huggingface/candle
1
0
Fork 0
mirror of https://github.com/huggingface/candle.git synced 2025-06-16 10:38:54 +00:00
Code Issues Packages Projects Releases Wiki Activity

Simplify Tensor::randn. (#255)

Browse Source 
* Simplify Tensor::randn.

* Also switch Tensor::rand to use a generic dtype.

* Support sampling for f16.

* Cleanup.
This commit is contained in:
Laurent Mazare
2023-07-27 07:40:36 +01:00
committed by GitHub
parent 89ba005962
commit 6475bfadfe
10 changed files with 111 additions and 72 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
README.md
View File

@ -2,10 +2,11 @@
ML framework for Rust
```rust
let a = Tensor::zeros((2, 3), DType::F32, &Device::Cpu)?;
let b = Tensor::zeros((3, 4), DType::F32, &Device::Cpu)?;
let a = Tensor::randn(0f32, 1., (2, 3), &Device::Cpu)?;
let b = Tensor::randn(0f32, 1., (3, 4), &Device::Cpu)?;
let c = a.matmul(&b)?;
println!("{c}");
```
## Check out our examples
@ -45,13 +46,15 @@ And then browse to
## Features
- Simple syntax (looks and like PyTorch)
- CPU and Cuda backends, m1, f16, bf16 (and tentatively wasm)
- Simple syntax, looks and like PyTorch.
- CPU and Cuda backends, m1, f16, bf16.
- Enable serverless (CPU), small and fast deployments
- Model training
- Distributed computing (NCCL).
- Models out of the box (Llama, Whisper, Falcon, ...)
- Emphasis on enabling users to use custom ops/kernels
- WASM support, run your models in a browser.
- Model training.
- Distributed computing using NCCL.
- Models out of the box: Llama, Whisper, Falcon, BERT...
- Embed user-defined ops/kernels, such as [flash-attention
v2](https://github.com/LaurentMazare/candle/blob/89ba005962495f2bfbda286e185e9c3c7f5300a3/candle-flash-attn/src/lib.rs#L152).
## How to use ?
@ -59,9 +62,7 @@ Cheatsheet:
| | Using PyTorch | Using Candle |
|------------|------------------------------------------|------------------------------------------------------------------|
| Creation | `torch.Tensor([[1, 2], [3, 4]])` | `Tensor::new(` |
| | | ` &[[1f32, 2.]], [3., 4.]],` |
| | | ` &Device::Cpu)?` |
| Creation | `torch.Tensor([[1, 2], [3, 4]])` | `Tensor::new(&[[1f32, 2.]], [3., 4.]], &Device::Cpu)?` |
| Indexing | `tensor[:, :4]` | `tensor.i((.., ..4))?` |
| Operations | `tensor.view((2, 2))` | `tensor.reshape((2, 2))?` |
| Operations | `a.matmul(b)` | `a.matmul(&b)?` |