Jay Gu

Another input I found that causes the stuck in while-loop is here: String test = "7h... nh? a nhìu th?t nhìu ^^~..........................................................................." Looks like the "....." pattern is the problem?

This is not the only bad example. Also I don't think the problem is the while loop, it might be due to some catastrophic regex matching which can take years.

Cool. I'm going to test it on the daily.10k tonight. On Tue, Oct 23, 2012 at 1:16 AM, brendano [email protected] wrote: > I just ran it on 1489999 or so...

It passed the test of daily.10k with 14720000 tweets. Will try on larger dataset later tomorrow. Jay On Tue, Oct 23, 2012 at 1:25 AM, Haijie Gu [email protected] wrote: >...

``` import torch from torch.nn import Parameter, Module aten = torch.ops.aten linear = aten.linear.default reshape = aten.reshape.default permute = aten.permute.default sdp = aten.scaled_dot_product_attention.default def mha(x, Wq, Wk, Wv, Wo, heads):...

This issue is caused by `builder.reset()`. I have a better repro: ```python import tensorrt as trt def test(use_reset = False): builder = trt.Builder(trt.Logger()) config = builder.create_builder_config() if use_reset: builder.reset() #...

@lix19937 Here's a repro shows that even setting np.float32 for `dummy_w`, `dummy_b`, `scale`, `bias`, it still error with ` Error Code 1: Cuda Runtime (misaligned address)`. ```python import tensorrt as...

> Can you use torch api to export an onnx of `affine_group_norm`, then use trtexec convert ? @haijieg @lix19937 No, I want to directly control how network is built using...

@EricLina thank you for your interest. We don't have plan to publish any benchmarks, but we have some [samples](https://github.com/NVIDIA/cutile-python/blob/7a86253d22446a24d68083584f01267325157891/samples/AttentionFMHA.py#L410) for you to get started on running your own comparison. The...

Hi @vgoklani, you can find a backward kernel example [here](https://github.com/NVIDIA/cutile-python/blob/main/samples/LayerNorm.py#L167).