-
## Fix the model test for `hf_Reformer.py`
1. setup env according to [Run a model under torch_xla2](https://github.com/pytorch/xla/blob/master/experimental/torch_xla2/docs/support_a_new_model.md)
…
-
### Issue description
Hydrogen reformer stage 2 does not work.......................Sorry ist ok its my mistake ,
I have not seen high pressure and extreme pressure !!!
### Reproduction steps
…
-
Found while running torchbench's `hf_Reformer` with thunderFX path (see below for the larger graph that fails).
```python
import torch
import thunder
def forward(x, double_scalar_tensor):
…
-
### 🐛 Describe the bug
Category | Model | Accuracy
-- | -- | --
torchbench_amp_bf16_training | basic_gnn_edgecnn | eager_two_runs_differ
torchbench_amp_bf16_training | …
-
## 一言でいうと
Transformerの演算を効率化する手法。Attentionの計算におけるQKではQに近いKだけ考慮される(=遠い所の計算は無駄)なので、ハッシュ関数を使ってソートしたうえでchunkに区切り、chunk内(似た者同士)/前chunk(接続)のみAttention計算を行う。またReversibleのResNetによる効率化も行っている。
 – 기근화
A man who want to be Batman
[https://ghk829.github.io/whitecross.github.com//reformer/](https://ghk829.github.io/whitecross.github.com//reformer/)
-
# Description:
Hello! I appreciate the excellent work on benchmarking Performer and Longformer against the base Transformer. I’d like to propose the implementation of additional efficient Transformer…