AI & ML interests
None yet
Organizations
None yet
clark12341/oct_11_v1_70B_Q4_1024seqLen_4bch_50epochs_32r16a_loss0-09_layersD_lr1e-3_mgn1.2_perDev4
Updated
clark12341/sep_30_v1_405B_Q4_1024seqLen_8bch_steps_16r32a_loss0-10_layersQKVO_lr1e-4
Updated
clark12341/sep_29_v1_8B_Q4_512seqLen_1bch_720steps_16r32a_loss0-10_layersQKVO_lr1e-4
Updated
clark12341/sep_26_v2_8B_Q4_1024seqLen_6bch_20steps_8r16a_loss0-15_layersQKVO_lr1e-3
Updated
clark12341/sep_26_v1_8B_Q4_1024seqLen_6bch_20steps_8r16a_loss0-19_layersQKVO_lr1e-3
Updated
clark12341/sep_26_v1_8B_Q4_1024seqLen_1bch_20steps_8r16a_loss0-52_layersQKVOUDG_lr1e-3
Updated
clark12341/sep_23_v10_70B_Q4_1024seqLen_1bch_144steps_8r16a_loss0-29_layersQKVO-up-down-gate_lr1e-3_H100
Updated
clark12341/sep_23_v9_70B_Q4_1024seqLen_1bch_144steps_8r16a_loss0-28_layersQKVO_lr1e-3_H100
Updated
clark12341/sep_23_v8_70B_Q4_1024seqLen_1bch_96steps_8r16a_loss0-36_layersQKVO_lr1e-3_H100
Updated
clark12341/sep_23_v7_70B_Q4_1024seqLen_1bch_48steps_8r16a_loss1-00_layersQKVO_lr1e-3_H100
Updated
clark12341/sep_23_v6_70B_Q4_1024seqLen_1bch_20steps_8r16a_loss0-08_layersQKVO_lr1e-3_H100
Updated
clark12341/sep_23_v4_70B_Q4_1024seqLen_1bch_20steps_8r16a_loss0-58_layersQKVO_lr1e-3
Updated
clark12341/sep_23_v5_70B_Q4_1024seqLen_1bch_20steps_8r16a_loss0-08_layersQKVO-up-down-gate_lr1e-3
Updated
clark12341/sep_23_v3_70B_Q4_1024seqLen_1bch_20steps_8r16a_loss0-18_layersQKVO-up-down-gate_lr1e-3
Updated
clark12341/sep_23_v2_70B_Q4_1024seqLen_6bch_20steps_8r16a_loss0-12_layersQKVO_lr1e-3
Updated
clark12341/sep_23_v1_405B_Q4_512seqLen_48bch_20steps_16r32a_loss0-21_layersAll-Linear_lr1e-4
Updated
clark12341/sep_22_v3Storage_70B_Q4_512seqLen_16bch_20steps_16r32a_loss0-06_layersAll-Linear_lr1e-3
Updated
clark12341/sep_22_v2Storage_70B_Q4_512seqLen_16bch_20steps_16r32a_loss0-06_layersAll-Linear_lr1e-3
Updated
clark12341/sep_22_v1_70B_Q4_512seqLen_16bch_20steps_16r32a_loss0-06_layersAll-Linear_lr1e-3
Updated
clark12341/sep_20_v12_70B_2048seqLen_48bch_300steps_16r32a_loss0-38_layersQVO_lr1e-5
Updated
clark12341/sep_20_v11_70B_2048seqLen_48bch_400steps_16r32a_loss0-41_layersQVO_lr1e-5
Updated
clark12341/sep_20_v10_70B_512seqLen_49bch_300steps_16r32a_loss0-60_layersQVO_lr1e-5
Updated
clark12341/sep_20_v9_70B_2048seqLen_48bch_300steps_32r64a_loss0-43_layersQVO_lr1e-5
Updated
clark12341/sep_20_v8_70B_2048seqLen_49bch_200steps_16r32a_loss0-15_layersQVO_lr5e-5
Updated
clark12341/sep_20_v7_70B_2048seqLen_49bch_120steps_16r32a_loss0-20_layersQVO_lr1e-4
Updated
clark12341/sep_20_v6_70B_2048seqLen_49bch_120steps_16r16a_loss0-50_layersQVO_lr1e-4
Updated
clark12341/sep_20_v5_70B_2048seqLen_49bch_60steps_16r32a_loss0-70_layersQVO_lr1e-4
Updated
clark12341/sep_20_v4_70B_2048seqLen_49bch_30steps_16r32a_loss1-85_layersQVO
Updated
clark12341/sep_20_v3_70B_2048seqLen_56bch_11steps_16r32a_loss0-56
Updated