{"id":"W4405270867","doi":"10.1109/tpds.2024.3515804","title":"UMPIPE: Unequal Microbatches-Based Pipeline Parallelism for Deep Neural Network Training","year":2024,"lang":"en","type":"article","venue":"IEEE Transactions on Parallel and Distributed Systems","topic":"Neural Networks and Reservoir Computing","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Victoria","funders":"","keywords":"Computer science; Pipeline (software); Parallelism (grammar); Parallel computing; Artificial neural network; Training (meteorology); Data parallelism; Computer architecture; Artificial intelligence; Operating system","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":[0.0004392405,0.0003689373,0.0004366961,0.0001122548,0.0005758431,0.0007335474,0.0004633935,0.0001817303,0.000003908762],"category_scores_gemma":[0.000004714847,0.0003048176,0.00027106,0.0006010922,0.0000638264,0.000272378,0.000006852646,0.0003885965,0.0000101401],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005417351,"about_ca_system_score_gemma":0.00007929643,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004969883,"about_ca_topic_score_gemma":0.00003030762,"domain_scores_codex":[0.997519,0.0001459043,0.000581958,0.0007555965,0.0002592242,0.0007383542],"domain_scores_gemma":[0.9985703,0.0006078818,0.00009301562,0.0003931235,0.00008507977,0.0002505531],"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":[0.00005455208,0.00004792377,0.00000327808,0.000149702,0.00005220052,0.00004270939,0.0001554347,0.9756075,0.0000796114,0.0009923661,0.002037257,0.02077743],"study_design_scores_gemma":[0.0007809303,0.0001824145,0.00001096595,0.0002138738,0.00003902011,0.00006063034,0.00007576441,0.991022,0.00002581239,0.0002676483,0.006951059,0.0003698801],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.002920547,0.002858979,0.9878315,0.001435412,0.00356622,0.0006312419,0.0001358574,0.000591665,0.00002852852],"genre_scores_gemma":[0.9930627,0.00004160876,0.005698459,0.0002150217,0.0005365472,0.0001536689,0.00006071031,0.00003176276,0.0001995096],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9901422,"threshold_uncertainty_score":0.9999404,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03361677045423552,"score_gpt":0.259957126064931,"score_spread":0.2263403556106955,"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."}}