TorchNeuron
Deepgram uses TorchNeuron TorchNeuron integrates with Neuron Compiler TorchNeuron supports torch.compile TorchNeuron provides a new native backend for PyTorch TorchNeuron replaces Optimum Neuron TorchNeuron uses the torch-neuronx package TorchNeuron uses the torch-neuronx package torch-neuronx is based on PyTorch XLA Transformers natively support TorchNeuron Amazon released TorchNeuron, its version of PyTorch that allows AI apps to be ported to Amazon’s AI chip Trainium. TorchNeuron allows AI apps to be ported to Amazon’s homegrown AI chip Trainium. TorchNeuron allows AI apps to be ported to AWS with no rewriting. i14 uses TorchNeuron Native TorchNeuron RUNS_ON AWS Trainium TorchNeuron INTEGRATES_WITH Neuron Kernel Interface PyTorch LOWERS_TO TorchNeuron TorchNeuron is the native PyTorch backend for AWS Trainium Ray Train Deep Learning Container has AWS OFI NCCL plugin in its stack· Entity summary generated by Graphiti
Relation timeline
BY VALID-FROM DATE- AWS Deep Learning Container (DLC)HASTorchNeuron
Ray Train Deep Learning Container has AWS OFI NCCL plugin in its stack
- TorchNeuronBACKEND FORAWS Trainium2
TorchNeuron is the native PyTorch backend for AWS Trainium
- TorchNeuronRUNS ONAWS Trainium2
TorchNeuron RUNS_ON AWS Trainium
- TorchNeuronINTEGRATES WITHNeuron Kernel Interface
TorchNeuron INTEGRATES_WITH Neuron Kernel Interface
- PyTorchLOWERS TOTorchNeuron
PyTorch LOWERS_TO TorchNeuron
- i14USESTorchNeuron
i14 uses TorchNeuron Native
- TorchNeuronPORTS TOAWS Trainium2
TorchNeuron allows AI apps to be ported to Amazon’s homegrown AI chip Trainium.
- TorchNeuronRUNS ONAmazon Web Services
TorchNeuron allows AI apps to be ported to AWS with no rewriting.
- AmazonRELEASESTorchNeuron
Amazon released TorchNeuron, its version of PyTorch that allows AI apps to be ported to Amazon’s AI chip Trainium.
- TorchNeuronINTEGRATES WITHNeuron Compiler
TorchNeuron integrates with Neuron Compiler
- DeepgramUSEDTorchNeuron
Deepgram uses TorchNeuron
- TorchNeuronSUPPORTStorch.compile
TorchNeuron supports torch.compile
- TransformersSUPPORTSTorchNeuron
Transformers natively support TorchNeuron
- TorchNeuronREPLACESApate.ai
TorchNeuron replaces Optimum Neuron
- TorchNeuronUSES PACKAGETorchNeuron
TorchNeuron uses the torch-neuronx package
- TorchNeuronBASED ONTrezor Suite
torch-neuronx is based on PyTorch XLA
- TorchNeuronPROVIDES NATIVE BACKEND FORPyTorch
TorchNeuron provides a new native backend for PyTorch
Relations are facts Graphiti extracted from reports, not verbatim quotes. An end date means later reporting showed the relation no longer holds.
Events 6
- Distributed training with Ray Train DLCs and EFA on EKS1 cited report · 2026-10-29
- Does Your Model Run on AWS Trainium? Look It Up2 cited reports · 2026-10-23
- Profiling TorchNeuron (Part 1): System-level profiling on AWS Trainium.1 cited report · 2026-10-22
- Beyond the KV cache: Training and serving i14’s Laplace 3B linear attention model on AWS Trainium2 with native PyTorch1 cited report · 2026-10-21
- Google has joined Amazon in attacking Nvidia’s stronghold with AI app developers1 cited report · 2026-10-21
- How Deepgram optimized voice AI at scale with AWS Trainium1 cited report · 2026-10-20
Source reports 7
- Distributed training with Ray Train DLCs and EFA on EKSAWS Machine Learning Blog · first-party · 2026-10-29
- Does Your Model Run on AWS Trainium? Look It UpHugging Face Blog · first-party · 2026-10-23
- Profiling TorchNeuron (Part 1): System-level profiling on AWS Trainium.Hugging Face Blog · first-party · 2026-10-22
- Beyond the KV cache: Training and serving i14’s Laplace 3B linear attention model on AWS Trainium2 with native PyTorchAWS Machine Learning Blog · first-party · 2026-10-21
- Google has joined Amazon in attacking Nvidia’s stronghold with AI app developersTechCrunch · AI · 2026-10-21
- How Deepgram optimized voice AI at scale with AWS TrainiumAWS Machine Learning Blog · first-party · 2026-10-20
- TorchNeuron: running HF models on AWS Trainium is easier than everHugging Face Blog · first-party · 2026-10-19