huggingface / optimum-neuron

Easy, fast and very cheap training and inference on AWS Trainium and Inferentia chips.

huggingface/optimum-neuron Issues

optimum-cli neuron consolidate -> OOM. Using only one Neuron Core
Updated 13 days ago2
The quality of generated image using NeuronStableDiffusionPipeline is very poor
Closed 16 days ago4
NeuronStableDiffusionXLPipeline is not available to run on inf2.xlarge if not using huggingface DLAMI
Updated 16 days ago
LLaVA support
Updated 22 days ago3
Add support to read different neuron cache from the local directory of the DLC based on ENVs
Updated 23 days ago8
optimum-neuron test results with torch-neuronx 2.1
Closed a month ago6
Python API - import NeuronStableDiffusionPipeline generates cannot import name 'retrieve_timesteps' errors
Closed a month ago1
Optimum Neuron Training Tests Failures
Updated a month ago1
Add optimum-neuron support for diffusers.StableDiffusionControlNetPipeline
Updated a month ago1
nvidia/Llama3-ChatQA-1.5-70B failing to start
Updated a month ago10
Bugs for TGI-NeuronX DLC for llama-2 70B on Trn1.32xlarge
Updated a month ago6
Clean up old AMIS
Updated a month ago2
T5 not running on summarization script
Updated a month ago2
Muti Node CLM Mistral training tutorial
Updated a month ago4
Request TGI NeuronX DLC to support Flan T5 models in SageMaker
Updated a month ago2
Unable to deploy TinyLlama in Amazon SageMaker using Optimum Neuron 0.0.20 w/ Neuronx 2.*
Updated a month ago4
caching diffusion models does not work
Closed a month ago3
Skip weight splitting for SafeTensors Weights
Updated a month ago3
Inf1: optimum neuron inference tests failing with import errors from transformers.utils and huggingface_hub
Closed a month ago5
Optimum neuron inference test results with NotImplementedError.
Updated 2 months ago4
sdxl compiled model load faild in sagemaker only (it works very well in EC2 instance)
Closed 2 months ago1
Poor performance to generate images with NeuronStableDiffusionPipeline
Closed 2 months ago3
Flan-UL2 compilation failure
Updated 2 months ago12
Add support for PyTorch 2.1
Closed 2 months ago
HF DL AMI v20240318 optimum-neuron install generates error on import
Closed 2 months ago4
Installation instructions are unclear
Closed 2 months ago3
optimum.neuron.modeling_decoder.get_available_cores returns a wrong result
Closed 2 months ago3
optimum-cli does not support exporting to Neuron by default in HF DL AMI v20240318
Updated 3 months ago
Failing Casual Language Modeling with Expanded Mistral-7B Model ( 1.75B Trainable Parameters)
Updated 3 months ago1
Optimum Neuron v 0.0.20 w/ Neuron 2.x taking too long to fine-tune TinyLlama in Amazon SageMaker
Updated 3 months ago2
Allow queueing in Neuron X TGI server beyond batch_size
Closed 3 months ago3
OpenSSL error during boto3 import
Closed 3 months ago1
TGI NeuronX DLC (Optimum-neuron) 0.0.20: SageMaker deployment failure with llama-2 7B
Updated 3 months ago14
RuntimeError: Failed to import optimum.neuron.modeling because of the following error (look up to see its traceback): No module named 'optimum.exporters.neuron'
Closed 3 months ago
Gradient Accumulation
Updated 3 months ago
ViT image classification example broken with error: cannot import name 'is_fairscale_available' from 'transformers.integrations'
Updated 3 months ago1
More than 1 Trainium Instance
Updated 3 months ago3
Support layoutlm models inference
Updated 3 months ago
Support owlv2 models inference
Updated 3 months ago
Support Marian models inference
Updated 3 months ago
Add support for phi-2
Closed 3 months ago2
Cache files ignored when running a multi-host distributed training
Updated 3 months ago
Hang while testing single and multi-node llama training with TP=8, PP=1
Closed 4 months ago
Support compilation on any instance, not just inf/trn
Closed 4 months ago1
Exporting mistral model to neuron fails with ModuleNotFoundError: No module named 'optimum.commands.env'
Closed 4 months ago5
Optimum-neuron 0.0.19: Error when exporting mistral model to inf2
Closed 4 months ago2
ON compile LCM-SD broken with cannot import NeuronLatentConsistencyModelPipeline error
Closed 4 months ago3
sd 1.5 inpainting model inference
Closed 4 months ago16
Failure to compile a model on Inf1 with optimum-cli due to lack of arguments
Closed 4 months ago2
Neuron export -> import `num_cores`
Closed 4 months ago2