Skip to content

Releases: pytorch/serve

TorchServe v0.12.0 Release Notes

30 Sep 22:46
@agunapal agunapal
v0.12.0
6bdb1ba
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Learn about vigilant mode.
Compare
Choose a tag to compare
TorchServe v0.12.0 Release Notes Latest
Latest

Highlights Include

  • GenAI updates
    • No code LLM deployments with TorchServe + vLLM & TensorRT-LLM using ts.llm_launcher script
    • OpenAI API support for TorchServe + vLLM
    • Integration of TensorRT-LLM engine
    • Stateful Inference on AWS Sagemaker (see blog)
  • Support for linux-aarch64
    • CI & nightly regression added
    • Publish docker & KServe images
  • PyTorch updates
    • Support for PyTorch 2.4
    • Deprecation of TorchText

PyTorch Updates

GenAI

  • Implement stateful inference session timeout by @namannandan in #3263
  • Use Case: Enhancing LLM Serving with Torch Compiled RAG on AWS Graviton by @agunapal in #3276
  • Feature add openai api for vllm integration by @mreso in #3287
  • Set vllm multiproc method to spawn by @mreso in #3310
  • TRT LLM Integration with LORA by @agunapal in #3305
  • Bump vllm from 0.5.0 to 0.5.5 in /examples/large_models/vllm by @dependabot in #3321
  • Use startup time in async worker thread instead of worker timeout by @mreso in #3315
  • Rename vllm dockerfile by @mreso in #3330

Support for linux-aarch64

Documentation

Improvements and Bug Fixing

New Contributors

Platform Support

Ubuntu 20.04 MacOS 10.14+, Windows 10 Pro, Windows Server 2019, Windows subsystem for Linux (Windows Server 2019, WSLv1, Ubuntu 18.0.4). TorchServe requires Python >= 3.8 and JDK17.

GPU Support Matrix

TorchServe version PyTorch version Python Stable CUDA Experimental CUDA
0.12.0 2.4.0 >=3.8, <=3.11 CUDA 11.8, CUDNN 8.7.0.84 CUDA 12.1, CUDNN 8.9.2.26
0.11.1 2.3.0 >=3.8, <=3.11 CUDA 11.8, CUDNN 8.7.0.84 CUDA 12.1, CUDNN 8.9.2.26
0.11.0 2.3.0 >=3.8, <=3.11 CUDA 11.8, CUDNN 8.7.0.84 CUDA 12.1, CUDNN 8.9.2.26
0.10.0 2.2.1 >=3.8, <=3.11 CUDA 11.8, CUDNN 8.7.0.84 CUDA 12.1, CUDNN 8.9.2.26
0.9.0 2.1 >=3.8, <=3.11 CUDA 11.8, CUDNN 8.7.0.84 CUDA 12.1, CUDNN 8.9.2.26
0.8.0 2.0 >=3.8, <=3.11 CUDA 11.7, CUDNN 8.5.0.96 CUDA 11.8, CUDNN 8.7.0.84
0.7.0 1.13 >=3.7, <=3.10 CUDA 11.6, CUDNN 8.3.2.44 CUDA 11.7, CUDNN 8.5.0.96

Inferentia2 Support Matrix

TorchServe version PyTorch version Python Neuron SDK
0.12.0 2.1 >=3.8, <=3.11 2.18.2+
0.11.1 2.1 >=3.8, <=3.11 2.18.2+
0.11.0 2.1 >=3.8, <=3.11 2.18.2+
0.10.0 1.13 >=3.8, <=3.11 2.16+
0.9.0 1.13 >=3.8, <=3.11 2.13.2+

Contributors

slashvar, tadayosi, and 16 other contributors
Assets 15

TorchServe v0.11.1 Release Notes

18 Jul 18:56
@agunapal agunapal
v0.11.1
9a0ab89
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Learn about vigilant mode.
Compare
Choose a tag to compare
TorchServe v0.11.1 Release Notes

This is the release of TorchServe v0.11.1.

Highlights Include

  • Security Updates
    • Token Authorization: TorchServe enforces token authorization by default which requires the correct token to be provided when calling a HTTP/S or gRPC API. This is a security feature which addresses the concern of unauthorized API calls. This is applicable in the scenario where an unauthorized user may try to access a running TorchServe instance. The default behavior is to enable this feature which creates a key file with the appropriate tokens to be used for API calls. Users have the option to disable this feature to prevent token authorization from being required for API calls. For more details, refer to the token authorization documentation: https://github.com/pytorch/serve/blob/master/docs/token_authorization_api.md
    • Model API Control: TorchServe disables the ability to register and delete models using HTTP/S or gRPC API calls by default once TorchServe is running. This is a security feature which addresses the concern of unintended registration and deletion of models once TorchServe has started. This is applicable in the scenario where a user may upload malicious code to the model server in the form of a model or where a user may delete a model that is being used. The default behavior prevents users from registering or deleting models once TorchServe is running. Model API control can be enabled to allow users to register and delete models using the TorchServe model load and delete APIs. For more details, refer to the model API control documentation: https://github.com/pytorch/serve/blob/master/docs/model_api_control.md
  • PyTorch 2.x updates
    • Standardized torch.compile configuration
    • Added examples for tensorrt & hpu backends
  • GenAI updates
    • Support continuous batching in sequence batch streaming
    • Asynchronous backend worker communication for continuous batching
    • No code LLM deployment
  • Support for Intel GPUs

Security Updates

  • Adding model-control-mode by @udaij12 in #3165
  • Enable Token Authorization by default by @udaij12 in #3163
  • Updating night CIs to account for model control and token auth by @udaij12 in #3188
  • Adding token auth and model api to workflow and https by @udaij12 in #3234
  • Enable token authorization and model control for gRPC by @namannandan in #3238

PyTorch 2.x Updates

GenAI

  • Support continuous batching in sequence batch streaming case by @lxning in #3160
  • GPT-FAST-MIXTRAL-MOE integration by @alex-kharlamov in #3151
  • clean a jobGroup immediately when it finished by @lxning in #3222
  • Asynchronous worker communication and vllm integration by @mreso in #3146
  • Add single command LLM deployment by @mreso in #3209
  • TensorRT-LLM Engine integration by @agunapal in #3228
  • Adds torch.compile documentation to alexnet example readme by @crmdias in #3227

Support for Intel GPUs

Documentation

Improvements and Bug Fixing

New Contributors

Platform Support

Ubuntu 20.04 MacOS 10.14+, Windows 10 Pro, Windows Server 2019, Windows subsystem for Linux (Windows Server 2019, WSLv1, Ubuntu 18.0.4). TorchServe requires Python >= 3.8 and JDK17.

GPU Support Matrix

TorchServe version PyTorch version Python Stable CUDA Experimental CUDA
0.11.1 2.3.0 >=3.8, <=3.11 CUDA 11.8, CUDNN 8.7.0.84 CUDA 12.1, CUDNN 8.9.2.26
0.11.0 2.3.0 >=3.8, <=3.11 CUDA 11.8, CUDNN 8.7.0.84 CUDA 12.1, CUDNN 8.9.2.26
0.10.0 2.2.1 >=3.8, <=3.11 CUDA 11.8, CUDNN 8.7.0.84 CUDA 12.1, CUDNN 8.9.2.26
0.9.0 2.1 >=3.8, <=3.11 CUDA 11.8, CUDNN 8.7.0.84 CUDA 12.1, CUDNN 8.9.2.26
0.8.0 2.0 >=3.8, <=3.11 CUDA 11.7, CUDNN 8.5.0.96 CUDA 11.8, CUDNN 8.7.0.84
0.7.0 1.13 >=3.7, <=3.10 CUDA 11.6, CUDNN 8.3.2.44 CUDA 11.7, CUDNN 8.5.0.96

Inferentia2 Support Matrix

TorchServe version PyTorch version Python Neuron SDK
0.11.1 2.1 >=3.8, <=3.11 2.18.2+
0.11.0 2.1 >=3.8, <=3.11 2.18.2+
0.10.0 1.13 >=3.8, <=3.11 2.16+
0.9.0 1.13 >=3.8, <=3.11 2.13.2+

Contributors

ijkilchenko, alex-kharlamov, and 13 other contributors
Assets 23
Loading

TorchServe v0.11.0 Release Notes

17 May 03:59
@lxning lxning
v0.11.0
34bc370
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Learn about vigilant mode.
Compare
Choose a tag to compare
TorchServe v0.11.0 Release Notes

This is the release of TorchServe v0.11.0.

Highlights Include

  • GenAI inference optimizations showcasing
    • torch.compile with OpenVINO backend for Stable Diffusion
    • Intel IPEX for Llama
  • Experimental support for Apple MPS and linux-aarch64
  • Security bug fixing

GenAI

  • Upgraded LLama2 examples to Llama3
    • Supported Llama3 in HuggingFace Accelerate Example #3108 @mreso
    • Supported Llama3 in chat bot #3131 @mreso
    • Supported Llama3 on inf2 Neuronx transformer using continuous batching or micro batching #3133 #3035 @lxning
  • Examples for LoRA and Mistral #3077 @lxning
  • IPEX LLM serving example with Intel AMX #3068 @bbhattar
  • Integration of Intel Openvino with TorchServe using torch.compile. Example showcase of openvino torch.compile backend with Stable Diffusion #3116 @suryasidd
  • Enabling retrieval of guaranteed sequential order of input sequences with low latency for stateful inference via HTTP extending this previously gRPC-only feature #3142 @lxning

Linux aarch64 Support:

TorchServe adds support for linux-aarch64 and shows an example working on AWS Graviton. This provides users with a new platform alternative for serving models on CPU.

Apple Silicon Support:

XGBoost Support:

With the XGBoost Classifier example, we show how to deploy any pickled model with TorchServe.

Security

The ability to bypass allowed_urls using relative paths has been fixed by ensuring preemptive check for relative paths prior to copying the model archive to the model store directory. Also, the default gRPC inference and management addresses are now set to localhost(127.0.0.1) to reduce scope of default access to gRPC endpoints.

C++ Backend

Documentation

Improvements and Bug Fixing

Platform Support

Ubuntu 20.04 MacOS 10.14+, Windows 10 Pro, Windows Server 2019, Windows subsystem for Linux (Windows Server 2019, WSLv1, Ubuntu 18.0.4). TorchServe now requires Python 3.8 and above, and JDK17.

GPU Support Matrix

TorchServe version PyTorch version Python Stable CUDA Experimental CUDA
0.11.0 2.3.0 >=3.8, <=3.11 CUDA 11.8, CUDNN 8.7.0.84 CUDA 12.1, CUDNN 8.9.2.26
0.10.0 2.2.1 >=3.8, <=3.11 CUDA 11.8, CUDNN 8.7.0.84 CUDA 12.1, CUDNN 8.9.2.26
0.9.0 2.1 >=3.8, <=3.11 CUDA 11.8, CUDNN 8.7.0.84 CUDA 12.1, CUDNN 8.9.2.26
0.8.0 2.0 >=3.8, <=3.11 CUDA 11.7, CUDNN 8.5.0.96 CUDA 11.8, CUDNN 8.7.0.84
0.7.0 1.13 >=3.7, <=3.10 CUDA 11.6, CUDNN 8.3.2.44 CUDA 11.7, CUDNN 8.5.0.96

Inferentia2 Support Matrix

TorchServe version PyTorch version Python Neuron SDK
0.11.0 2.1 >=3.8, <=3.11 2.18.2+
0.10.0 1.13 >=3.8, <=3.11 2.16+
0.9.0 1.13 >=3.8, <=3.11 2.13.2+
Assets 19
Loading

TorchServe v0.10.0 Release Notes

15 Mar 00:03
@lxning lxning
v0.10.0
4a630cb
Compare
Choose a tag to compare
TorchServe v0.10.0 Release Notes

This is the release of TorchServe v0.10.0.

Highlights include

  • Extended support for PyTorch 2.x inference
  • C++ backend
  • GenAI fast series torch.compile showcase examples
  • Token authentication support for enhanced security.

C++ Backend

TorchServe presented the experimental C++ backend at the PyTorch Conference 2022. Similar to the Python backend, C++ backend also runs as a process and utilizes the BaseHandler to define APIs for customizing the handler. By providing a backend and handler written in pure C++ for TorchServe, it is now possible to deploy PyTorch models without any Python overhead. This release officially promoted the experimental branch to the master branch and included additional examples and Docker images for development.

torch.compile

With the launch of PT2 Inference at the PyTorch Conference 2023, we have added several key examples showcasing out-of-box speedups for torch.compile and AOT Compile. Since there is no new development being done in TorchScript, starting this release, TorchServe is preparing the migration path for customers to switch from TorchScript to torch.compile.

GenAI torch.compile series

The fast series GenAI models - GPTFast, SegmentAnythingFast, DiffusionFast with 3-10x speedups using torch.compile and native PyTorch optimizations:

Cold start problem solution

To address cold start problems, there is an example included to show how torch._export.aot_load (experimental API) can be used to load a pre-compiled model. TorchServe has also started benchmarking models with torch.compile and tracking their performance compared to TorchScript.

The new TorchServe C++ backend also includes torch.compile and AOTInductor related examples for ResNet50, BERT and Llama2.

  1. torch.compile
    a. Example torch.compile with image classifier model densenet161 #2915 @agunapal
    b. Example torch._export.aot_compile with image classification model ResNet-18 #2832 #2906 #2932 #2948 @agunapal
    c. Example torch inductor fx graph caching with image classification model densenet161 #2925 @agunapal

  2. C++ AOTInductor
    a. Example AOT Inductor with Llama2 #2913 @mreso
    b. Example AOT Inductor with ResNet-50 #2944 @lxning
    c. Example AOT Inductor with BERTSequenceClassification #2931 @lxning

Gen AI

  • Supported sequence batching for stateful inference in gRPC bi-directional streaming #2513 @lxning
  • The fast series Gen AI models using torch.compile and native PyTorch optimizations.
  • Example Mistral 7B with vLLM #2781 @agunapal
  • Example PyTorch native tensor parallel with Llama2 with continuous batching #2709 @mreso @HamidShojanazeri
  • Supported inf2 Neuronx transformer continuous batching for both no coding style and advanced customers with Llama2-70B example #2803 #3016 @lxning
  • Example deepspeed mii fastgen with Llama2-13B #2779 @lxning

Security

TorchServe has implemented token authentication for management and inference APIs. This is an optional config and can be enabled using torchserve-endpoint-plugin. This plugin can be downloaded from maven. This further strengthens TorchServe’s capability as a secure model serving solution. The security features of TorchServe are documented here

Apple Silicon Support

TorchServe is now supported on Apple Silicon mac. The current support is for CPU only. We have also posted an RFC for the deprecation of x86 mac support.

KServe Updates

While serving large models, model loading can take some time even though the pod is running. Even though TorchServe is up, the worker is not ready till the model is loaded. To address this, TorchServe now sets the model ready status in KServe after the model has been loaded on workers. TorchServe also includes native open inference protocol support in gRPC. This is an experiment feature.

  • Supported native KServe open inference protocol in gRPC #2609 @andyi2it
  • Refactored TorchServe configuration in KServe #2995 @sgaist
  • Improved KServe protocol version handling #2957 @sgaist
  • Updated KServe test script to return model version #2973 @agunapal
  • Set model status using TorchServe API in KServe #1878 @byeongjokim
  • Supported no-archive model archiver in KServe #2839 @agunapal
  • How to deploy MNIST using KServe with minikube #2718 @agunapal
  • Changes to support no-model archive mode with KServe #2839 @agunpal

Metrics Updates

In order to extend backwards compatibility support for metrics, auto-detection of backend metrics enables the flexibility to publish custom model metrics without having to explicitly specify them in the metrics configuration file. Furthermore, a customized script to collect system metrics is also now supported.

Improvements and Bug Fixing

Documentation

Platform Support

Ubuntu 20.04 MacOS 10.14+, Windows 10 Pro, Windows Server 2019, Windows subsystem for Linux (Windows Server 2019, WSLv1, Ubuntu 18.0.4). TorchServe now requires Python 3.8 and above, and JDK17.

GPU Support Matrix

TorchServe version PyTorch version Python Stable CUDA Experimental CUDA
0.10.0 2.2.1 >=3.8, <=3.11 CUDA 11.8, CUDNN 8.7.0.84 CUDA 12.1, CUDNN 8.9.2.26
0.9.0 2.1 >=3.8, <=3.11 CUDA 11.8, CUDNN 8.7.0.84 CUDA 12.1, CUDNN 8.9.2.26
0.8.0 2.0 >=3.8, <=3.11 CUDA 11.7, CUDNN 8.5.0.96 CUDA 11.8, CUDNN 8.7.0.84
0.7.0 1.13 >=3.7, <=3.10 CUDA 11.6, CUDNN ...
Read more

Contributors

sgaist, nathanweeks, and 20 other contributors
Assets 19
Loading

TorchServe v0.9.0 Release Notes

13 Oct 00:21
@lxning lxning
v0.9.0
db47936
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.
Compare
Choose a tag to compare
TorchServe v0.9.0 Release Notes

This is the release of TorchServe v0.9.0.

Security

Our security process is documented here

We rely heavily on automation to improve the security of torchserve namely by

  1. On a monthly basis updating our gradle and pip dependencies
  2. Docker scanning via Snyk
  3. Code analysis via CodeQL

A key point to remember is that torchserve will allow you to configure things in an unsecure way so make sure to read our security docs and relevant security warnings to make sure your product is secure in production. In general we do not encourage you to download untrusted mar files from the internet, running a .mar file effectively is running arbitrary python code so make sure to unzip mar files and validate whether they are doing anything suspicious.

Code scanning fixes

  1. Used Sha-256 in ziputils #2629 @msaroufim
  2. Verified default hostname in Test #2631 @msaroufim
  3. Fixed zip slip error #2634 @msaroufim
  4. Used string array as Process arguments input #2632 #2635 @msaroufim
  5. Enabled Netty HTTP header validation as default #2630 @msaroufim
  6. Verified 3rd party package installation path #2687 @lxning
  7. Allowed url validation #2685 @lxning including
  • Disabled loading TS_ALLOWED_URLS from env by default.
  • Moved the model url validation to last step.
  • Sanity check model archive name to guard Uncontrolled data used in path expression

Address configuration updates

  1. Updated default address from 0.0.0.0 to 127.0.0.1 #2624 #2704 @namannandan @agunapal
  2. Bind container ports to localhost ports #2646 @namannandan

Documentation improvements

  1. Updated security readme #2643 #2690 @msaroufim @agunapal
  2. Updated security guidance in docker readme #2669 @agunapal

Dependency improvements

  1. Created dependabot.yml #2642 #2675 @msaroufim
  2. Bumped packaging from 23.1 to 23.2
  3. Bumped pygit2 from 1.21.1 to 1.13.1
  4. Bumped com.github.spotbugs from 4.0.2 to 5.1.3
  5. Bumped ONNX from 1.14.0 to 1.14.1
  6. Bumped Pillow from 9.3.0 to 10.0.1
  7. Bumped Bump com.amazonaws:DynamoDBLocal from 1.13.2 to 2.0.0
  8. Upgraded node to version 18 #2663 @agunapal

Blogs

New Features

New Examples

  1. Deploy Llama2 on Inferentia2 #2458 @namannandan
  2. Using TorchServe on SageMaker Inf2.24xlarge with Llama2-13B @lxning
  3. PyTorch tensor parallel on Llama2 example #2623 #2689 @HamidShojanazeri
  4. Enabled better transformer (ie. flash attention 2) on Llama2 #2700 @HamidShojanazeri @lxning
  5. Llama2 Chatbot on Mac #2618 @agunapal
  6. ASR speech recognition example #2047 @husenzhang

Improvements

Documentation

Platform Support

Ubuntu 16.04, Ubuntu 18.04, Ubuntu 20.04 MacOS 10.14+, Windows 10 Pro, Windows Server 2019, Windows subsystem for Linux (Windows Server 2019, WSLv1, Ubuntu 18.0.4). TorchServe now requires Python 3.8 and above, and JDK17.

GPU Support

Torch 2.1.0 + Cuda 11.8, 12.1
Torch 2.0.1 + Cuda 11.7
Torch 2.0.0 + Cuda 11.7
Torch 1.13 + Cuda 11.7
Torch 1.11 + Cuda 10.2, 11.3, 11.6
Torch 1.9.0 + Cuda 11.1
Torch 1.8.1 + Cuda 9.2

Contributors

abergmeier, msaroufim, and 14 other contributors
Assets 17
Loading

TorchServe v0.8.2 Release Notes

28 Aug 23:20
@lxning lxning
v0.8.2
04e0b37
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.
Compare
Choose a tag to compare
TorchServe v0.8.2 Release Notes

This is the release of TorchServe v0.8.2.

Security

  • Updated snakeyaml version to v2 #2523 @nskool
  • Added warning about model allowed urls when default value is applied #2534 @namannandan

Custom metrics backwards compatibility

  • add_metric is now backwards compatible with versions [< v0.6.1] but the default metric type is inferred to be COUNTER. If the metric is of a different type, it will need to be specified in the call to add_metric as follows:
    metrics.add_metric(name='GenericMetric', value=10, unit='count', dimensions=[...], metric_type=MetricTypes.GAUGE)
  • When upgrading from versions [v0.6.1 - v0.8.1] to v0.8.2, replace the call to add_metric with add_metric_to_cache.
  • All custom metrics updated in the custom handler will need to be included in the metrics configuration file for them to be emitted by Torchserve. This is shown here.
  • A detailed upgrade guide is included in the metrics documentation.

New Features

New Examples

  1. Example LLama v2 70B chat using HuggingFace Accelerate #2494 @lxning @HamidShojanazeri @agunapal

  2. large model example OPT-6.7B on Inferentia2 #2399 @namannandan

    • This example demonstrates how NeuronX compiles the model , detects neuron core availability and runs the inference.

  3. DeepSpeed deferred init with OPT-30B #2419 @agunapal

    • This PR added feature deferred model init in OPT-30B example by leveraging DeepSpeed new version. This feature is able to significantly reduce model loading latency.

  4. Torch TensorRT example #2483 @agunapal

    • This PR uses Resnet-50 as an example to demonstrate Torch TensorRT.

  5. K8S mnist example using minikube #2323 @agunapal

    • This example shows how to use a pre-trained custom MNIST model to performing real time Digit recognition via K8S.

  6. Example for custom metrics #2516 @namannandan

  7. Example for object detection with ultralytics YOLO v8 model #2508 @agunapal

Improvements

Documentation

Platform Support

Ubuntu 16.04, Ubuntu 18.04, Ubuntu 20.04 MacOS 10.14+, Windows 10 Pro, Windows Server 2019, Windows subsystem for Linux (Windows Server 2019, WSLv1, Ubuntu 18.0.4). TorchServe now requires Python 3.8 and above, and JDK17.

GPU Support

Torch 2.0.1 + Cuda 11.7, 11.8
Torch 2.0.0 + Cuda 11.7, 11.8
Torch 1.13 + Cuda 11.7, 11.8
Torch 1.11 + Cuda 10.2, 11.3, 11.6
Torch 1.9.0 + Cuda 11.1
Torch 1.8.1 + Cuda 9.2

Contributors

bryanwweber, namannandan, and 13 other contributors
Assets 12
Loading

TorchServe v0.8.1 Release Notes

14 Jun 23:54
@lxning lxning
v0.8.1
c2cdcfb
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.
Compare
Choose a tag to compare
TorchServe v0.8.1 Release Notes

This is the release of TorchServe v0.8.1.

New Features

  1. Supported microbatch in handler to parallel process a batch request from frontend. #2210 @mreso

Because pre- and post- processing are often carried out on the CPU the GPU sits idle until the two CPU bound steps are executed and the worker receives a new batch. Microbatch in handler is able to parallel process inference, pre- and post- processing for a batch request from frontend.

  1. Supported job ticket #2350 @lxning

This feature help with use cases where inference latency can be high, such as generative models, auto regressive decoder models like chatGPT. Applications can take effective actions, for example, routing the rejected request to a different server, or scaling up model server capacity, based on the business requirements.

  1. Supported job queue size configuration per model #2350 @lxning

New Examples

This example demonstrates creative content assisted by generative AI by using TorchServe on SageMaker MME.

Improvements

  • Upgraded to PyTorch 2.0.1 #2374 @namannandan

  • Significant reduction in Docker Image Size

    • Reduce GPU docker image size by 3GB #2392 @agunapal
    • Reduced dependency installation time and decrease docker image size #2364 @mreso 
        GPU
  pytorch/torchserve   0.8.1-gpu   04eef250c14e   4 hours ago     2.34GB
  pytorch/torchserve   0.8.0-gpu   516bb13a3649   4 weeks ago     5.86GB
  pytorch/torchserve   0.6.0-gpu   fb6d4b85847d   12 months ago   2.13GB

        CPU
  pytorch/torchserve   0.8.1-cpu   68a3fcae81af   4 hours ago     662MB
  pytorch/torchserve   0.8.0-cpu   958ef6dacea2   4 weeks ago     2.37GB
  pytorch/torchserve   0.6.0-cpu   af91330a97bd   12 months ago   496MB

  • Updated CPU information for IPEX #2372 @min-jean-cho

  • Fixed inf2 example handler #2378 @namannandan

  • Added inf2 nightly benchmark #2283 @namannandan

  • Fixed archiver tgz format model directory structure mismatch on SageMaker #2405 @lxning

  • Fixed model archiver to fail if extra files are missing #2212 @mreso

  • Fixed device type setting in model config yaml #2408 @lxning

  • Fixed batchsize in config.properties not honored #2382 @lxning

  • Upgraded torchrun argument names and fixed backend tcp port connection #2377 @lxning

  • Fixed error thrown while loading multiple models in KServe #2235 @jagadeeshi2i

  • Fixed KServe fastapi migration issues #2175 @jagadeeshi2i

  • Added type annotation in model_server.py #2384 @josephcalise

  • Speed up unit test by removing sleep in start/stop torchserve #2383 @mreso

  • Removed cu118 from regression tests #2380 @agunapal

  • Enabled ONNX CI test #2363 @msaroufim

  • Removed session_mocker usage to prevent test cross talking #2375 @mreso

  • Enabled regression test in CI #2370 @msaroufim

  • Fixed regression test failures #2371 @namannandan

  • Bump up transformers version from 4.28.1 to 4.30.0 #2410

Documentation

Platform Support

Ubuntu 16.04, Ubuntu 18.04, Ubuntu 20.04 MacOS 10.14+, Windows 10 Pro, Windows Server 2019, Windows subsystem for Linux (Windows Server 2019, WSLv1, Ubuntu 18.0.4). TorchServe now requires Python 3.8 and above, and JDK17.

GPU Support

Torch 2.0.1 + Cuda 11.7, 11.8
Torch 2.0.0 + Cuda 11.7, 11.8
Torch 1.13 + Cuda 11.7, 11.8
Torch 1.11 + Cuda 10.2, 11.3, 11.6
Torch 1.9.0 + Cuda 11.1
Torch 1.8.1 + Cuda 9.2

Contributors

msaroufim, namannandan, and 7 other contributors
Assets 12
Loading

TorchServe v0.8.0 Release Notes

12 May 23:01
@lxning lxning
v0.8.0
35fb574
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.
Compare
Choose a tag to compare
TorchServe v0.8.0 Release Notes

This is the release of TorchServe v0.8.0.

New Features

  1. Supported large model inference in distributed environment #2193 #2320 #2209 #2215 #2310 #2218 @lxning @HamidShojanazeri

TorchServe added the deep integration to support large model inference. It provides PyTorch native large model inference solution by integrating PiPPy. It also provides the flexibility and extensibility to support other popular libraries such as Microsoft Deepspeed, and HuggingFace Accelerate.

  1. Supported streaming response for GRPC #2186 and HTTP #2233 @lxning

To improve UX in Generative AI inference, TorchServe allows for sending intermediate token response to client side by supporting GRPC server side streaming and HTTP 1.1 chunked encoding .

  1. Supported PyTorch/XLA on GPU and TPU #2182 @morgandu

By leveraging torch.compile it's now possible to run torchserve using XLA which is optimized for both GPU and TPU deployments.

  1. Implemented New Metrics platform #2199 #2190 #2165 @namannandan @lxning

TorchServe fully supports metrics in Prometheus mode or Log mode. Both frontend and backend metrics can be configured in a central metrics YAML file.

  1. Supported map based model config YAML file. #2193 @lxning

Added config-file option for model config to model archiver tool. Users is able to flexibly define customized parameters in this YAML file, and easily access them in backend handler via variable context.model_yaml_config. This new feature also made TorchServe easily support the other new features and enhancements.

  1. Refactored PT2.0 support #2222 @msaroufim

We've refactored our model optimization utilities, improved logging to help debug compilation issues. We've also now deprecated compile.json in favor of using the new YAML config format, follow our guide here to learn more https://github.com/pytorch/serve/blob/master/examples/pt2/README.md the main difference is while archiving a model instead of passing in compile.json via --extra-files we can pass in a --config-file model_config.yaml

  1. Supported user specified gpu deviceIds for a model #2193 @lxning

By default, TorchServe uses a round-robin algorithm to assign GPUs to a worker on a host. Starting from v0.8.0, TorchServe allows users to define deviceIds in the model_config.yaml. to assign GPUs to a model.

  1. Supported cpu model on a GPU host #2193 @lxning

TorchServe supports hybrid mode on a GPU host. Users are able to define deviceType in model config YAML file to deploy a model on CPU of a GPU host.

  1. Supported Client Timeout #2267 @lxning

TorchServe allows users to define clientTimeoutInMills in a model config YAML file. TorchServe calculates the expired timestamp of an incoming inference request if clientTimeoutInMills is set, and drops the request once it is expired.

  1. Updated ping endpoint default behavior #2254 @lxning

Supported maxRetryTimeoutInSec, which defines the max maximum time window of recovering a dead backend worker of a model, in model config YAML file. The default value is 5 min. Users are able to adjust it in model config YAML file. The ping endpoint returns 200 if all models have enough healthy workers (ie, equal or larger the minWorkers); otherwise returns 500.

New Examples

Improvements

TorchServe can be used with Intel® Extension for PyTorch* to give performance boost on Intel hardware. Intel® Extension for PyTorch* is a Python package extending PyTorch with up-to-date features optimizations that take advantage of AVX-512 Vector Neural Network Instructions (AVX512 VNNI), Intel® Advanced Matrix Extensions (Intel® AMX), and more.

dashboard

Enabling core pinning in TorchServe CPU nightly benchmark shows significant performance speedup. This feature is implemented via a script under PyTorch Xeon backend, initiated from Intel® Extension for PyTorch*. To try out core pinning on your workload, add cpu_launcher_enable=true in config.properties.

To try out more optimizations with Intel® Extension for PyTorch*, install Intel® Extension for PyTorch* and add ipex_enable=true in config.properties.

In case of OOM , return error code 507 instead of generic code 503

a). Added wildcard file search in model archiver --extra-file #2142 @gustavhartz
b). Added zip-store option to model archiver tool #2196 @mreso
c). Made model archiver tests runnable from any directory #2191 @mreso
d). Supported tgz format model decompression in TorchServe frontend #2214 @lxning

Automatically flag deviation of metrics from the average of last 30 runs

Dependency Upgrades

Documentation

This study compares TPS b/w TorchServe with Nvidia MPS enabled and TorchServe without Nvidia MPS enabled on P3 and G4. It can help to the decision in enabling MPS for your deployment or not.

Platform Support

Ubuntu 16.04, Ubuntu 18.04, Ubuntu 20.04 MacOS 10.14+, Windows 10 Pro, Windows Server 2019, Windows subsystem for Linux (Windows Server 2019, WSLv1, Ubuntu 18.0.4). TorchServe now requires Python 3.8 and above, and JDK17.

GPU Support

Torch 2.0.0 + Cuda 11.7, 11.8
Torch 1.13 + Cuda 11.7, 11.8
Torch 1.11 + Cuda 10.2, 11.3, 11.6
Torch 1.9.0 + Cuda 11.1
Torch 1.8.1 + Cuda 9.2

Contributors

tmc, msaroufim, and 12 other contributors
Assets 14
Loading

TorchServe v0.7.1 Release Notes

09 Feb 00:00
@lxning lxning
v0.7.1
5140113
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.
Compare
Choose a tag to compare
TorchServe v0.7.1 Release Notes

This is the release of TorchServe v0.7.1.

Security

Dependency Upgrades

Improvements

Documentation

Deprecation

Platform Support

Ubuntu 16.04, Ubuntu 18.04, Ubuntu 20.04 MacOS 10.14+, Windows 10 Pro, Windows Server 2019, Windows subsystem for Linux (Windows Server 2019, WSLv1, Ubuntu 18.0.4). TorchServe now requires Python 3.8 and above, and JDK17.

GPU Support

Torch 1.13 + Cuda 11.7
Torch 1.11 + Cuda 10.2, 11.3, 11.6
Torch 1.9.0 + Cuda 11.1
Torch 1.8.1 + Cuda 9.2

Contributors

blackrabbit, MasoudKaviani, and 8 other contributors
Assets 14
Loading

TorchServe v0.7.0 Release Notes

13 Dec 22:15
@lxning lxning
v0.7.0
7845403
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.
Compare
Choose a tag to compare
TorchServe v0.7.0 Release Notes

This is the release of TorchServe v0.7.0.

New Examples

Better Transformer / Flash Attention & Xformer Memory Efficient provides out of box performance with major speed ups for PyTorch Transformer encoders. This has been integrated into Torchserve HF Transformer example, please read more about this integration here.

Main speed ups in Better Transformers comes from exploiting sparsity on padded inputs and kernel fusions. As a result you would see the biggest gains when dealing with larger workloads, such sequences with longer paddings and larger batch sizes.

In our benchmarks on P3 instances with 4 V100 GPUs, using Torchserve benchmarking workloads, throughput has shown significant improvement with large batch sizes. 45.5% increase with batch size 8; 50.8% increase with batch size 16; 45.2% increase with batch size 32; 47.2% increase with batch size 64. and 17.2 increase with batch size 4. These number can vary based on your workload (batch size , padding percentage) and your hardware. Please look up some other benchmarks in the blog post.

We've added experimental support for PT 2.0 as in torch.compile() support within torchserve. To use it you need to supply a file compile.json when archiving your model to specify which backend you want. We've also enabled by default mode=reduce-overhead which is ideally suited for smaller batch sizes which are more common for inference. We recommend for now to leverage GPUs with tensor cores available like A10G or A100 since you're likely to see the greatest speedups there.

On training we've seen speedups ranging from 30% to 2x https://pytorch.org/get-started/pytorch-2.0/ but we haven't ran any performance benchmarks yet for inference. Until then we recommend you continue leveraging other runtimes like TensorRT or IPEX for accelerated inference which we highlight in our performance_guide.md. There are a few important caveats to consider when you're using torch.compile: changes in batch sizes will cause recompilations so make sure to leverage a small batch size, there will be additional overhead to start a model since you need to compile it first and you'll likely still see the largest speedups with TensorRT.

However, we hope that adding this support will make it easier for you to benchmark and try out PT 2.0. Learn more here https://github.com/pytorch/serve/tree/master/examples/pt2

Dependency Upgrades

Improvements

Documentation

Platform Support

Ubuntu 16.04, Ubuntu 18.04, MacOS 10.14+, Windows 10 Pro, Windows Server 2019, Windows subsystem for Linux (Windows Server 2019, WSLv1, Ubuntu 18.0.4). TorchServe now requires Python 3.8 and above, and JDK17.

GPU Support

Torch 1.13 + Cuda 11.7
Torch 1.11 + Cuda 10.2, 11.3, 11.6
Torch 1.9.0 + Cuda 11.1
Torch 1.8.1 + Cuda 9.2

Contributors

jagadeesh, msaroufim, and 4 other contributors
Assets 14
Loading