Skip to content

Commit d6d53b3

Browse files
author
MichaelHirn
committed
docs/readme: add cuDNN v4 to feature matrix
CLOSE: #43
1 parent 080ffe6 commit d6d53b3

File tree

2 files changed

+37
-36
lines changed

2 files changed

+37
-36
lines changed

‎README.md

+18-18
Original file line numberDiff line numberDiff line change
@@ -24,24 +24,24 @@ This Plugins provides the following operations to the Collenchyma Backend.
2424
Every Operation includes forward + backward. A `-` means not yet implemented.
2525
More information can be found in the [Documentation][docs-ops].
2626

27-
| Operation | CUDA | OpenCL | Native |
28-
|--- |--- |--- |--- |
29-
| Sigmoid | cuDNN v3 | - | Rust |
30-
| SigmoidPointwise | cuDNN v3 | - | |
31-
| ReLU | cuDNN v3 | - | Rust |
32-
| ReLUPointwise | cuDNN v3 | - | |
33-
| Tanh | cudNN v3 | - | Rust |
34-
| TanhPointwise | cudNN v3 | - | |
35-
| | | | |
36-
| Normalization (LRN) | cudNN v3 | - | - |
37-
| | | | |
38-
| Convolution | cudNN v3 | - | - |
39-
| | | | |
40-
| Softmax | cudNN v3 | - | Rust |
41-
| LogSoftmax | cudNN v3 | - | Rust |
42-
| | | | |
43-
| Pooling Max | cudNN v3 | - | - |
44-
| Pooling Avg | cudNN v3 | - | - |
27+
| Operation | CUDA | OpenCL | Native |
28+
|--- |--- |--- |--- |
29+
| Sigmoid | { cuDNN v3, v4 } | - | Rust |
30+
| SigmoidPointwise | { cuDNN v3, v4 } | - | |
31+
| ReLU | { cuDNN v3, v4 } | - | Rust |
32+
| ReLUPointwise | { cuDNN v3, v4 } | - | |
33+
| Tanh | { cuDNN v3, v4 } | - | Rust |
34+
| TanhPointwise | { cuDNN v3, v4 } | - | |
35+
| | | | |
36+
| Normalization (LRN) | { cuDNN v3, v4 } | - | - |
37+
| | | | |
38+
| Convolution | { cuDNN v3, v4 } | - | - |
39+
| | | | |
40+
| Softmax | { cuDNN v3, v4 } | - | Rust |
41+
| LogSoftmax | { cuDNN v3, v4 } | - | Rust |
42+
| | | | |
43+
| Pooling Max | { cuDNN v3, v4 } | - | - |
44+
| Pooling Avg | { cuDNN v3, v4 } | - | - |
4545

4646
Kudos to [ehiggs][ehiggs], for implementing the native Rust operations.
4747

‎src/lib.rs

+19-18
Original file line numberDiff line numberDiff line change
@@ -80,24 +80,25 @@
8080
//! This Plugins provides the following operations. (Forward + Backward)
8181
//! A `-` means not yet implemented.
8282
//!
83-
//! | Operation | CUDA | OpenCL | Native |
84-
//! |--- |--- |--- |--- |
85-
//! | Sigmoid | cuDNN v3 | - | Rust |
86-
//! | SigmoidPointwise | cuDNN v3 | - | |
87-
//! | ReLU | cuDNN v3 | - | Rust |
88-
//! | ReLUPointwise | cuDNN v3 | - | |
89-
//! | Tanh | cudNN v3 | - | Rust |
90-
//! | TanhPointwise | cuDNN v3 | - | |
91-
//! | | | | |
92-
//! | Normalization (LRN) | cudNN v3 | - | - |
93-
//! | | | | |
94-
//! | Convolution | cudNN v3 | - | - |
95-
//! | | | | |
96-
//! | Softmax | cudNN v3 | - | Rust |
97-
//! | LogSoftmax | cudNN v3 | - | Rust |
98-
//! | | | | |
99-
//! | Pooling Max | cudNN v3 | - | - |
100-
//! | Pooling Avg | cudNN v3 | - | - |
83+
84+
//! | Operation | CUDA | OpenCL | Native |
85+
//! |--- |--- |--- |--- |
86+
//! | Sigmoid | { cuDNN v3, v4 } | - | Rust |
87+
//! | SigmoidPointwise | { cuDNN v3, v4 } | - | |
88+
//! | ReLU | { cuDNN v3, v4 } | - | Rust |
89+
//! | ReLUPointwise | { cuDNN v3, v4 } | - | |
90+
//! | Tanh | { cuDNN v3, v4 } | - | Rust |
91+
//! | TanhPointwise | { cuDNN v3, v4 } | - | |
92+
//! | | | | |
93+
//! | Normalization (LRN) | { cuDNN v3, v4 } | - | - |
94+
//! | | | | |
95+
//! | Convolution | { cuDNN v3, v4 } | - | - |
96+
//! | | | | |
97+
//! | Softmax | { cuDNN v3, v4 } | - | Rust |
98+
//! | LogSoftmax | { cuDNN v3, v4 } | - | Rust |
99+
//! | | | | |
100+
//! | Pooling Max | { cuDNN v3, v4 } | - | - |
101+
//! | Pooling Avg | { cuDNN v3, v4 } | - | - |
101102
//!
102103
//! [collenchyma]: https://github.com/autumnai/collenchyma
103104
//! [collenchyma-docs]: http://autumnai.github.io/collenchyma

0 commit comments

Comments
 (0)