Skip to content

1602AnsselAttn1511

Petr Baudis edited this page Mar 17, 2016 · 5 revisions

Answer Sentence Selection attn1511 Parameters

attn1511 is a near-state-of-art model that models attention to select important tokens (well, effectively single one) in the answer.

yodaqa-curatedv2

1a51 config:

{"Ddim": "2", "adim": "0.5", "attn_mode": "sum", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "1", "focus_act": "softmax", "inp_e_dropout": "0", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xd551938>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xd55b2a8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "2"}

3x epoch_fract=1 ay_1a51d0_ef1 - 0.365413 (95% [0.231826, 0.499000]):

10658869.arien.ics.muni.cz.ay_1a51d0_ef1 etc.
[0.297544, 0.369631, 0.429065, ]

8x ay_1a51d0 - 0.467227 (95% [0.447812, 0.486642]):

10739077.arien.ics.muni.cz.ay_1a51d0 etc.
[0.468351, 0.474777, 0.423373, 0.462561, 0.457953, 0.454790, 0.507381, 0.488632, ]

17x ay_1a51_i12d12 - 0.475377 (95% [0.462617, 0.488138]):

10739085.arien.ics.muni.cz.ay_1a51_i12d12 etc.
[0.534213, 0.465073, 0.506191, 0.502081, 0.477935, 0.453235, 0.470745, 0.475528, 0.445433, 0.491368, 0.451581, 0.484796, 0.485237, 0.449552, 0.492834, 0.435048, 0.460565, ]

8x ay_1a51_i34d34 - 0.435022 (95% [0.395684, 0.474361]):

10776831.arien.ics.muni.cz.ay_1a51_i34d34 etc.
[0.444848, 0.516539, 0.364797, 0.459191, 0.445569, 0.451148, 0.434366, 0.363721, ]

8x ay_1a51_i45d45 - 0.430352 (95% [0.399448, 0.461255]):

10776811.arien.ics.muni.cz.ay_1a51_i45d45 etc.
[0.474579, 0.435509, 0.413227, 0.376304, 0.482622, 0.460141, 0.387999, 0.412432, ]

Dropout is not beneficial!

16x transfer learning baseline ay_1a51d0_c12s12dot - 0.336538 (95% [0.322145, 0.350931]):

10783229.arien.ics.muni.cz.ay_1a51d0_c12s12dot etc.
[0.343768, 0.330794, 0.299370, 0.336806, 0.373948, 0.346185, 0.313222, 0.359498, 0.393426, 0.350097, 0.330460, 0.340535, 0.290558, 0.293632, 0.350445, 0.331868, ]

Non-softmax focus experiments:

1x ay_1a51_i34d34 - 0.508993

1x ay_1a51_i34d34_sigmoid/norm (attn1511-4231da2347d7c633) focus_act="sigmoid/norm" - 0.408475

1x ay_1a51_i34d34_tanh/norm (attn1511-62ded3032528cf69) focus_act="tanh/norm" - 0.431351

1x ay_1a51_i34d34_dot (attn1511--5a2f60ee436e45d6) attn_mode='dot' - 0.418623

1x ay_1a51_i34d34_dot_tanh/norm (attn1511-49faaf7a6f7a336) focus_act="tanh/norm" "attn_mode='dot' - 0.434465

(focus_act='relu/norm' doesn't work at all, nan loss)


Some automated tuning:

rs = RandomSearch(modelname+'_rlog.txt',
                  dropout=[1/2], inp_e_dropout=[1/2], l2reg=[1e-6, 1e-5, 1e-4],
                  sdim=[1/6, 1/2, 1, 2, 3], cdim=[1, 2, 3],
                  project=[True, True, False], adim=[1/4, 1/2, 1],
                  ptscorer=[B.dot_ptscorer, B.mlp_ptscorer], Ddim=[1, 2])
-237586d4ccf6a502 0.300199 {"Ddim": "1", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xa7772a8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "2"}
-2d46e3e448e57755 0.303680 {"Ddim": "2", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "3"}
682ce75464858c0f 0.310376 {"Ddim": "1", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xa7772a8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "3"}
-44441b8fb5458b32 0.314066 {"Ddim": "2", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xb6dbc80>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xb6db578>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.166666666667"}
-6a7288b1b658445c 0.314850 {"Ddim": "1", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "2"}
221bcc94e030be75 0.314899 {"Ddim": "2", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xb0f1b18>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xb0f1410>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "3"}
17df4c702195bc47 0.317641 {"Ddim": "1", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function dot_ptscorer at 0xa7772a8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"}
-4214b98d70b16bba 0.321702 {"Ddim": "1", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function dot_ptscorer at 0xa7772a8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}
-4fd9ac7c3b435902 0.323867 {"Ddim": "1", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xb6dbc80>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xb6db578>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.166666666667"}
4ff090875d342b58 0.330390 {"Ddim": "1", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xa7772a8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "3"}
451c3abdcddccd02 0.331499 {"Ddim": "2", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "3"}
43a05a72ac4dc532 0.334182 {"Ddim": "2", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xb0f1b18>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function dot_ptscorer at 0xb0f1410>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"}
-526f5f1539629b83 0.334696 {"Ddim": "1", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xb0f1b18>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xb0f1410>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "2"}
-21aeb0cd3f4821ae 0.341496 {"Ddim": "1", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xb0f1b18>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xb0f1410>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "3"}
-6c9f22de476cd549 0.342109 {"Ddim": "2", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xb6dbc80>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function dot_ptscorer at 0xb6db578>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}
-402e30232b47d8a5 0.342497 {"Ddim": "1", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xb0f1b18>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xb0f1410>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}
-6f943055a7e6162d 0.343604 {"Ddim": "1", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function dot_ptscorer at 0xa7772a8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}
3afbd3c20fd52270 0.345473 {"Ddim": "2", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function dot_ptscorer at 0xa7772a8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}
-6139554d840d39d8 0.348934 {"Ddim": "1", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xa7772a8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}
-6f266fdac13ac89c 0.352246 {"Ddim": "1", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xb6dbc80>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xb6db578>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}
43900904c414e5e7 0.357368 {"Ddim": "2", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xb6dbc80>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xb6db578>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "2"}
6b3e9804a758510a 0.360251 {"Ddim": "2", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xa7772a8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "2"}
655686d098d09eb3 0.365905 {"Ddim": "1", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xa7772a8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}
-21b52b521afc381c 0.373282 {"Ddim": "2", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xb0f1b18>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xb0f1410>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.166666666667"}
-8edc729396e59c3 0.375685 {"Ddim": "2", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function dot_ptscorer at 0xa7772a8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.166666666667"}
1d2e5b5d20a9aad 0.383040 {"Ddim": "1", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "3"}
-59b810b573d5a42e 0.392960 {"Ddim": "1", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xa7772a8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"}
-179a38f149519152 0.406989 {"Ddim": "2", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xb6dbc80>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xb6db578>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}
-6673426e5de51073 0.408651 {"Ddim": "2", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xa7772a8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"}
-44a410b23968e36f 0.418191 {"Ddim": "2", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "3"}
-48e3733f7483046 0.428235 {"Ddim": "1", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xb6dbc80>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xb6db578>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"}
-20e0c6b30c9cdc2 0.431799 {"Ddim": "1", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xa7772a8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}
67a08ea3cbf71376 0.438215 {"Ddim": "2", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xb6dbc80>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xb6db578>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"}
6d982a0e7d128bc9 0.440476 {"Ddim": "2", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}
7d62578040c55a2e 0.441889 {"Ddim": "2", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "3"}
-407acc38fbe57a28 0.444542 {"Ddim": "1", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"}
6723cc7495a1ff5f 0.445879 {"Ddim": "1", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xa7772a8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.166666666667"}
-76e2cbd987a8c77c 0.451182 {"Ddim": "1", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xb0f1b18>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xb0f1500>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}
-882641795baf27b 0.452118 {"Ddim": "2", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xb0f1b18>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function mlp_ptscorer at 0xb0f1500>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.166666666667"}
-7f3b7db72a81bd8c 0.453427 {"Ddim": "1", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "2"}
7ab011a76ba8e2b4 0.457835 {"Ddim": "2", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function dot_ptscorer at 0xa7772a8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "2"}
-72f708c4fd5566b0 0.458563 {"Ddim": "1", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xb6dbc80>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xb6db668>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.166666666667"}
58b0577191159c4b 0.459125 {"Ddim": "2", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xb0f1b18>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function mlp_ptscorer at 0xb0f1500>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "2"}
-61fb87a960a785d7 0.459712 {"Ddim": "2", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "3"}
70699bd6b5ca09c5 0.460222 {"Ddim": "2", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.166666666667"}
d44d426669c9384 0.462679 {"Ddim": "2", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xb6dbc80>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function mlp_ptscorer at 0xb6db668>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.166666666667"}
-4bb2ace040dba258 0.463887 {"Ddim": "2", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "2"}
-3685de36feed67ca 0.465561 {"Ddim": "2", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"}
5a803581bf4fb39f 0.466202 {"Ddim": "2", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xb6dbc80>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xb6db668>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}
-48e3733f7483046 0.467167 {"Ddim": "1", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xb6dbc80>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xb6db578>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"}
3733da9c8d94769c 0.469074 {"Ddim": "2", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.166666666667"}
-424b1351aad34ae7 0.472965 {"Ddim": "1", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function dot_ptscorer at 0xa7772a8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "3"}
10c7b53d29bf5925 0.475236 {"Ddim": "1", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function dot_ptscorer at 0xa7772a8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "2"}
4c6e9aafb15292a1 0.475781 {"Ddim": "1", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "2"}
7bc85bc78bbed603 0.479035 {"Ddim": "2", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"}
2b2d2d4eb74c5388 0.482915 {"Ddim": "1", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xb6dbc80>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function mlp_ptscorer at 0xb6db668>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"}
-4132a603e4a3826 0.484734 {"Ddim": "2", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.166666666667"}
-5968e1e13a961db1 0.487529 {"Ddim": "1", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "False", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "3"}
-3cdb04537d0386f 0.491044 {"Ddim": "2", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}
3cb54f5e10b89fd2 0.492298 {"Ddim": "2", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}
-7d76d38159e390e8 0.494640 {"Ddim": "1", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}
-49b46d2eddffa2f6 0.495505 {"Ddim": "1", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xb0f1b18>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xb0f1500>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}
2cba6135cac89b16 0.496766 {"Ddim": "2", "adim": "0.25", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-06", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}
-198090a409940443 0.511067 {"Ddim": "1", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xb6dbc80>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xb6db668>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"}
3cb54f5e10b89fd2 0.515061 {"Ddim": "2", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xa7779b0>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xa777398>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}
-1e21422e8afa5ada 0.515602 {"Ddim": "1", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "1e-05", "loss": "<function ranknet at 0xb6dbc80>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xb6db668>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}

This one is interesting because it works nice with sdim=1/2.

-198090a409940443 0.511067 {"Ddim": "1", "adim": "1", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "3", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xb6dbc80>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xb6db668>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"}

17x ay_1a51_i12d12 - 0.475377 (95% [0.462617, 0.488138]):

10739085.arien.ics.muni.cz.ay_1a51_i12d12 etc.
[0.534213, 0.465073, 0.506191, 0.502081, 0.477935, 0.453235, 0.470745, 0.475528, 0.445433, 0.491368, 0.451581, 0.484796, 0.485237, 0.449552, 0.492834, 0.435048, 0.460565, ]

20x ay_1a51_i12d12_s1D1 - 0.484051 (95% [0.478339, 0.489763]):

10739132.arien.ics.muni.cz.ay_1a51_i12d12_s1D1 etc.
[0.475571, 0.478771, 0.481291, 0.477475, 0.494344, 0.465826, 0.503417, 0.478698, 0.486257, 0.500414, 0.488591, 0.485898, 0.478274, 0.471880, 0.491617, 0.484207, 0.508996, 0.456956, 0.479755, 0.492782, ]

22x ay_1a51_i12d12_s1D1_crelu - 0.476372 (95% [0.471518, 0.481226]):

10739140.arien.ics.muni.cz.ay_1a51_i12d12_s1D1_crelu etc.
[0.480890, 0.476020, 0.475552, 0.478026, 0.490553, 0.474001, 0.487312, 0.456299, 0.474825, 0.480515, 0.476974, 0.482320, 0.470621, 0.497697, 0.469605, 0.450156, 0.487766, 0.465605, 0.491329, 0.476710, 0.471843, 0.465563, ]

wang

Baseline - 2a51:

{"Ddim": "2", "adim": "0.5", "attn_mode": "sum", "balance_class": "False", "batch_size": "160", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0.5", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0.5", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "<function ranknet at 0xb8c4b18>", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xb8cc500>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "1"}

Baseline attn1511--48015bb276a5a773: 0.848168

No dropout attn1511-3555c0fde94bb456: 0.876154

1/2 dropout, transfer config attn1511--2bc024977a852a88 attn1511-48fbff404382d333: 0.770385, 0.807015

No dropout, transfer config (cdim=1/2 ptscorer=B.dot_ptscorer sdim=1/2) attn1511-747fbbd194b07973 attn1511--28de41e5fb6aa5f1: [0.825641, 0.784033]

ubuntu

10639380.arien.ics.muni.cz.a51_sCdot_a1

    RunID: attn1511-11e382196688634d  ({"Ddim": "2", "adim": "1", "attn_mode": "sum", "batch_size": "192", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0", "e_add_flags": "True", "embdim": "300", "focus_act": "softmax", "inp_e_dropout": "0.75", "l2reg": "0.0001", "loss": "binary_crossentropy", "mlpsum": "sum", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xd0da5f0>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"})

200064/200000 [==============================] - 2461s - loss: 0.3899                                                         val mrr 0.677996
    Predict&Eval (best epoch)
    Val MRR: 0.693203
    Val 2-R@1: 0.863037
    Val 10-R@1: 0.536912  10-R@2: 0.707464  10-R@5: 0.924284

10639381.arien.ics.muni.cz.a51_sCdot_c12

    RunID: attn1511--64338ac592ae6282  ({"Ddim": "2", "adim": "0.5", "attn_mode": "sum", "batch_size": "192", "cdim": "0.5", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0", "e_add_flags": "True", "embdim": "300", "focus_act": "softmax", "inp_e_dropout": "0.75", "l2reg": "0.0001", "loss": "binary_crossentropy", "mlpsum": "sum", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xbd915f0>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"})

200064/200000 [==============================] - 2350s - loss: 0.5379                                                         val mrr 0.570424
    Predict&Eval (best epoch)
    Val MRR: 0.609489
    Val 2-R@1: 0.775051
    Val 10-R@1: 0.461810  10-R@2: 0.585787  10-R@5: 0.810992

10649013.arien.ics.muni.cz.a51d0_sCdot_c2

    RunID: attn1511-711caef77096d728  ({"Ddim": "2", "adim": "0.5", "attn_mode": "sum", "batch_size": "192", "cdim": "2", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0", "e_add_flags": "True", "embdim": "300", "focus_act": "softmax", "inp_e_dropout": "0", "l2reg": "0.0001", "loss": "binary_crossentropy", "mlpsum": "sum", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xc5eac80>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"})

200064/200000 [==============================] - 2857s - loss: 0.1968                                                         val mrr 0.741777
    Predict&Eval (best epoch)
    Val MRR: 0.774761
    Val 2-R@1: 0.902096
    Val 10-R@1: 0.655470  10-R@2: 0.792280  10-R@5: 0.949796

10649015.arien.ics.muni.cz.a51d0_sCdot

    RunID: attn1511--156abb54ad7724db  ({"Ddim": "2", "adim": "0.5", "attn_mode": "sum", "batch_size": "192", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0", "e_add_flags": "True", "embdim": "300", "focus_act": "softmax", "inp_e_dropout": "0", "l2reg": "0.0001", "loss": "binary_crossentropy", "mlpsum": "sum", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xcba4050>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"})

200064/200000 [==============================] - 2406s - loss: 0.2013                                                         val mrr 0.743263
    Predict&Eval (best epoch)
    Val MRR: 0.774890
    Val 2-R@1: 0.902249
    Val 10-R@1: 0.656748  10-R@2: 0.792280  10-R@5: 0.947751

10649016.arien.ics.muni.cz.a51d0_sCdot_c12

    RunID: attn1511--56ec61ba4b2fffb5  ({"Ddim": "2", "adim": "0.5", "attn_mode": "sum", "batch_size": "192", "cdim": "0.5", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0", "e_add_flags": "True", "embdim": "300", "focus_act": "softmax", "inp_e_dropout": "0", "l2reg": "0.0001", "loss": "binary_crossentropy", "mlpsum": "sum", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xd9fa050>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"})

200064/200000 [==============================] - 2167s - loss: 0.1955                                                         val mrr 0.741723
    Predict&Eval (best epoch)
    Val MRR: 0.772967
    Val 2-R@1: 0.902658
    Val 10-R@1: 0.654039  10-R@2: 0.788088  10-R@5: 0.948364

10649019.arien.ics.muni.cz.a51d0_sCdot_a1

    RunID: attn1511--51a91352567c4134  ({"Ddim": "2", "adim": "1", "attn_mode": "sum", "batch_size": "192", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0", "e_add_flags": "True", "embdim": "300", "focus_act": "softmax", "inp_e_dropout": "0", "l2reg": "0.0001", "loss": "binary_crossentropy", "mlpsum": "sum", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function dot_ptscorer at 0xcb48050>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"})

200064/200000 [==============================] - 2537s - loss: 0.1934                                                         val mrr 0.740856
    Predict&Eval (best epoch)
    Val MRR: 0.776363
    Val 2-R@1: 0.901892
    Val 10-R@1: 0.658436  10-R@2: 0.793200  10-R@5: 0.949591

Our choice is 10649016 a51d0_sCdot_c12 because it seems about as good as the best models, but trains somewhat quicker: inp_e_dropout=0 dropout=0 cdim=1/2 ptscorer=B.dot_ptscorer sdim=1/2

Transfer learning for anssel may want MLP - re-benchmark with MLP classifier and no dropout: 10784608.arien.ics.muni.cz.a51d0_s12c11

RunID: attn1511--2484866139c00683  ({"Ddim": "2", "adim": "0.5", "attn_mode": "sum", "batch_size": "192", "cdim": "1", "cfiltlen": "3", "cnnact": "tanh", "cnninit": "glorot_uniform", "dropout": "0", "dropoutfix_inp": "0", "dropoutfix_rec": "0", "e_add_flags": "True", "embdim": "300", "epoch_fract": "0.25", "focus_act": "softmax", "inp_e_dropout": "0", "inp_w_dropout": "0", "l2reg": "0.0001", "loss": "binary_crossentropy", "mlpsum": "sum", "nb_epoch": "16", "pool_layer": "<class 'keras.layers.convolutional.MaxPooling1D'>", "project": "True", "ptscorer": "<function mlp_ptscorer at 0xc4f8de8>", "rnn": "<class 'keras.layers.recurrent.GRU'>", "rnnact": "tanh", "rnnbidi": "True", "rnnbidi_mode": "sum", "rnninit": "glorot_uniform", "rnnlevels": "1", "sdim": "0.5"})

200064/200000 [==============================] - 2480s - loss: 0.1920                                                         val mrr 0.734266
Predict&Eval (best epoch)
Val MRR: 0.755622
Val 2-R@1: 0.890542
Val 10-R@1: 0.632566  10-R@2: 0.767638  10-R@5: 0.936299
Clone this wiki locally