-
-
### Description
In Arxiv html format, we see "Anonymous Authors" in the webpage. Wonder how to fix this issue (previously we used MLSys format, which has the option of using "Anonymous Authors" in th…
-
### 🐛 Describe the bug
Fake tensor fake mode mismatch
repro
```python
def create_sd2():
from diffusers import StableDiffusionPipeline, EulerDiscreteScheduler
model_id = "stabilityai/st…
-
1. Improve the existing simple_inference application with a pipeline scheduling solution.
2. Define the tensor to be transmitted using `MPI_Datatype`.
3. The current version does not support GDR, so…
-
-
与@luomai 老师交流后，我很希望能够Co-Lead/参与到这一拓展章节的撰写过程中，我对Graph, GNN等比较熟悉，刚开始了解MLSys + Graph这一有趣的方向。
-
1. Abstract implementation of memory operations for heterogeneous devices.
2. High-level memory data struct. such as ring buffer, memory pool.
-
https://proceedings.mlsys.org/static/paper_files/mlsys/2020/94-Paper.pdf
来源：MLSys'20 Schedule
提出了一个新的超参数搜索算法 ASHA
-
Asynchronous Successive Halving Algorithm (ASHA) is, as far as I can understand, a tweaked version of the Hyperband multi-fidelity intenstifier (aka [ early stopping | pruning | scheduler ] algorithm)…
-
### 🐛 Describe the bug
Maybe we need to forbid it similar to how we forbid shared tensors with `PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True`
Logs
```
Traceback (most recent call last):
F…