jburtoft commited on
Commit
a11adad
1 Parent(s): 58872e7

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +34 -0
README.md ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ To run:
2
+
3
+ Stand up Inferentia inf2.8xlarge with Hugging Face AMI.
4
+
5
+ Update to 2.16:
6
+ ```
7
+ sudo apt-get update -y \
8
+ && sudo apt-get install -y --no-install-recommends \
9
+ aws-neuronx-dkms=2.15.9.0 \
10
+ aws-neuronx-collectives=2.19.7.0-530fb3064 \
11
+ aws-neuronx-runtime-lib=2.19.5.0-97e2d271b \
12
+ aws-neuronx-tools=2.16.1.0
13
+
14
+ pip3 install --upgrade \
15
+ neuronx-cc==2.12.54.0 \
16
+ torch-neuronx==1.13.1.1.13.0 \
17
+ transformers-neuronx==0.9.474 \
18
+ --extra-index-url=https://pip.repos.neuron.amazonaws.com
19
+
20
+ python3 -m pip install git+https://github.com/aws-neuron/transformers-neuronx.git
21
+
22
+ pip install git+https://github.com/huggingface/optimum.git
23
+ pip install git+https://github.com/huggingface/optimum-neuron.git
24
+ ```
25
+
26
+ Then you should be able to run it right from the repository:
27
+ ```
28
+ from optimum.neuron import NeuronStableDiffusionXLPipeline
29
+ pipeline = NeuronStableDiffusionXLPipeline.from_pretrained("jburtoft/TDXL", device_ids=[0, 1])
30
+ prompt = "a photo of an astronaut riding a horse on mars"
31
+ # generate image
32
+ image = pipeline(prompt).images[0]
33
+ image.save("generated_image.jpg")
34
+ ```