-
In #3396, I'm running into some issues migrating the [refbackend pipeline](https://github.com/llvm/torch-mlir/blob/a5d3b546f86f9547049b5ee4562bb4698de50b38/projects/pt1/python/torch_mlir_e2e_test/lina…
-
Hi, I am facing an annoying bug when cleaning up the memory after the run (core-dump). The bug occurs only on PPC64le architecture (IBM power9) when running on the GPU (Tesla V100) via the PortableCL …
-
## 🐛 Bug
I am training my ConvNet model with OCT Data and analysing the privacy spent using Opacus by implementing the [Random Sparsification](https://github.com/JunyiZhu-AI/RandomSparsification)…
-
**Describe the bug**
Performing https://github.com/vllm-project/llm-compressor/tree/main/examples/quantization_24_sparse_w4a16
```python
import os
import torch
from llmcompressor.transforme…
-
**Describe the bug**
RUN llm-compressor/examples/quantization_w8a8_fp8$ python llama3_example.py
save safetensors :KeyError: torch.float8_e4m3fn
**Expected behavior**
A clear and concise descri…
-
# Paper Information
- **Paper Title**: MeanSparse: Post-Training Robustness Enhancement Through Mean-Centered Feature Sparsification
- **Paper URL**: https://arxiv.org/pdf/2406.05927
- **Paper au…
-
### 🚀 The feature, motivation and pitch
I am trying to **train** ultra-sparse linear layers with as low as 0.1% of nonzero elements. Forward propagation is successful, however propagating the loss …
-
Hi, Ziming!
While trying your very detailed tutorial, I’ve found a severe issue which may undermine the effectiveness of integrating KAN into other regular neural networks.
For instance, in tutori…
-
At this moment, we only support `Dense` layers. This is becuase `SparseArrays.jl` currently supports only 1D vectors or 2D matrices, so we could try transforming `Conv` layers into dense matrix multip…
-
## Keyword: sgd
There is no result
## Keyword: optimization
### Multi-Target Decision Making under Conditions of Severe Uncertainty
- **Authors:** Authors: Christoph Jansen, Georg Schollmeyer, Thoma…