Comments (2)
ezyang
commented on May 24, 2024
This one is size oblivious flavor: while technically the split tensor will be contiguous if we have a split with size 100, we just want to treat all of them as non-contiguous.
This is reminiscent of the problem at #120288
from pytorch.
ezyang
commented on May 24, 2024
It looks like asserting that you are not size 100 is enough
import torch
import torch.functional
torch._dynamo.config.capture_scalar_outputs = True
@torch.compile(fullgraph=True, dynamic=True)
def f(x, yt):
ys = yt.tolist()
for y in ys:
torch._check(x.size(1) != y)
return torch.functional.split(x.clone(), ys, dim=1)
f(torch.randn(80, 100), torch.tensor([20, 40, 40]))
from pytorch.
Related Issues (20)
- [dynamo] nn module parameterization can cause excessive restarts HOT 2
- got Unexpected error from cudaGetDeviceCount() after calling torch.cuda.is_available() HOT 2
- [Compiled autograd] Support user-defined triton kernel for which speculate_subgraph on bw failed
- torch_randn_like function potentially giving different results with torch.manual_seed()? HOT 1
- Error calling to_sparse_coo() on any subclass of Tensor
- C++ libtorch x86_64 Darwin deprecation note missing
- [NestedTensor] chunk fails under DEBUG=1 builds HOT 1
- DISABLED test_ring_attention_compile_attention_fn0 (__main__.RingAttentionTest) HOT 1
- DISABLED test_full_symbolic_value_cuda (__main__.TestInductorDynamicCUDA) HOT 2
- Eager/FX backend sometimes uses more memory than eager HOT 1
- scheduler.step() HOT 1
- Inductor generated Triton kernel spends double time from Llama2 to Llama 3 HOT 1
- `Dockerfile` should set the `syntax` directive to v1 HOT 2
- [DCP] DCP load for non-tensor values HOT 3
- Support loading with map_location on xpu backend HOT 4
- `LRScheduler` Refactoring HOT 6
- Why I can't differentiate the output of ODE solver(Scipy) by autograd.grad HOT 2
- ❓Different results between normal batching and `vmap` while using lower precision (e.g., bfloat16)
- aten::linalg_svd on mps backend
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from pytorch.