<img src="https://raw.githubusercontent.com/neuralmagic/deepsparse/main/docs/source/multi-stream.png" alt="multi stream diagram" />
216-
217-
PRO TIP: The most common use cases for the multi-stream scheduler are where parallelism is low with respect to core count, and where requests need to be made asynchronously without time to batch them.
218-
219-
3 ⚡ Elastic scheduling: requests execute in parallel, but not multiplexed on individual NUMA nodes.
220-
221-
Use Case: A workload that might benefit from the elastic scheduler is one in which multiple requests need to be handled simultaneously, but where performance is hindered when those requests have to share an L3 cache.
211+
3. Elastic scheduling requests execute in parallel, but not multiplexed on individual NUMA nodes.
212+
<br>Use Case: A workload that might benefit from the elastic scheduler is one in which multiple requests need to be handled simultaneously, but where performance is hindered when those requests have to share an L3 cache.
222213

223214
## Resources
224215
#### Libraries
225216
- [DeepSparse](https://docs.neuralmagic.com/deepsparse/)
226-
227217
- [SparseML](https://docs.neuralmagic.com/sparseml/)
228-
229218
- [SparseZoo](https://docs.neuralmagic.com/sparsezoo/)
230-
231219
- [Sparsify](https://docs.neuralmagic.com/sparsify/)
232220

233221

234222
#### Versions
235223
- [DeepSparse](https://pypi.org/project/deepsparse) | stable
236-
237224
- [DeepSparse-Nightly](https://pypi.org/project/deepsparse-nightly/) | nightly (dev)
238-
239225
- [GitHub](https://github.com/neuralmagic/deepsparse/releases) | releases
240226

241227
#### Info
242228

243229
- [Blog](https://www.neuralmagic.com/blog/)
244-
245230
- [Resources](https://www.neuralmagic.com/resources/)
246231

247232

@@ -252,16 +237,14 @@ Use Case: A workload that might benefit from the elastic scheduler is one in whi
252237

253238
Contribute with code, examples, integrations, and documentation as well as bug reports and feature requests! [Learn how here.](https://github.com/neuralmagic/deepsparse/blob/main/CONTRIBUTING.md)
254239

255-
For user help or questions about DeepSparse, sign up or log in to our **[Deep Sparse Community Slack](https://join.slack.com/t/discuss-neuralmagic/shared_invite/zt-q1a1cnvo-YBoICSIw3L1dmQpjBeDurQ)**. We are growing the community member by member and happy to see you there. Bugs, feature requests, or additional questions can also be posted to our [GitHub Issue Queue.](https://github.com/neuralmagic/deepsparse/issues) You can get the latest news, webinar and event invites, research papers, and other ML Performance tidbits by [subscribing](https://neuralmagic.com/subscribe/) to the Neural Magic community.
240+
For user help or questions about DeepSparse, sign up or log into our [Deep Sparse Community Slack](https://join.slack.com/t/discuss-neuralmagic/shared_invite/zt-q1a1cnvo-YBoICSIw3L1dmQpjBeDurQ). We are growing the community member by member and happy to see you there. Bugs, feature requests, or additional questions can also be posted to our [GitHub Issue Queue.](https://github.com/neuralmagic/deepsparse/issues) You can get the latest news, webinar and event invites, research papers, and other ML performance tidbits by [subscribing](https://neuralmagic.com/subscribe/) to the Neural Magic community.
256241

257242
For more general questions about Neural Magic, complete this [form.](http://neuralmagic.com/contact/)
258243

259244
### License
260245

261246
DeepSparse Community is licensed under the [Neural Magic DeepSparse Community License.](https://github.com/neuralmagic/deepsparse/blob/main/LICENSE-NEURALMAGIC)
262-
Some source code, example files, and scripts included in the deepsparse GitHub repository or directory are licensed under the [Apache License Version 2.0](https://github.com/neuralmagic/deepsparse/blob/main/LICENSE) as noted.
263-
264-
[DeepSparse Enterprise](https://docs.neuralmagic.com/products/deepsparse-ent) requires a Trial License or [can be fully licensed](https://neuralmagic.com/legal/master-software-license-and-service-agreement/) for production, commercial applications.
247+
Some source code, example files, and scripts included in the DeepSparse GitHub repository or directory are licensed under the [Apache License Version 2.0](https://github.com/neuralmagic/deepsparse/blob/main/LICENSE), as noted.
265248

266249
### Cite
267250

0 commit comments

Comments
 (0)