char-fork/llama.cpp

git clone https://git.t4t.associates/char-fork/llama.cpp

Charlotte Somenc/dec in server context8300f1411

char/aliceai-t5-moe
180 files

files

abs.cl3.2 KiB
add.cl5.0 KiB
add_id.cl1.4 KiB
argsort.cl2.7 KiB
clamp.cl649 B
concat.cl6.6 KiB
conv2d.cl6.9 KiB
conv2d_f16_f32.cl6.5 KiB
cpy.cl8.6 KiB
cumsum.cl3.8 KiB
cvt.cl89.5 KiB
diag.cl834 B
diag_mask_inf.cl1.6 KiB
div.cl3.7 KiB
embed_kernel.py477 B
exp.cl3.4 KiB
expm1.cl3.2 KiB
fill.cl465 B
flash_attn_f16.cl16.2 KiB
flash_attn_f32.cl16.1 KiB
flash_attn_f32_f16.cl103.9 KiB
flash_attn_f32_q4_0.cl79.7 KiB
flash_attn_f32_q8_0.cl72.0 KiB
flash_attn_pre_f16.cl5.6 KiB
gated_delta_net.cl9.0 KiB
gelu.cl2.5 KiB
gemm_moe_mxfp4_f32.cl6.2 KiB
gemm_moe_mxfp4_f32_ns.cl17.5 KiB
gemm_moe_mxfp4_q8_1_dp4a.cl7.8 KiB
gemm_moe_q4_0_f32_ns.cl15.5 KiB
gemm_moe_q4_0_q8_1_dp4a.cl6.8 KiB
gemm_moe_q4_1_f32_ns.cl15.5 KiB
gemm_moe_q4_k_f32_ns.cl15.8 KiB
gemm_moe_q4_k_q8_1_dp4a.cl8.4 KiB
gemm_moe_q5_0_f32_ns.cl16.4 KiB
gemm_moe_q5_1_f32_ns.cl16.5 KiB
gemm_moe_q5_k_f32_ns.cl16.9 KiB
gemm_moe_q6_k_f32_ns.cl16.6 KiB
gemm_moe_q6_k_q8_1_dp4a.cl8.1 KiB
gemm_moe_q8_0_f32_ns.cl9.9 KiB
gemm_moe_q8_1_dp4a.cl10.8 KiB
gemm_noshuffle_iq4_nl_f32.cl5.5 KiB
gemm_noshuffle_iq4_nl_q8_1_dp4a.cl5.7 KiB
gemm_noshuffle_q1_0_f32.cl3.0 KiB
gemm_noshuffle_q4_0_f32.cl6.0 KiB
gemm_noshuffle_q4_0_q8_1_dp4a.cl5.3 KiB
gemm_noshuffle_q4_1_f32.cl4.7 KiB
gemm_noshuffle_q4_k_f32.cl18.7 KiB
gemm_noshuffle_q4_k_q8_1_dp4a.cl11.7 KiB
gemm_noshuffle_q5_0_f32.cl5.5 KiB
gemm_noshuffle_q5_0_q8_1_dp4a.cl9.5 KiB
gemm_noshuffle_q5_1_f32.cl5.7 KiB
gemm_noshuffle_q5_k_f32.cl7.4 KiB
gemm_noshuffle_q5_k_q8_1_dp4a.cl6.5 KiB
gemm_noshuffle_q6_k_f32.cl10.3 KiB
gemm_noshuffle_q6_k_f32_tiled.cl5.8 KiB
gemm_noshuffle_q6_k_q8_1_dp4a.cl5.9 KiB
gemm_noshuffle_q8_0_f32.cl3.8 KiB
gemm_noshuffle_q8_0_q8_1_dp4a.cl8.1 KiB
gemm_xmem_f16_f32_os8.cl8.7 KiB
gemv_moe_mxfp4_f32.cl5.8 KiB
gemv_moe_mxfp4_f32_ns.cl8.9 KiB
gemv_moe_q4_0_f32_ns.cl3.8 KiB
gemv_moe_q4_1_f32_ns.cl3.9 KiB
gemv_moe_q4_k_f32_ns.cl8.6 KiB
gemv_moe_q5_0_f32_ns.cl4.2 KiB
gemv_moe_q5_1_f32_ns.cl4.4 KiB
gemv_moe_q5_k_f32_ns.cl5.6 KiB
gemv_moe_q6_k_f32_ns.cl5.4 KiB
gemv_noshuffle_iq4_nl_f32.cl18.1 KiB
gemv_noshuffle_q1_0_f32.cl6.5 KiB
gemv_noshuffle_q4_0_f32.cl21.1 KiB
gemv_noshuffle_q4_0_f32_32b_trans.cl6.1 KiB
gemv_noshuffle_q4_0_f32_spec.cl16.0 KiB
gemv_noshuffle_q4_1_f32.cl21.0 KiB
gemv_noshuffle_q4_k_f32.cl44.6 KiB
gemv_noshuffle_q4_k_f32_32b_trans.cl4.5 KiB
gemv_noshuffle_q4_k_f32_o4.cl20.0 KiB
gemv_noshuffle_q4_k_f32_tiled.cl4.6 KiB
gemv_noshuffle_q5_0_f32.cl21.1 KiB
gemv_noshuffle_q5_1_f32.cl21.9 KiB
gemv_noshuffle_q5_k_f32.cl28.3 KiB
gemv_noshuffle_q6_k_f32.cl30.0 KiB
gemv_noshuffle_q6_k_f32_o4.cl26.3 KiB
gemv_noshuffle_q6_k_f32_tiled.cl7.9 KiB
gemv_noshuffle_q8_0_f32.cl12.8 KiB
get_rows.cl5.4 KiB
glu.cl14.4 KiB
group_norm.cl3.4 KiB
im2col_f16.cl1.3 KiB
im2col_f32.cl1.3 KiB
l2_norm.cl2.1 KiB
mean.cl3.3 KiB
moe_add_id_glu.cl3.0 KiB
moe_combine.cl3.8 KiB
moe_reorder_b.cl717 B
moe_reorder_quant_a_q8_1.cl2.2 KiB
moe_sort_by_expert.cl4.3 KiB
mul.cl4.0 KiB
mul_mat_f16_f32.cl4.5 KiB
mul_mm_f16_f32_kq_kqv.cl11.0 KiB
mul_mm_f16_f32_l4_lm.cl4.7 KiB
mul_mm_f32_f32_l4_lm.cl6.9 KiB
mul_mm_iq4_nl_f32_l4_lm.cl5.9 KiB
mul_mm_q1_0_f32_l4_lm.cl5.4 KiB
mul_mm_q4_0_f32_l4_lm.cl5.6 KiB
mul_mm_q4_1_f32_l4_lm.cl5.6 KiB
mul_mm_q4_k_f32_l4_lm.cl6.3 KiB
mul_mm_q5_0_f32_l4_lm.cl5.9 KiB
mul_mm_q5_1_f32_l4_lm.cl6.0 KiB
mul_mm_q5_k_f32_l4_lm.cl6.8 KiB
mul_mm_q6_k_f32_l4_lm.cl5.3 KiB
mul_mm_q8_0_f32_l4_lm.cl4.9 KiB
mul_mv_f16_f16.cl4.0 KiB
mul_mv_f16_f32.cl4.1 KiB
mul_mv_f16_f32_1row.cl3.4 KiB
mul_mv_f16_f32_l4.cl37.0 KiB
mul_mv_f16_f32_mrow.cl16.1 KiB
mul_mv_f32_f32.cl3.5 KiB
mul_mv_id_mxfp4_f32.cl5.6 KiB
mul_mv_id_mxfp4_f32_flat.cl5.7 KiB
mul_mv_id_q4_0_f32_8x_flat.cl8.4 KiB
mul_mv_id_q8_0_f32.cl3.9 KiB
mul_mv_id_q8_0_f32_flat.cl6.5 KiB
mul_mv_iq4_nl_f32.cl4.7 KiB
mul_mv_iq4_nl_f32_flat.cl6.4 KiB
mul_mv_mxfp4_f32.cl4.5 KiB
mul_mv_mxfp4_f32_flat.cl5.3 KiB
mul_mv_q1_0_f32.cl4.2 KiB
mul_mv_q1_0_f32_flat.cl7.8 KiB
mul_mv_q4_0_f32.cl6.2 KiB
mul_mv_q4_0_f32_1d_16x_flat.cl10.1 KiB
mul_mv_q4_0_f32_1d_8x_flat.cl8.1 KiB
mul_mv_q4_0_f32_8x_flat.cl8.1 KiB
mul_mv_q4_0_f32_v.cl7.2 KiB
mul_mv_q4_1_f32.cl5.7 KiB
mul_mv_q4_1_f32_flat.cl6.4 KiB
mul_mv_q4_k_f32.cl5.7 KiB
mul_mv_q4_k_f32_flat.cl7.2 KiB
mul_mv_q5_0_f32.cl6.7 KiB
mul_mv_q5_0_f32_flat.cl7.0 KiB
mul_mv_q5_1_f32.cl6.7 KiB
mul_mv_q5_1_f32_flat.cl7.2 KiB
mul_mv_q5_k_f32.cl6.1 KiB
mul_mv_q5_k_f32_flat.cl9.0 KiB
mul_mv_q6_k_f32.cl7.2 KiB
mul_mv_q6_k_f32_flat.cl11.0 KiB
mul_mv_q8_0_f32.cl3.6 KiB
mul_mv_q8_0_f32_flat.cl5.9 KiB
neg.cl3.3 KiB
norm.cl5.7 KiB
pad.cl1.4 KiB
quant_a_q8_1.cl1.3 KiB
relu.cl529 B
repeat.cl1.0 KiB
rms_norm.cl10.8 KiB
rope.cl24.5 KiB
scale.cl774 B
sdpa_xmem_f32_f16_os8.cl38.8 KiB
set_rows.cl20.3 KiB
sigmoid.cl876 B
silu.cl897 B
softmax_4_f16.cl3.1 KiB
softmax_4_f32.cl3.1 KiB
softmax_f16.cl3.1 KiB
softmax_f32.cl3.1 KiB
softplus.cl3.7 KiB
solve_tri.cl1.6 KiB
sqr.cl1.3 KiB
sqrt.cl1.4 KiB
ssm_conv.cl1.8 KiB
ssm_scan.cl8.7 KiB
sub.cl3.7 KiB
sum_rows.cl3.3 KiB
tanh.cl3.0 KiB
transpose.cl4.9 KiB
tri.cl898 B
tsembd.cl1.4 KiB
unary_ext.cl6.2 KiB
upscale.cl3.7 KiB