-
I'm using the Stable Diffusion Configuration and modified it in order to run on an inferentia2 instance.
My inference script is as below
```
from io import BytesIO
from fastapi import FastAPI
f…
-
I am trying to finetune llama3-70B on trn132xlarge using distributed training. It failed with following error:
Container image: f"763104351884.dkr.ecr.{region}.amazonaws.com/pytorch-training-neur…
-
Model generates only garbage.
Sample: https://github.com/aws-neuron/aws-neuron-samples/blob/master/torch-neuronx/transformers-neuronx/inference/llama-3-8b-32k-sampling.ipynb
NeuronSDK2.19 PyTorc…
-
some logs:
#source shell_scripts/export_llama.sh qwen2 "" "" "" llama3.txt
checkpoint_files: ['models/llm_models/weights/Qwen2.5-0.5B-Instruct/model.safetensors']
Preparing Model Calibration In…
-
Hi, I am exporting a model using torch neuron but I can't find any reference to save a custom attribute in the model.
For instance I would like to save the dimension of input image as an integer so…
-
#### Summary
Hi again! I encountered a bug while playing with attention and sharding in JAX. The issue occurs with specific sharding setups and fails under certain core configurations.
#### Steps …
-
**General**
As presented, I don’t feel like it’s made clear what this tutorial adds to the tutorials. Setting aside the “Getting Started” Wilson-Cowan model, I think this is the first real introduct…
-
This issue occurs when building a layer with the maximum number of possible unqiue connections. For example, a layer with three neurons should have three possible connections: (0, 1), (1, 2) and (0, 2…
-
### Your current environment
The output of `python collect_env.py`
```text
Model name: Intel(R) Xeon(R) Platinum 8375C CPU @ 2.90GHz
CPU family: 6
Mo…
-
The description of the neuron model gif_psc_exp states that the adaptive current q_stc is given as a list of nA. I think this has to be pA.
I can not find a conversion between nA and pA in models/gif…