-
https://tone.aliyun-inc.com/ws/xesljfzh/test_result/344452?tab=1
【环境准备】
```
BINARY_URL=oss://compiler-ci-bucket/dragonwell8/CI/tar/20240712-144239-511-#464-linux.x86_64.fastdebug.master-f827bf4b4…
-
### 🐛 Describe the bug
**Description:**
I have encountered some precision differences when using the `__add__`, `layer_norm` and `linear` in the following way.
**Code to Reproduce:**
```pyth…
-
### 🐛 Describe the bug
**Description:**
I have encountered some precision differences when using the `__add__`, `layer_norm` and `__add__` in the following way.
**Code to Reproduce:**
```pyt…
-
### 🐛 Describe the bug
**Description:**
I have encountered some precision differences when using the `conv2d` and `batch_norm` in the following way.
**Code to Reproduce:**
```python
import …
-
### Your current environment
```text
PyTorch version: 2.0.1+cu117
Is debug build: False
CUDA used to build PyTorch: 11.7
ROCM used to build PyTorch: N/A
OS: Ubuntu 22.04.3 LTS (x86_64)
GCC ve…
-
### 🐛 Describe the bug
**Description:**
I have encountered some precision differences when using the `gelu`, `dropout` and `linear` in the following way.
**Code to Reproduce:**
```python
im…
-
### 🐛 Describe the bug
**Description:**
I have encountered some precision differences when using the `batch_norm` and `conv2d` in the following way.
**Code to Reproduce:**
```python
import …
-
Hi guys,
I cannot serve an vllm server with a config.yaml file using
`vllm serve model --config config.yaml`
Regardsless of the specified path, I always get:
`vllm: error: unrecognized ar…
-
This bug report may mainly talk about, how to perform benchmark correctly. I use both code from cargo book and from internet searching, the result is quite different among them.
I tried to impl…
-
Since the new `1.14.1` release, symbols mode no longer works for me.
Other modes work perfectly fine, so maybe it has something to do with the new **AVX2** optims for symbols?
I'm on Arch Linux.…
folke updated
2 months ago