Llama-4-Maverick-17B-128E-Instruct-FP8 · original · l1234_popcorn_default_gpt

Llama-4-Maverick-17B-128E-Instruct-FP8

l1234_popcorn_default_gpt · original · 18 trajectories

← all variants ← originalgenerated 2026-05-28 03:33:15 UTC

L1 P8 8_Matmul_with_irregular_shapes_.

compile_fail —

L1 P27 27_SELU_.py

compile_fail —

L1 P29 29_Softplus.py

correct 0.84x 5.6μs

L1 P50 50_conv_standard_2D__square_inpu

in_progress —

L1 P79 79_conv_transposed_1D_asymmetric

compile_fail —

L1 P82 82_conv_depthwise_2D_square_inpu

compile_fail —

L1 P85 85_conv_depthwise_2D_asymmetric_

compile_fail —

L1 P93 93_masked_cumsum.py

compile_fail —

L2 P48 48_Conv3d_Scaling_Tanh_Multiply_

compile_fail —

L2 P51 51_Gemm_Subtract_GlobalAvgPool_L

in_progress —

L2 P66 66_Matmul_Dropout_Softmax.py

compile_fail —

L2 P71 71_Conv2d_Divide_LeakyReLU.py

in_progress —

L2 P88 88_Gemm_GroupNorm_Swish_Multiply

compile_fail —

L2 P94 94_Gemm_BiasAdd_Hardtanh_Mish_Gr

compile_fail —

L3 P7 7_GoogleNetInceptionV1.py

compile_fail —

L3 P19 19_MobileNetV1.py

in_progress —

L3 P49 49_Mamba2ReturnFinalState.py

compile_fail —

L4 P8 8_facebook-opt-1p3b_bs512_seq32.

compile_fail —