huggingface/candle
1
0
Fork 0
mirror of https://github.com/huggingface/candle.git synced 2025-06-20 04:00:28 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add support for Llama 3.1 (#2359)

Browse Source 
* Add Llama 3.1 rope

* Clippy

* Format

* Clippy

* Add support for multiple eos tokens:

* Untagged either

* Remove either dep and fix settings.json

* Make the max positional embeddings configurable
This commit is contained in:
Eric Buehler
2024-07-26 15:32:26 -04:00
committed by GitHub
parent ddafc61055
commit 0f5cbb08b3
24 changed files with 165 additions and 71 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
candle-core/benches/benchmarks/qmatmul.rs
View File

@ -7,7 +7,7 @@ use criterion::{black_box, criterion_group, Criterion, Throughput};
use std::time::Instant;
fn run(matmul: &QMatMul, x: &Tensor) {
matmul.forward(&x).unwrap();
matmul.forward(x).unwrap();
}
fn run_bench(c: &mut Criterion, device: &Device, dtype: GgmlDType) {
@ -50,7 +50,7 @@ fn run_bench(c: &mut Criterion, device: &Device, dtype: GgmlDType) {
fn criterion_benchmark(c: &mut Criterion) {
let handler = BenchDeviceHandler::new().unwrap();
for device in handler.devices {
for dtype in vec![
for dtype in [
GgmlDType::F32,
GgmlDType::F16,
GgmlDType::Q4_0,