-
{noformat}Caused by: java.lang.AssertionError
at water.fvec.Vec$ESPC.rowLayout(Vec.java:1726)
at water.fvec.AppendableVec.compute_rowLayout(AppendableVec.java:95)
at water.fvec.AppendableVec.layout_an…
-
How to understand waitPeer and postPeer function in prims_simple.h?
-
## 🚀 Memory clearing mechanism with torch.autograd.Function integration
Today we pass "saved for backward" tensors to the generated backward function inside a Python list and within the generated f…
-
## 🐛 Bug
Repro script from @nikitaved
```
import itertools
import torch
from torch.utils.benchmark import Timer, Compare
import thunder
#dtypes = [torch.bfloat16, torch.float, torch.double…
-
### 🐛 Describe the bug
Not sure whether this kind of thing is not supposed to be supported, but trying to get to the base tensor of a view using `as_strided` operation fails when trying to `torch.c…
-
From the example below, the autocast is applied only for device cuda, however thunder.jit still applies it to CPU inputs.
```python
import thunder
import torch
def foo(x, w):
return torch…
-
## 🐛 Bug
```py
import thunder
import torch
@thunder.jit
def normal(shape, device):
return torch.randn(tuple(shape), device=device)
shape = [1, 4, 40, 84, 84]
t1 = normal(shape, devic…
-
The class labels primitives are being moved over to RAFT quickly so that they can be used to support SLHC in 0.19. Just to avoid any surprises, we're keeping the function signatures compatible with th…
-
The given solution is breaking for the following test case on gfg
10 11
0 1 7
0 4 4
0 7 8
0 9 3
1 7 7
1 8 3
2 5 8
2 7 10
3 4 3
4 6 10
5 7 8
-
E:\work\ncnn\build\tools\onnx\pnnx>pnnx bestyb2.onnx inputshape=[1,3,640,640] fp16=1
pnnxparam = bestyb2.pnnx.param
pnnxbin = bestyb2.pnnx.bin
pnnxpy = bestyb2_pnnx.py
pnnxonnx = bestyb2.pnnx.onnx…