-
```cpp
import jittor as jt
jt.flags.use_cuda = 1
a = jt.float32([1,2,3]*10)
b = jt.float32([4,5,6]*10)
for i in range(1):
c = a*b
print(c)
print(c.data)
```
currently I use your …
-
Running full style avatar but it seems to resize images the wrong way. It zooms out a lot on the face for each image (even though it was 1536 crop size from faceverse and there is a lot of space aroun…
-
Loading config from: ./projects/ngp/configs/ngp_base.py
load train data
Traceback (most recent call last):
File "/home/featurize/JNeRF/python/jnerf/utils/registry.py", line 33, in build_from_cfg…
-
切图时报错,完整信息如下
```
[i 0908 02:20:53.712638 68 compiler.py:955] Jittor(1.3.5.12) src: /usr/local/miniconda3/envs/jit/lib/python3.8/site-packages/jittor
[i 0908 02:20:53.722398 68 compiler.py:956] g++ …
-
-
An error occurs when training the isaid dataset. The environment configuration and data set processing are performed according to your instructions:
![image](https://github.com/Jittor/JSeg/assets/828…
-
我首先使用系统自带的gcc 11.3.0 报错
![image](https://github.com/Jittor/JittorLLMs/assets/41530341/ed43383b-2e9b-448a-8177-ff60a477732c)
我尝试了 conda install gcc=8 gxx=8, 报错为GOMP的链接错误
![image](https://github.co…
-
## Describe the bug
在使用`nn.CrossEntropyLoss()`时,指定`weight`参数为`jitter.ones(num_classes, dtype='float32')`和置为`None`的运行结果不同。且指定包括全一在内的值后会带来梯度爆炸。
## Full Log
这是指定了`weight`具体值的训练曲线。
![accuracy](https:/…
-
unrecognized relocation (0x2a) in section `.text'
cat /etc/os-release
NAME="Tencent tlinux"
VERSION="2.2 (Final)"
ID="tlinux"
ID_LIKE="rhel fedora centos"
VERSION_ID="2.2"
PRETTY_NAME="Ten…
-