You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Jun 3, 2025. It is now read-only.
Copy file name to clipboardExpand all lines: src/content/details/faqs.mdx
+29-29Lines changed: 29 additions & 29 deletions
Original file line number
Diff line number
Diff line change
@@ -1,8 +1,8 @@
1
1
---
2
2
title: "FAQs"
3
3
metaTitle: "FAQs"
4
-
metaDescription: "FAQs for the DeepSparse product from Neural Magic"
5
-
index: 2000
4
+
metaDescription: "FAQs for the Neural Magic Platform"
5
+
index: 4000
6
6
---
7
7
8
8
# FAQs
@@ -11,22 +11,22 @@ index: 2000
11
11
12
12
**What is Neural Magic?**
13
13
14
-
Founded by a team of award-winning MIT computer scientists and funded by Amdocs, Andreessen Horowitz, Comcast Ventures, NEA, Pillar VC, and
15
-
Ridgeline Partners, Neural Magic is the creator and maintainer of the Deep Sparse Platform. It has several components, including the
16
-
[DeepSparse Engine,](/products/deepsparse) a CPU runtime that runs sparse models at GPU speeds. To enable companies the ability to use
17
-
ubiquitous and unconstrained CPU resources, Neural Magic includes [SparseML](/products/sparseml) and the [SparseZoo,](/products/sparsezoo)
18
-
open-sourced model optimization technologies that allow users to achieve performance breakthroughs, at scale, with all the flexibility of software.
14
+
Neural Magic was founded by a team of award-winning MIT computer scientists and is funded by Amdocs, Andreessen Horowitz, Comcast Ventures, NEA, Pillar
15
+
VC, and Ridgeline Partners. The Neural Magic Platform includes several components, including [DeepSparse,](/products/deepsparse), [SparseML]
16
+
(/products/sparseml), and [SparseZoo[(/products/sparsezoo). DeepSparse is an inference runtime offering GPU-class performance on CPUs and tooling to
17
+
integrate ML into your application. [SparseML](/products/sparseml) and [SparseZoo,](/products/sparsezoo) are and open-source tooling and model repository
18
+
combination that enable you to create an inference-optimized sparse-model for deployment with DeepSparse.
19
19
20
-
**What is the DeepSparse Engine?**
20
+
Together, these components remove the tradeoff between performance and the simplicity and scalability of software-delivered deployments.
21
21
22
-
The DeepSparse Engine, created by Neural Magic, is a general purpose engine for machine learning, enabling machine learning to be practically
23
-
run in new places, on new kinds of workloads. It delivers state of art, GPU-class performance for the deep learning applications running on x86
24
-
CPUs. The DeepSparse Engine achieves its performance using breakthrough algorithms that reduce the computation needed for neural network execution
25
-
and accelerate the resulting memory-bound computation.
22
+
**What is DeepSparse?**
23
+
24
+
DeepSparse, created by Neural Magic, is an inference runtime for deep learning models. It delivers state of art, GPU-class performance on commodity CPUs
25
+
as well as tooling for integrating a model into an application and monitoring models in production.
26
26
27
27
**Why Neural Magic?**
28
28
29
-
Learn more about Neural Magic and the DeepSparse Engine (formerly known as the Neural Magic Inference Engine).
29
+
Learn more about Neural Magic and DeepSparse (formerly known as the Neural Magic Inference Engine).
30
30
[Watch the Why Neural Magic video](https://youtu.be/zJy_8uPZd0o)
31
31
32
32
**How does Neural Magic make it work?**
@@ -44,8 +44,8 @@ for our end users to train and infer on for their deep learning needs, and have
44
44
Our inference engine supports all versions of TensorFlow <=2.0;supportfortheKerasAPIisthroughTensorFlow2.0.
Copy file name to clipboardExpand all lines: src/content/details/glossary.mdx
+1-1Lines changed: 1 addition & 1 deletion
Original file line number
Diff line number
Diff line change
@@ -124,7 +124,7 @@ The machine learning community includes a vast array of terminology that can hav
124
124
</tr>
125
125
<tr>
126
126
<td>Unstructured pruning</td>
127
-
<td>A method for compressing a neural network. Unstructured pruning removes individual weight connections from a trained network. Software like Neural Magic's DeepSparse Engine runs these pruned networks faster.</td>
127
+
<td>A method for compressing a neural network. Unstructured pruning removes individual weight connections from a trained network. Software like Neural Magic's DeepSparse runs these pruned networks faster.</td>
Copy file name to clipboardExpand all lines: src/content/get-started/deploy-a-model/cv-object-detection.mdx
+4-4Lines changed: 4 additions & 4 deletions
Original file line number
Diff line number
Diff line change
@@ -9,9 +9,9 @@ index: 2000
9
9
10
10
This page walks through an example of deploying an object detection model with DeepSparse Server.
11
11
12
-
The DeepSparse Server is a server wrapper around `Pipelines`, including the object detection pipeline. As such,
12
+
DeepSparse Server is a server wrapper around `Pipelines`, including the object detection pipeline. As such,
13
13
the server provides and HTTP interface that accepts images and image files as inputs and outputs the labeled predictions.
14
-
With all of this built on top of the DeepSparse Engine, the simplicity of servable pipelines is combined with GPU-class performance on CPUs for sparse models.
14
+
In this way, DeepSparse combines the simplicity of servable pipelines with GPU-class performance on CPUs for sparse models.
15
15
16
16
## Install Requirements
17
17
@@ -20,12 +20,12 @@ This example requires [DeepSparse Server+YOLO Install](/get-started/install/deep
20
20
## Start the Server
21
21
22
22
Before starting the server, the model must be set up in the format expected for DeepSparse `Pipelines`.
23
-
See an example of how to setup `Pipelines` in the [Try a Model](../../try-a-model) section.
23
+
See an example of how to setup `Pipelines` in the [Use a Model](../../use-a-model) section.
24
24
25
25
Once the `Pipelines` are set up, the `deepsparse.server` command launches a server with the model at `--model_path` inside. The `model_path` can either
26
26
be a SparseZoo stub or a path to a local `model.onnx` file.
27
27
28
-
The command below shows how to start up the DeepSparse Server for a sparsified YOLOv5l model trained on the COCO dataset from the SparseZoo.
28
+
The command below shows how to start up DeepSparse Server for a sparsified YOLOv5l model trained on the COCO dataset from the SparseZoo.
29
29
The output confirms the server was started on port `:5543` with a `/docs` route for general info and a `/predict/from_files` route for inference.
Copy file name to clipboardExpand all lines: src/content/get-started/deploy-a-model/nlp-text-classification.mdx
+4-4Lines changed: 4 additions & 4 deletions
Original file line number
Diff line number
Diff line change
@@ -9,9 +9,9 @@ index: 1000
9
9
10
10
This page walks through an example of deploying a text-classification model with DeepSparse Server.
11
11
12
-
The DeepSparse Server is a server wrapper around `Pipelines`, including the sentiment analysis pipeline. As such,
12
+
DeepSparse Server is a server wrapper around `Pipelines`, including the sentiment analysis pipeline. As such,
13
13
the server provides an HTTP interface that accepts raw text sequences as inputs and responds with the labeled predictions.
14
-
With all of this built on top of the DeepSparse Engine, the simplicity of servable pipelines is combined with GPU-class performance on CPUs for sparse models.
14
+
In this way, DeepSparse combines the simplicity of servable pipelines with GPU-class performance on CPUs for sparse models.
15
15
16
16
## Install Requirements
17
17
@@ -20,12 +20,12 @@ This example requires [DeepSparse Server Install](/get-started/install/deepspars
20
20
## Start the Server
21
21
22
22
Before starting the server, the model must be set up in the format expected for DeepSparse `Pipelines`.
23
-
See an example of how to set up `Pipelines` in the [Try a Model](../../try-a-model) section.
23
+
See an example of how to set up `Pipelines` in the [Use a Model](../../use-a-model) section.
24
24
25
25
Once the `Pipelines` are set up, the `deepsparse.server` command launches a server with the model at `--model_path` inside. The `model_path` can either
26
26
be a SparseZoo stub or a local model path.
27
27
28
-
The command below starts up the DeepSparse Server for a sparsified DistilBERT model (from the SparseZoo) trained on the SST2 dataset for sentiment analysis.
28
+
The command below starts up DeepSparse Server for a sparsified DistilBERT model (from the SparseZoo) trained on the SST2 dataset for sentiment analysis.
29
29
The output confirms the server was started on port `:5543` with a `/docs` route for general info and a `/predict` route for inference.
Copy file name to clipboardExpand all lines: src/content/get-started/install/deepsparse-ent.mdx
+12-8Lines changed: 12 additions & 8 deletions
Original file line number
Diff line number
Diff line change
@@ -1,17 +1,21 @@
1
1
---
2
2
title: "DeepSparse Enterprise"
3
3
metaTitle: "DeepSparse Enterprise Installation"
4
-
metaDescription: "Installation instructions for the DeepSparse Engine enabling performant neural network deployments"
4
+
metaDescription: "Installation instructions for DeepSparse enabling performant neural network deployments"
5
5
index: 2000
6
6
---
7
7
8
-
# DeepSparse Enterprise Edition Installation
8
+
# DeepSparse Enterprise Installation
9
9
10
-
The [DeepSparse Engine](/products/deepsparse-ent) enables GPU-class performance on CPUs, leveraging sparsity within models to reduce FLOPs and the unique cache hierarchy on CPUs to reduce memory movement.
11
-
The engine accepts models in the open-source [ONNX format](https://onnx.ai/), which are easily created from PyTorch and TensorFlow models.
10
+
[DeepSparse Enterprise](/products/deepsparse-ent) enables GPU-class performance on CPUs.
12
11
13
-
Currently, DeepSparse is tested on Python 3.7-3.10, ONNX 1.5.0-1.10.1, ONNX opset version 11+ and is [manylinux compliant](https://peps.python.org/pep-0513/).
14
-
It is limited to Linux systems running on x86 CPU architectures.
12
+
Currently, DeepSparse is tested on Python 3.7-3.10, ONNX 1.5.0-1.10.1, ONNX opset version 11+, and [manylinux compliant systems](https://peps.python.org/pep-0513/).
13
+
14
+
We currently support x86 CPU architectures.
15
+
16
+
DeepSparse is available in two versions:
17
+
1.[**DeepSparse Community**](/products/deepsparse) is free for evaluation, research, and non-production use with our [DeepSparse Community License](https://neuralmagic.com/legal/engine-license-agreement/).
18
+
2.[**DeepSparse Enterprise**](/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.
15
19
16
20
## Installing DeepSparse Enterprise
17
21
@@ -23,7 +27,7 @@ pip install deepsparse-ent
23
27
24
28
## Installing the Server
25
29
26
-
The [DeepSparse Server](/use-cases/deploying-deepsparse/deepsparse-server) allows you to serve models and pipelines through an HTTP interface using the deepsparse.server CLI.
30
+
[DeepSparse Server](/user-guide/deploying-deepsparse/deepsparse-server) allows you to serve models and pipelines through an HTTP interface using the deepsparse.server CLI.
27
31
To install, use the following extra option:
28
32
29
33
```bash
@@ -37,6 +41,6 @@ To use YOLO models, install with the following extra option:
37
41
38
42
```bash
39
43
pip install deepsparse-ent[yolo] # just yolo requirements
40
-
pip install deepsparse-ent[yolo,server] # both yolo + server requirements
44
+
pip install deepsparse-ent[yolo,server] # both yolo + server requirements
0 commit comments