Skip to content

fix: Revert rayon from weight conversion (63% slower) #63

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
shift merged 1 commit into from
Apr 22, 2026
Merged

fix: Revert rayon from weight conversion (63% slower) #63

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 1 addition & 2 deletions src/model/cpu_block_attn_res.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,5 @@
use crate::model::cpu_linear::{CpuLinear, CpuRmsNorm};
use crate::model::cpu_moe::CpuMoELayer;
use rayon::prelude::*;
use crate::model::gemma_mapper::{matmul, rms_norm, apply_rope, apply_rope_gqa, gelu_tanh};
use crate::model::gemma_mapper::{MappedGemma4Model, Gemma4FfnWeights};

Expand Down Expand Up @@ -2148,7 +2147,7 @@ pub fn gemma4_to_block_attnres(teacher: &MappedGemma4Model) -> CpuBlockAttnResMo
let vs = config.vocab_size;
let first_shared_layer = config.num_layers.saturating_sub(config.num_kv_shared_layers);

let layers: Vec<CpuBlockAttnResLayer> = teacher.layers.par_iter().enumerate().map(|(layer_idx, layer_weights)| {
let layers: Vec<CpuBlockAttnResLayer> = teacher.layers.iter().enumerate().map(|(layer_idx, layer_weights)| {
let layer_head_dim = layer_weights.attn.head_dim;
let layer_q_dim = layer_weights.attn.q_dim;
let layer_kv_dim = layer_weights.attn.kv_dim;
Expand Down
Loading