Docs.rs
rullama-0.5.0
Platform
x86_64-unknown-linux-gnu
Feature flags
docs.rs
About docs.rs
Badges
Builds
Metadata
Shorthand URLs
Download
Rustdoc JSON
Build queue
Privacy policy
Rust
Rust website
The Book
Standard Library API Reference
Rust by Example
The Cargo Guide
Clippy Documentation
rullama 0.5.0
Browser-resident Gemma 4 inference: pure Rust → WebAssembly + WebGPU. Loads Ollama's on-disk GGUF blobs and runs the forward pass on the local GPU via hand-written WGSL.
Crate
Source
Builds
Feature flags
Documentation
..
adain.wgsl
adam_step.wgsl
add_bias_batched.wgsl
attention.wgsl
attention_backward_dkv.wgsl
attention_backward_dq.wgsl
attention_probs.wgsl
avg_pool2d.wgsl
avg_pool2d_half_chf.wgsl
bf16_matmul.wgsl
bf16_matmul_batched.wgsl
bf16_matmul_batched_tiled.wgsl
bf16_matmul_batched_tiled_v2.wgsl
bf16_matmul_batched_tiled_v3.wgsl
bf16_matmul_batched_tiled_v3_f16lds.wgsl
block_local_attention.wgsl
clamp.wgsl
conv1d.wgsl
conv1d_f16.wgsl
conv2d.wgsl
conv2d_chf.wgsl
conv2d_chf_f16.wgsl
conv_transpose1d.wgsl
conv_transpose1d_f16.wgsl
cross_entropy_backward.wgsl
depthwise_conv1d.wgsl
diffusion_attention.wgsl
f16_matmul.wgsl
f16_matmul_batched.wgsl
f16_matmul_batched_tiled.wgsl
f16_matmul_batched_tiled_v2.wgsl
f16_matmul_batched_tiled_v3.wgsl
f16_matmul_batched_tiled_v3_f16lds.wgsl
f16_matmul_batched_tiled_v4.wgsl
geglu.wgsl
geglu_backward.wgsl
gelu_exact.wgsl
glu_split.wgsl
half_residual_add.wgsl
istft.wgsl
layernorm_affine.wgsl
leaky_relu.wgsl
lora_embed_col_read.wgsl
lora_embed_col_scatter_add.wgsl
lora_matmul_col.wgsl
lora_matmul_fused.wgsl
lora_matmul_fused_f16b.wgsl
lora_matmul_row.wgsl
lora_outer_add.wgsl
matmul_q4_0_backward_input.wgsl
matmul_q4_k_backward_input.wgsl
matmul_q6_k_backward_input.wgsl
moe_combine.wgsl
moe_combine_batched.wgsl
moe_expert_matmul_batched_q4_k.wgsl
moe_expert_matmul_batched_q5_0.wgsl
moe_expert_matmul_batched_q8_0.wgsl
moe_expert_matmul_q4_k.wgsl
moe_expert_matmul_q5_0.wgsl
moe_expert_matmul_q8_0.wgsl
moe_geglu_halves.wgsl
moe_geglu_halves_batched.wgsl
moe_router.wgsl
moe_router_batched.wgsl
nearest_upsample2x.wgsl
pos_embed_add.wgsl
q4_0_dequant_matmul.wgsl
q4_k_dequant_matmul.wgsl
q4_k_dequant_matmul_f16lds.wgsl
q4_k_dequant_matmul_tiled.wgsl
q4_k_dequant_matmul_wg256.wgsl
q5_0_dequant_matmul.wgsl
q6_k_dequant_matmul.wgsl
q6_k_dequant_matmul_f16lds.wgsl
q6_k_dequant_matmul_tiled.wgsl
q8_0_dequant_matmul.wgsl
quick_geglu.wgsl
residual_add.wgsl
rmsnorm.wgsl
rmsnorm_backward.wgsl
rmsnorm_per_row.wgsl
rmsnorm_per_row_backward.wgsl
rope_2d.wgsl
rope_neox.wgsl
rope_neox_backward.wgsl
scale.wgsl
scale_per_inner_dim.wgsl
silu.wgsl
snake.wgsl
softcap.wgsl
spec_phase.wgsl
sum_of_squares.wgsl
transpose2d.wgsl
transpose_hpd_to_phd.wgsl
transpose_phd_to_hpd.wgsl
vision_attention.wgsl
vision_attention_flash.wgsl
vision_attention_flash_hpd_f16.wgsl
vision_attention_flash_hpd_f16_q4.wgsl
vision_attention_flash_hpd_f16_wg128.wgsl
vision_attention_flash_hpd_f16_wg32.wgsl
vision_attention_flash_q16.wgsl
vision_attention_flash_q4.wgsl
vision_attention_flash_q8.wgsl
vision_attention_flash_sub_hpd.wgsl
vision_attention_flash_sub_hpd_f16.wgsl
vision_attention_flash_sub_hpd_f16_q16.wgsl
vision_attention_flash_sub_t64.wgsl
vision_attention_flash_subgroup.wgsl