# Accelerate sees only one GPU on multi-GPU Sagemaker instance

**URL:** <https://discuss.huggingface.co/t/accelerate-sees-only-one-gpu-on-multi-gpu-sagemaker-instance/37137>\
**Category:** 🤗Accelerate\
**Created:** [April 19, 2023, 4:48pm UTC](https://discuss.huggingface.co/t/accelerate-sees-only-one-gpu-on-multi-gpu-sagemaker-instance/37137 "2023-04-19T16:48:47Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![alexcarterkarsus](https://avatars.discourse-cdn.com/v4/letter/a/9dc877/32.png) [@alexcarterkarsus](https://discuss.huggingface.co/u/alexcarterkarsus)\
**Post date:** [April 19, 2023, 4:48pm UTC](https://discuss.huggingface.co/t/accelerate-sees-only-one-gpu-on-multi-gpu-sagemaker-instance/37137/1 "2023-04-19T16:48:47Z")

</div>

Hello,  
I am trying to use Accelerate on Sagemaker training instance (p3.8xlarge) with 4 GPUs, in the hope that Accelerate can automatically leverage the multiple GPU, but it seems it can’t detect them.  
so, while setting up HuggingFace estimator for the Sagemaker instance, I can set the distribution parameter as

```auto
distribution = {
    
        'pytorchddp': {'enabled': True }
}

```

,  
and use it in HF estimator:

```auto
pytorch_estimator = HuggingFace(entry_point='my_script.py',
                            role=role,
                            instance_count=1,
                            instance_type='ml.p3.8xlarge',
                            py_version='py38',
                            pytorch_version="1.10.2",
                            transformers_version="4.17.0",
                            volume_size=60,
                            source_dir= '/my_path/',
                            tags = aws_tags,
                            max_run=3*24 * 60 * 60,
                           distribution = distribution,
                            output_path = "s3://my_s3/output",
                            checkpoint_s3_uri=checkpoint_s3_bucket,
                            checkpoint_local_path=checkpoint_local_path
                           )

```

inside “my\_script.py” I can verify if PyTorch can see all the GPUs:

```auto
import torch
available_gpus = [torch.cuda.device(i) for i in range(torch.cuda.device_count())]
print(available_gpus)

```

and indeed I can see there are 4 GPUs from the output:  
[\<torch.cuda.device object at 0x7f479f0b8820\>, \<torch.cuda.device object at 0x7f479f0b8760\>, \<torch.cuda.device object at 0x7f479f0b87c0\>, \<torch.cuda.device object at 0x7f479f0b8730\>]

however, I then set up Accelerate, and let it print out the available processes it sees,

```auto
accelerator = Accelerator(mixed_precision=args.mixed_precision)
...
print("Accelerator has determined the num processes to be: ", accelerator.num_processes)

```

and it sees only 1:  
“Accelerator has determined the num processes to be: 1”

This is further corroborate by the fact that running the same training code on a p3.2xlarge instance takes as much time as on this p3.8xlarge instance, i.e., only one GPU is working (presumably because Accelerate only sees one GPU).  
So my question is: is there some special way to make Accelerate aware of all the GPUs in a Sagemaker instance? There were some similar discussion on Github([Cannot run distributed training on sagemaker · Issue #492 · huggingface/accelerate · GitHub](https://github.com/huggingface/accelerate/issues/492)) and I thought the issue was resolved…  
Thanks!

---

<div class="post-metadata">

**Author:** ![smangrul](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/smangrul/32/3791_2.png) [@smangrul](https://discuss.huggingface.co/u/smangrul)\
**Post date:** [May 2, 2023, 7:39am UTC](https://discuss.huggingface.co/t/accelerate-sees-only-one-gpu-on-multi-gpu-sagemaker-instance/37137/2 "2023-05-02T07:39:58Z")

</div>

Hello @alexcarterkarsus,

Are you following this [notebooks/sagemaker/22\_accelerate\_sagemaker\_examples at main · huggingface/notebooks · GitHub](https://github.com/huggingface/notebooks/tree/main/sagemaker/22_accelerate_sagemaker_examples)

You might be launching or using the integration in wrong way. Please refer the above repo for concrete examples and try running them and then adapting it for your usecases.