Comments (5)
apaszke
commented on April 29, 2024
I agree that full isn't the best name (see full/valid/same convolution e.g. in here), but why transpose? Is there an intuitive explanation?
from pytorch.
jcjohnson
commented on April 29, 2024
There is a good argument for convolution transpose in the appendix to this paper:
https://arxiv.org/abs/1602.05110
from pytorch.
rbgirshick
commented on April 29, 2024
The MatConvNet manual also has a good explanation: http://www.vlfeat.org/matconvnet/matconvnet-manual.pdf. IIRC, if you think about doing convolution by toeplitz matrix multiplication, then the so-called "deconvolution" operator is doing matrix multiplication with the transpose of a toeplitz matrix.
from pytorch.
apaszke
commented on April 29, 2024
Thanks for the pointers! I've actually found the explanation in theano docs a couple days ago: http://deeplearning.net/software/theano_versions/dev/tutorial/conv_arithmetic.html#transposed-convolution.
from pytorch.
soumith
commented on April 29, 2024
fixed via #113
from pytorch.
Related Issues (20)
- DISABLED test_batched_mm_float32_bs_1_cuda_float32 (__main__.TestDecompCUDA) HOT 2
- Request for modifying native allocator to use nvidia unified memory `cudaMallocManage` HOT 3
- [Dynamo] Fix refleak in 3.12+ and Dynamic Shapes test_parameter_free
- Many tests under test/distributed/elastic not running in OSS CI
- DISABLED test_grad_to_node_inplace (__main__.TestAutogradWithCompiledAutograd) HOT 6
- DISABLED test_contiguous (__main__.TestPartitioning) HOT 3
- device_mesh / fsdp issue with _get_device_handle HOT 2
- PreDispatch tracing doesn't produce correct "val" for Sparse Fake tensors. HOT 1
- `test_triton_scaled_dot_product_attention_block_size_16_cuda_bfloat16` is broken on A100 HOT 2
- Data corrupted when transferring over 4GiB to MPS HOT 3
- Dynamo Export: Support for PixelShuffle HOT 2
- mypy.ini isn't valid HOT 1
- Inconsistent argument parsing behavior for torch.ops bindings between Tensor and FakeTensor
- DISABLED test_grad_unreachable (__main__.TestAutogradWithCompiledAutograd) HOT 3
- [RFC] efficient nan-checker for distributed-collectives HOT 5
- BUG Report using `backend:cudaMallocAsync`
- DISABLED test_binary_op_float_inf_nan__foreach_add_cuda_float32 (__main__.TestForeachCUDA) HOT 1
- DISABLED test_default_partitioner_output_tensor_shape_tensor (__main__.TestPartitioning) HOT 3
- torch.compile produces wrong result for changing attributes in python class instances HOT 1
- libtorch_cuda_linalg.so missing
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.