{"id":"W3022810465","doi":"10.1109/micro50266.2020.00071","title":"GOBO: Quantizing Attention-Based NLP Models for Low Latency and Energy Efficient Inference","year":2020,"lang":"en","type":"preprint","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":11,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science; Quantization (signal processing); Computation; Inference; Latency (audio); Language model; Computer engineering; Parallel computing; Computer hardware; Algorithm; Artificial intelligence","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0002392914,0.0003075638,0.000370906,0.0001321933,0.000124524,0.0003755156,0.0009872703,0.0002087819,0.000005283979],"category_scores_gemma":[0.00006177508,0.0002929407,0.0001603649,0.0001300839,0.0000312641,0.0001282256,0.001336166,0.0002506972,0.00000393601],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005202094,"about_ca_system_score_gemma":0.0002987998,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000352188,"about_ca_topic_score_gemma":0.00002855017,"domain_scores_codex":[0.9976737,0.000048741,0.0004550399,0.001159095,0.0003159399,0.0003474939],"domain_scores_gemma":[0.9984432,0.0002222759,0.000186809,0.0008041793,0.0001813491,0.0001622128],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000003838923,0.00003467812,0.00006353914,0.0002125208,0.00001537917,0.000003788551,0.0001325422,0.7051737,0.0001790914,0.2839743,0.00003488198,0.01017171],"study_design_scores_gemma":[0.0002668783,0.00002951954,0.0000474668,0.0001810927,0.00001247554,7.291446e-7,0.00000752695,0.9582473,0.0002162867,0.040597,0.00006257265,0.0003311741],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.009700606,0.0002006634,0.9862713,0.001879639,0.0006503268,0.0002467697,0.000009799576,0.0003250992,0.0007158319],"genre_scores_gemma":[0.7543646,0.00001906683,0.2448341,0.0004739932,0.00006808042,0.00008120115,0.00001493977,0.0000154848,0.0001285165],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.744664,"threshold_uncertainty_score":0.9999523,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06401963398422118,"score_gpt":0.2784773016417433,"score_spread":0.2144576676575221,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}