quantization2 / lm-quant-toolkit /data-vis /data /stor /mxq /mxq1 /result-eval_model_mxq-misc.csv
chen459664's picture
Add files using upload-large-folder tool
21ad80b verified
Raw
History Blame Contribute Delete
13 kB
model,algo,config,config_detail,quant_duration,model_storage_size,load_mem_allot,load_mem_reserved,ppl_mem_allot,ppl_mem_reserved,leaderboard_mem_allot,leaderboard_mem_reserved,ppl_wikitext,ppl_c4,duration_wikitext,duration_c4,duration_leaderboard,ifeval,bbh,mathlevel5,gpqa,musr,mmlupro
Llama-2-7b-hf,mxq,2_48,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 2.48}",0,2615709056,2615461888,2634022912,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Llama-2-7b-hf,mxq,2_75,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 2.75}",0,2920633216,2920379392,2931818496,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Llama-2-7b-hf,mxq,3_00,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 3.0}",0,3196746496,3196483584,3210739712,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Llama-2-7b-hf,mxq,3_50,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 3.5}",0,3684375616,3684128768,3695181824,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Llama-2-7b-hf,mxq,3_76,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 3.76}",0,3888163904,3887900672,3900702720,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Llama-2-7b-hf,mxq,4_01,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 4.01}",0,3978525952,3978259456,3980394496,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Llama-2-7b-hf,mxq,4_50,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 4.5}",0,4167958272,4167675904,4177526784,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Llama-2-7b-hf,mxq,5_00,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 5.0}",0,4580368432,4580072448,4588568576,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Llama-2-13b-hf,mxq,2_48,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 2.48}",0,4749733824,4749479936,4762632192,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Llama-2-13b-hf,mxq,2_75,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 2.75}",0,5374828140,5374548992,5391777792,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Llama-2-13b-hf,mxq,3_00,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 3.0}",0,5911590440,5911293952,5918162944,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Llama-2-13b-hf,mxq,3_50,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 3.5}",0,6839601836,6839320576,6845104128,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Llama-2-13b-hf,mxq,3_76,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 3.76}",0,7260904572,7260601344,7266631680,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Llama-2-13b-hf,mxq,4_01,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 4.01}",0,7511320636,7511015424,7522484224,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Llama-2-13b-hf,mxq,4_50,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 4.5}",0,7793715072,7793388544,7799308288,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Llama-2-13b-hf,mxq,4_75,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 4.75}",0,8189094420,8188795904,8204058624,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Llama-2-13b-hf,mxq,5_00,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 5.0}",0,8586606072,8586331136,8604614656,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Meta-Llama-3-8B,mxq,2_48,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 2.48}",0,4340458828,4340162560,4355784704,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Meta-Llama-3-8B,mxq,2_75,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 2.75}",0,4684537156,4684237824,4697620480,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Meta-Llama-3-8B,mxq,3_00,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 3.0}",0,4914445264,4914144256,4924112896,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Meta-Llama-3-8B,mxq,3_50,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 3.5}",0,5564376980,5564070912,5582618624,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Meta-Llama-3-8B,mxq,3_76,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 3.76}",0,5687041832,5686735872,5697961984,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Meta-Llama-3-8B,mxq,4_01,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 4.01}",0,6004516788,6004209664,6010437632,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Meta-Llama-3-8B,mxq,4_50,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 4.5}",0,6028258732,6027952128,6029312000,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Meta-Llama-3-8B,mxq,4_75,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 4.75}",0,6246399508,6246091776,6253707264,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
Meta-Llama-3-8B,mxq,5_00,"{'weight_quant_params': {'nbits': 4, 'channel_wise': True, 'group_size': 64, 'optimize': True, 'round_zero': True, 'axis': 0, 'view_as_float': False}, 'scale_quant_params': {'nbits': 8, 'channel_wise': True, 'group_size': 128, 'optimize': False}, 'zero_quant_params': {'nbits': 8, 'channel_wise': False, 'group_size': None, 'optimize': False}, 'offload_meta': False, 'mixed': True, 'budget': 5.0}",0,6464412128,6464103424,6469713920,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0