# python3 -m espnet2.bin.s2t_train --use_preprocessor true --bpemodel data/token_list/bpe_unigram50000/bpe.model --token_type bpe --token_list data/token_list/bpe_unigram50000/tokens.txt --non_linguistic_symbols none --cleaner none --g2p none --valid_data_path_and_name_and_type dump/raw/cgn_valid/wav.scp,speech,kaldi_ark --valid_shape_file /ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//valid/speech_shape --resume true --fold_length 80000 --output_dir exp/s2t_lora_2 --config conf/lora_medium_2.yaml --frontend_conf fs=16k --normalize=global_mvn --normalize_conf stats_file=/ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//train/feats_stats.npz --train_data_path_and_name_and_type dump/raw/cgn_train/wav.scp,speech,kaldi_ark --train_shape_file /ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//train/speech_shape --fold_length 150 --train_data_path_and_name_and_type dump/raw/cgn_train/text.prev,text_prev,text --train_shape_file /ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//train/text_prev_shape.bpe --fold_length 150 --train_data_path_and_name_and_type dump/raw/cgn_train/text.ctc,text_ctc,text --train_shape_file /ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//train/text_ctc_shape.bpe --fold_length 150 --train_data_path_and_name_and_type dump/raw/cgn_train/text,text,text --train_shape_file /ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//train/text_shape.bpe --valid_data_path_and_name_and_type dump/raw/cgn_valid/text.prev,text_prev,text --valid_shape_file /ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//valid/text_prev_shape.bpe --valid_data_path_and_name_and_type dump/raw/cgn_valid/text.ctc,text_ctc,text --valid_shape_file /ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//valid/text_ctc_shape.bpe --valid_data_path_and_name_and_type dump/raw/cgn_valid/text,text,text --valid_shape_file /ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//valid/text_shape.bpe --model_conf extract_feats_in_collect_stats=false --ngpu 1 --multiprocessing_distributed True 
# Started at Wed May 21 13:18:27 EDT 2025
#
/ocean/projects/cis210027p/pwang10/miniconda3/envs/rota/bin/python3 /ocean/projects/cis210027p/pwang10/rota/espnet/espnet2/bin/s2t_train.py --use_preprocessor true --bpemodel data/token_list/bpe_unigram50000/bpe.model --token_type bpe --token_list data/token_list/bpe_unigram50000/tokens.txt --non_linguistic_symbols none --cleaner none --g2p none --valid_data_path_and_name_and_type dump/raw/cgn_valid/wav.scp,speech,kaldi_ark --valid_shape_file /ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//valid/speech_shape --resume true --fold_length 80000 --output_dir exp/s2t_lora_2 --config conf/lora_medium_2.yaml --frontend_conf fs=16k --normalize=global_mvn --normalize_conf stats_file=/ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//train/feats_stats.npz --train_data_path_and_name_and_type dump/raw/cgn_train/wav.scp,speech,kaldi_ark --train_shape_file /ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//train/speech_shape --fold_length 150 --train_data_path_and_name_and_type dump/raw/cgn_train/text.prev,text_prev,text --train_shape_file /ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//train/text_prev_shape.bpe --fold_length 150 --train_data_path_and_name_and_type dump/raw/cgn_train/text.ctc,text_ctc,text --train_shape_file /ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//train/text_ctc_shape.bpe --fold_length 150 --train_data_path_and_name_and_type dump/raw/cgn_train/text,text,text --train_shape_file /ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//train/text_shape.bpe --valid_data_path_and_name_and_type dump/raw/cgn_valid/text.prev,text_prev,text --valid_shape_file /ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//valid/text_prev_shape.bpe --valid_data_path_and_name_and_type dump/raw/cgn_valid/text.ctc,text_ctc,text --valid_shape_file /ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//valid/text_ctc_shape.bpe --valid_data_path_and_name_and_type dump/raw/cgn_valid/text,text,text --valid_shape_file /ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//valid/text_shape.bpe --model_conf extract_feats_in_collect_stats=false --ngpu 1 --multiprocessing_distributed True
[w005] 2025-05-21 13:21:28,896 (s2t:462) INFO: Vocabulary size: 50002
[w005] 2025-05-21 13:21:33,567 (e_branchformer_encoder:445) INFO: Gradient checkpoint layers: []
[w005] 2025-05-21 13:21:34,046 (transformer_decoder:103) INFO: Gradient checkpoint layers: []
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.embed.conv.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.embed.conv.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.embed.conv.2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.embed.conv.2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.embed.out.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.embed.out.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.cgmlp.channel_proj1.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.cgmlp.channel_proj1.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.cgmlp.csgu.norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.cgmlp.csgu.norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.cgmlp.csgu.conv.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.cgmlp.csgu.conv.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.cgmlp.channel_proj2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.cgmlp.channel_proj2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.feed_forward_macaron.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.feed_forward_macaron.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,978 (abs_task:1370) INFO: Setting encoder.encoders.0.feed_forward_macaron.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.0.feed_forward_macaron.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.0.norm_ff.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.0.norm_ff.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.0.norm_ff_macaron.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.0.norm_ff_macaron.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.0.norm_mha.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.0.norm_mha.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.0.norm_mlp.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.0.norm_mlp.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.0.norm_final.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.0.norm_final.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.0.depthwise_conv_fusion.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.0.depthwise_conv_fusion.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.0.merge_proj.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.0.merge_proj.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.cgmlp.channel_proj1.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.cgmlp.channel_proj1.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.cgmlp.csgu.norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.cgmlp.csgu.norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.cgmlp.csgu.conv.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.cgmlp.csgu.conv.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.cgmlp.channel_proj2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.cgmlp.channel_proj2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.feed_forward_macaron.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.feed_forward_macaron.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.feed_forward_macaron.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.feed_forward_macaron.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.norm_ff.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.norm_ff.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.norm_ff_macaron.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.norm_ff_macaron.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.norm_mha.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.norm_mha.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.norm_mlp.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.norm_mlp.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.norm_final.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.norm_final.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.depthwise_conv_fusion.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.depthwise_conv_fusion.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.merge_proj.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.1.merge_proj.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.2.attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.2.attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.2.attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.2.attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.2.attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.2.attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,979 (abs_task:1370) INFO: Setting encoder.encoders.2.attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.cgmlp.channel_proj1.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.cgmlp.channel_proj1.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.cgmlp.csgu.norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.cgmlp.csgu.norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.cgmlp.csgu.conv.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.cgmlp.csgu.conv.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.cgmlp.channel_proj2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.cgmlp.channel_proj2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.feed_forward_macaron.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.feed_forward_macaron.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.feed_forward_macaron.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.feed_forward_macaron.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.norm_ff.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.norm_ff.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.norm_ff_macaron.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.norm_ff_macaron.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.norm_mha.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.norm_mha.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.norm_mlp.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.norm_mlp.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.norm_final.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.norm_final.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.depthwise_conv_fusion.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.depthwise_conv_fusion.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.merge_proj.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.2.merge_proj.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.cgmlp.channel_proj1.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.cgmlp.channel_proj1.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.cgmlp.csgu.norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.cgmlp.csgu.norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.cgmlp.csgu.conv.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.cgmlp.csgu.conv.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.cgmlp.channel_proj2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.cgmlp.channel_proj2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.feed_forward_macaron.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.feed_forward_macaron.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.feed_forward_macaron.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.feed_forward_macaron.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.norm_ff.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.norm_ff.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.norm_ff_macaron.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.norm_ff_macaron.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.norm_mha.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,980 (abs_task:1370) INFO: Setting encoder.encoders.3.norm_mha.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.3.norm_mlp.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.3.norm_mlp.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.3.norm_final.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.3.norm_final.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.3.depthwise_conv_fusion.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.3.depthwise_conv_fusion.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.3.merge_proj.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.3.merge_proj.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.cgmlp.channel_proj1.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.cgmlp.channel_proj1.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.cgmlp.csgu.norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.cgmlp.csgu.norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.cgmlp.csgu.conv.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.cgmlp.csgu.conv.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.cgmlp.channel_proj2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.cgmlp.channel_proj2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.feed_forward_macaron.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.feed_forward_macaron.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.feed_forward_macaron.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.feed_forward_macaron.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.norm_ff.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.norm_ff.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.norm_ff_macaron.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.norm_ff_macaron.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.norm_mha.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.norm_mha.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.norm_mlp.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.norm_mlp.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.norm_final.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.norm_final.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.depthwise_conv_fusion.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.depthwise_conv_fusion.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.merge_proj.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.4.merge_proj.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.5.attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.5.attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.5.attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.5.attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.5.attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.5.attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.5.attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.5.attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.5.cgmlp.channel_proj1.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.5.cgmlp.channel_proj1.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.5.cgmlp.csgu.norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.5.cgmlp.csgu.norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.5.cgmlp.csgu.conv.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.5.cgmlp.csgu.conv.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,981 (abs_task:1370) INFO: Setting encoder.encoders.5.cgmlp.channel_proj2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.cgmlp.channel_proj2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.feed_forward_macaron.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.feed_forward_macaron.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.feed_forward_macaron.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.feed_forward_macaron.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.norm_ff.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.norm_ff.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.norm_ff_macaron.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.norm_ff_macaron.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.norm_mha.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.norm_mha.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.norm_mlp.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.norm_mlp.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.norm_final.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.norm_final.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.depthwise_conv_fusion.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.depthwise_conv_fusion.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.merge_proj.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.5.merge_proj.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.cgmlp.channel_proj1.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.cgmlp.channel_proj1.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.cgmlp.csgu.norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.cgmlp.csgu.norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.cgmlp.csgu.conv.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.cgmlp.csgu.conv.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.cgmlp.channel_proj2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.cgmlp.channel_proj2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.feed_forward_macaron.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.feed_forward_macaron.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.feed_forward_macaron.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.feed_forward_macaron.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.norm_ff.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.norm_ff.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.norm_ff_macaron.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.norm_ff_macaron.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.norm_mha.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.norm_mha.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.norm_mlp.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.norm_mlp.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.norm_final.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.norm_final.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.depthwise_conv_fusion.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.depthwise_conv_fusion.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.merge_proj.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.6.merge_proj.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.7.attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,982 (abs_task:1370) INFO: Setting encoder.encoders.7.attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.cgmlp.channel_proj1.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.cgmlp.channel_proj1.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.cgmlp.csgu.norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.cgmlp.csgu.norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.cgmlp.csgu.conv.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.cgmlp.csgu.conv.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.cgmlp.channel_proj2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.cgmlp.channel_proj2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.feed_forward_macaron.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.feed_forward_macaron.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.feed_forward_macaron.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.feed_forward_macaron.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.norm_ff.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.norm_ff.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.norm_ff_macaron.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.norm_ff_macaron.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.norm_mha.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.norm_mha.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.norm_mlp.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.norm_mlp.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.norm_final.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.norm_final.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.depthwise_conv_fusion.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.depthwise_conv_fusion.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.merge_proj.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.7.merge_proj.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.cgmlp.channel_proj1.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.cgmlp.channel_proj1.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.cgmlp.csgu.norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.cgmlp.csgu.norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.cgmlp.csgu.conv.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.cgmlp.csgu.conv.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.cgmlp.channel_proj2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.cgmlp.channel_proj2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.feed_forward_macaron.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.feed_forward_macaron.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.feed_forward_macaron.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.feed_forward_macaron.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.norm_ff.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,983 (abs_task:1370) INFO: Setting encoder.encoders.8.norm_ff.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.8.norm_ff_macaron.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.8.norm_ff_macaron.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.8.norm_mha.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.8.norm_mha.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.8.norm_mlp.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.8.norm_mlp.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.8.norm_final.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.8.norm_final.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.8.depthwise_conv_fusion.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.8.depthwise_conv_fusion.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.8.merge_proj.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.8.merge_proj.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.cgmlp.channel_proj1.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.cgmlp.channel_proj1.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.cgmlp.csgu.norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.cgmlp.csgu.norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.cgmlp.csgu.conv.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.cgmlp.csgu.conv.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.cgmlp.channel_proj2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.cgmlp.channel_proj2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.feed_forward_macaron.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.feed_forward_macaron.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.feed_forward_macaron.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.feed_forward_macaron.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.norm_ff.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.norm_ff.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.norm_ff_macaron.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.norm_ff_macaron.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.norm_mha.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.norm_mha.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.norm_mlp.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.norm_mlp.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.norm_final.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.norm_final.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.depthwise_conv_fusion.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.depthwise_conv_fusion.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.merge_proj.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.9.merge_proj.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.10.attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.10.attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.10.attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.10.attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.10.attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.10.attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.10.attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.10.attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.10.cgmlp.channel_proj1.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.10.cgmlp.channel_proj1.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,984 (abs_task:1370) INFO: Setting encoder.encoders.10.cgmlp.csgu.norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.cgmlp.csgu.norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.cgmlp.csgu.conv.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.cgmlp.csgu.conv.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.cgmlp.channel_proj2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.cgmlp.channel_proj2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.feed_forward_macaron.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.feed_forward_macaron.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.feed_forward_macaron.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.feed_forward_macaron.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.norm_ff.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.norm_ff.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.norm_ff_macaron.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.norm_ff_macaron.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.norm_mha.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.norm_mha.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.norm_mlp.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.norm_mlp.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.norm_final.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.norm_final.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.depthwise_conv_fusion.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.depthwise_conv_fusion.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.merge_proj.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.10.merge_proj.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.cgmlp.channel_proj1.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.cgmlp.channel_proj1.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.cgmlp.csgu.norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.cgmlp.csgu.norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.cgmlp.csgu.conv.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.cgmlp.csgu.conv.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.cgmlp.channel_proj2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.cgmlp.channel_proj2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.feed_forward_macaron.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.feed_forward_macaron.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.feed_forward_macaron.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.feed_forward_macaron.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.norm_ff.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.norm_ff.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.norm_ff_macaron.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.norm_ff_macaron.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.norm_mha.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.norm_mha.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.norm_mlp.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.norm_mlp.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.norm_final.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.norm_final.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.depthwise_conv_fusion.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,985 (abs_task:1370) INFO: Setting encoder.encoders.11.depthwise_conv_fusion.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.11.merge_proj.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.11.merge_proj.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.cgmlp.channel_proj1.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.cgmlp.channel_proj1.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.cgmlp.csgu.norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.cgmlp.csgu.norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.cgmlp.csgu.conv.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.cgmlp.csgu.conv.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.cgmlp.channel_proj2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.cgmlp.channel_proj2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.feed_forward_macaron.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.feed_forward_macaron.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.feed_forward_macaron.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.feed_forward_macaron.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.norm_ff.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.norm_ff.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.norm_ff_macaron.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.norm_ff_macaron.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.norm_mha.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.norm_mha.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.norm_mlp.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.norm_mlp.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.norm_final.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.norm_final.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.depthwise_conv_fusion.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.depthwise_conv_fusion.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.merge_proj.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.12.merge_proj.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.cgmlp.channel_proj1.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.cgmlp.channel_proj1.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.cgmlp.csgu.norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.cgmlp.csgu.norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.cgmlp.csgu.conv.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.cgmlp.csgu.conv.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.cgmlp.channel_proj2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.cgmlp.channel_proj2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,986 (abs_task:1370) INFO: Setting encoder.encoders.13.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.13.feed_forward_macaron.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.13.feed_forward_macaron.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.13.feed_forward_macaron.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.13.feed_forward_macaron.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.13.norm_ff.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.13.norm_ff.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.13.norm_ff_macaron.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.13.norm_ff_macaron.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.13.norm_mha.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.13.norm_mha.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.13.norm_mlp.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.13.norm_mlp.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.13.norm_final.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.13.norm_final.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.13.depthwise_conv_fusion.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.13.depthwise_conv_fusion.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.13.merge_proj.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.13.merge_proj.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.cgmlp.channel_proj1.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.cgmlp.channel_proj1.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.cgmlp.csgu.norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.cgmlp.csgu.norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.cgmlp.csgu.conv.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.cgmlp.csgu.conv.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.cgmlp.channel_proj2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.cgmlp.channel_proj2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.feed_forward_macaron.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.feed_forward_macaron.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.feed_forward_macaron.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.feed_forward_macaron.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.norm_ff.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.norm_ff.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.norm_ff_macaron.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.norm_ff_macaron.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.norm_mha.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.norm_mha.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.norm_mlp.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.norm_mlp.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.norm_final.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.norm_final.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.depthwise_conv_fusion.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.depthwise_conv_fusion.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.merge_proj.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.14.merge_proj.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.15.attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.15.attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.15.attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.15.attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,987 (abs_task:1370) INFO: Setting encoder.encoders.15.attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.cgmlp.channel_proj1.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.cgmlp.channel_proj1.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.cgmlp.csgu.norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.cgmlp.csgu.norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.cgmlp.csgu.conv.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.cgmlp.csgu.conv.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.cgmlp.channel_proj2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.cgmlp.channel_proj2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.feed_forward_macaron.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.feed_forward_macaron.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.feed_forward_macaron.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.feed_forward_macaron.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.norm_ff.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.norm_ff.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.norm_ff_macaron.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.norm_ff_macaron.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.norm_mha.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.norm_mha.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.norm_mlp.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.norm_mlp.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.norm_final.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.norm_final.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.depthwise_conv_fusion.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.depthwise_conv_fusion.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.merge_proj.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.15.merge_proj.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.cgmlp.channel_proj1.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.cgmlp.channel_proj1.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.cgmlp.csgu.norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.cgmlp.csgu.norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.cgmlp.csgu.conv.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.cgmlp.csgu.conv.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.cgmlp.channel_proj2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.cgmlp.channel_proj2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.feed_forward_macaron.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.feed_forward_macaron.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.feed_forward_macaron.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.feed_forward_macaron.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.norm_ff.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.norm_ff.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.norm_ff_macaron.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,988 (abs_task:1370) INFO: Setting encoder.encoders.16.norm_ff_macaron.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.16.norm_mha.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.16.norm_mha.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.16.norm_mlp.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.16.norm_mlp.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.16.norm_final.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.16.norm_final.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.16.depthwise_conv_fusion.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.16.depthwise_conv_fusion.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.16.merge_proj.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.16.merge_proj.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.cgmlp.channel_proj1.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.cgmlp.channel_proj1.0.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.cgmlp.csgu.norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.cgmlp.csgu.norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.cgmlp.csgu.conv.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.cgmlp.csgu.conv.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.cgmlp.channel_proj2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.cgmlp.channel_proj2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.feed_forward_macaron.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.feed_forward_macaron.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.feed_forward_macaron.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.feed_forward_macaron.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.norm_ff.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.norm_ff.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.norm_ff_macaron.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.norm_ff_macaron.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.norm_mha.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.norm_mha.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.norm_mlp.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.norm_mlp.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.norm_final.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.norm_final.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.depthwise_conv_fusion.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.depthwise_conv_fusion.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.merge_proj.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.encoders.17.merge_proj.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.after_norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,989 (abs_task:1370) INFO: Setting encoder.after_norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.embed.0.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.after_norm.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.after_norm.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.output_layer.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.output_layer.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.self_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.self_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.self_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.self_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.self_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.self_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.self_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.self_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.src_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.src_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.src_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.src_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.src_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.src_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.src_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.src_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.norm1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.norm1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.norm2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.norm2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.norm3.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.0.norm3.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.1.self_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.1.self_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.1.self_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.1.self_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.1.self_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.1.self_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.1.self_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.1.self_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.1.src_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.1.src_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.1.src_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.1.src_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,991 (abs_task:1370) INFO: Setting decoder.decoders.1.src_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.1.src_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.1.src_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.1.src_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.1.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.1.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.1.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.1.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.1.norm1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.1.norm1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.1.norm2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.1.norm2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.1.norm3.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.1.norm3.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.self_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.self_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.self_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.self_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.self_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.self_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.self_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.self_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.src_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.src_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.src_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.src_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.src_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.src_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.src_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.src_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.norm1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.norm1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.norm2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.norm2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.norm3.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.2.norm3.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.self_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.self_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.self_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.self_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.self_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.self_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.self_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.self_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.src_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.src_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.src_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.src_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.src_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.src_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.src_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.src_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.norm1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.norm1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,992 (abs_task:1370) INFO: Setting decoder.decoders.3.norm2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.3.norm2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.3.norm3.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.3.norm3.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.self_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.self_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.self_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.self_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.self_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.self_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.self_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.self_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.src_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.src_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.src_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.src_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.src_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.src_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.src_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.src_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.norm1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.norm1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.norm2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.norm2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.norm3.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.4.norm3.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.self_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.self_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.self_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.self_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.self_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.self_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.self_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.self_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.src_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.src_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.src_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.src_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.src_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.src_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.src_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.src_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.norm1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.norm1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.norm2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.norm2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.norm3.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.5.norm3.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.6.self_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.6.self_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.6.self_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.6.self_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.6.self_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.6.self_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,993 (abs_task:1370) INFO: Setting decoder.decoders.6.self_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.6.self_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.6.src_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.6.src_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.6.src_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.6.src_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.6.src_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.6.src_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.6.src_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.6.src_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.6.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.6.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.6.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.6.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.6.norm1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.6.norm1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.6.norm2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.6.norm2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.6.norm3.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.6.norm3.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.self_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.self_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.self_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.self_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.self_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.self_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.self_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.self_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.src_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.src_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.src_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.src_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.src_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.src_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.src_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.src_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.norm1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.norm1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.norm2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.norm2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.norm3.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.7.norm3.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.8.self_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.8.self_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.8.self_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.8.self_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.8.self_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.8.self_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.8.self_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.8.self_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.8.src_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.8.src_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.8.src_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.8.src_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.8.src_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.8.src_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.8.src_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,994 (abs_task:1370) INFO: Setting decoder.decoders.8.src_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.8.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.8.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.8.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.8.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.8.norm1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.8.norm1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.8.norm2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.8.norm2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.8.norm3.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.8.norm3.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.self_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.self_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.self_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.self_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.self_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.self_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.self_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.self_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.src_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.src_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.src_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.src_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.src_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.src_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.src_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.src_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.norm1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.norm1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.norm2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.norm2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.norm3.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.9.norm3.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.self_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.self_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.self_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.self_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.self_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.self_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.self_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.self_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.src_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.src_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.src_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.src_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.src_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.src_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.src_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.src_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.norm1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.norm1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.norm2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.norm2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,995 (abs_task:1370) INFO: Setting decoder.decoders.10.norm3.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.10.norm3.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.self_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.self_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.self_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.self_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.self_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.self_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.self_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.self_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.src_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.src_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.src_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.src_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.src_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.src_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.src_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.src_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.norm1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.norm1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.norm2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.norm2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.norm3.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.11.norm3.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.self_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.self_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.self_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.self_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.self_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.self_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.self_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.self_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.src_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.src_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.src_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.src_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.src_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.src_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.src_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.src_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.norm1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.norm1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.norm2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.norm2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.norm3.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.12.norm3.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.13.self_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.13.self_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.13.self_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.13.self_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.13.self_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,996 (abs_task:1370) INFO: Setting decoder.decoders.13.self_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.self_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.self_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.src_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.src_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.src_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.src_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.src_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.src_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.src_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.src_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.norm1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.norm1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.norm2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.norm2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.norm3.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.13.norm3.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.self_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.self_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.self_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.self_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.self_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.self_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.self_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.self_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.src_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.src_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.src_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.src_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.src_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.src_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.src_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.src_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.norm1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.norm1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.norm2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.norm2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.norm3.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.14.norm3.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.15.self_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.15.self_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.15.self_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.15.self_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.15.self_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.15.self_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.15.self_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.15.self_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.15.src_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.15.src_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.15.src_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.15.src_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.15.src_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.15.src_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.15.src_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.15.src_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,997 (abs_task:1370) INFO: Setting decoder.decoders.15.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.15.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.15.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.15.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.15.norm1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.15.norm1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.15.norm2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.15.norm2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.15.norm3.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.15.norm3.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.self_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.self_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.self_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.self_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.self_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.self_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.self_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.self_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.src_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.src_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.src_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.src_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.src_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.src_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.src_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.src_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.norm1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.norm1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.norm2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.norm2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.norm3.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.16.norm3.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.self_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.self_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.self_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.self_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.self_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.self_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.self_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.self_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.src_attn.linear_q.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.src_attn.linear_q.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.src_attn.linear_k.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.src_attn.linear_k.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.src_attn.linear_v.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.src_attn.linear_v.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.src_attn.linear_out.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.src_attn.linear_out.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.feed_forward.w_1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.feed_forward.w_1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.feed_forward.w_2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.feed_forward.w_2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.norm1.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.norm1.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.norm2.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.norm2.bias.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.norm3.weight.requires_grad = False
[w005] 2025-05-21 13:21:36,998 (abs_task:1370) INFO: Setting decoder.decoders.17.norm3.bias.requires_grad = False
[w005] 2025-05-21 13:21:37,001 (abs_task:1370) INFO: Setting ctc.ctc_lo.weight.requires_grad = False
[w005] 2025-05-21 13:21:37,001 (abs_task:1370) INFO: Setting ctc.ctc_lo.bias.requires_grad = False
[w005] 2025-05-21 13:21:57,578 (abs_task:1398) INFO: pytorch.version=2.1.2, cuda.available=True, cudnn.version=8902, cudnn.benchmark=False, cudnn.deterministic=True
[w005] 2025-05-21 13:21:57,589 (abs_task:1399) INFO: Model structure:
ESPnetS2TModel(
  (frontend): DefaultFrontend(
    (stft): Stft(n_fft=512, win_length=400, hop_length=160, center=True, normalized=False, onesided=True)
    (frontend): Frontend()
    (logmel): LogMel(sr=16000, n_fft=512, n_mels=80, fmin=0, fmax=8000.0, htk=False)
  )
  (specaug): SpecAug(
    (freq_mask): MaskAlongAxis(mask_width_range=[0, 27], num_mask=2, axis=freq)
    (time_mask): MaskAlongAxisVariableMaxWidth(mask_width_ratio_range=[0.0, 0.05], num_mask=10, axis=time)
  )
  (normalize): GlobalMVN(stats_file=/ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium/train/feats_stats.npz, norm_means=True, norm_vars=True)
  (encoder): EBranchformerEncoder(
    (embed): Conv2dSubsampling(
      (conv): Sequential(
        (0): Conv2d(1, 1024, kernel_size=(3, 3), stride=(2, 2))
        (1): ReLU()
        (2): Conv2d(1024, 1024, kernel_size=(3, 3), stride=(2, 2))
        (3): ReLU()
      )
      (out): Sequential(
        (0): Linear(in_features=19456, out_features=1024, bias=True)
        (1): PositionalEncoding(
          (dropout): Dropout(p=0.1, inplace=False)
        )
      )
    )
    (encoders): MultiSequential(
      (0): EBranchformerEncoderLayer(
        (attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (cgmlp): ConvolutionalGatingMLP(
          (channel_proj1): Sequential(
            (0): Linear(in_features=1024, out_features=4096, bias=True)
            (1): GELU(approximate='none')
          )
          (csgu): ConvolutionalSpatialGatingUnit(
            (norm): LayerNorm((2048,), eps=1e-12, elementwise_affine=True)
            (conv): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
            (act): Identity()
            (dropout): Dropout(p=0.1, inplace=False)
          )
          (channel_proj2): Linear(in_features=2048, out_features=1024, bias=True)
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (feed_forward_macaron): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (norm_ff): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_ff_macaron): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mha): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mlp): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_final): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
        (depthwise_conv_fusion): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
        (merge_proj): Linear(in_features=2048, out_features=1024, bias=True)
      )
      (1): EBranchformerEncoderLayer(
        (attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (cgmlp): ConvolutionalGatingMLP(
          (channel_proj1): Sequential(
            (0): Linear(in_features=1024, out_features=4096, bias=True)
            (1): GELU(approximate='none')
          )
          (csgu): ConvolutionalSpatialGatingUnit(
            (norm): LayerNorm((2048,), eps=1e-12, elementwise_affine=True)
            (conv): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
            (act): Identity()
            (dropout): Dropout(p=0.1, inplace=False)
          )
          (channel_proj2): Linear(in_features=2048, out_features=1024, bias=True)
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (feed_forward_macaron): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (norm_ff): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_ff_macaron): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mha): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mlp): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_final): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
        (depthwise_conv_fusion): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
        (merge_proj): Linear(in_features=2048, out_features=1024, bias=True)
      )
      (2): EBranchformerEncoderLayer(
        (attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (cgmlp): ConvolutionalGatingMLP(
          (channel_proj1): Sequential(
            (0): Linear(in_features=1024, out_features=4096, bias=True)
            (1): GELU(approximate='none')
          )
          (csgu): ConvolutionalSpatialGatingUnit(
            (norm): LayerNorm((2048,), eps=1e-12, elementwise_affine=True)
            (conv): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
            (act): Identity()
            (dropout): Dropout(p=0.1, inplace=False)
          )
          (channel_proj2): Linear(in_features=2048, out_features=1024, bias=True)
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (feed_forward_macaron): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (norm_ff): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_ff_macaron): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mha): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mlp): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_final): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
        (depthwise_conv_fusion): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
        (merge_proj): Linear(in_features=2048, out_features=1024, bias=True)
      )
      (3): EBranchformerEncoderLayer(
        (attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (cgmlp): ConvolutionalGatingMLP(
          (channel_proj1): Sequential(
            (0): Linear(in_features=1024, out_features=4096, bias=True)
            (1): GELU(approximate='none')
          )
          (csgu): ConvolutionalSpatialGatingUnit(
            (norm): LayerNorm((2048,), eps=1e-12, elementwise_affine=True)
            (conv): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
            (act): Identity()
            (dropout): Dropout(p=0.1, inplace=False)
          )
          (channel_proj2): Linear(in_features=2048, out_features=1024, bias=True)
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (feed_forward_macaron): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (norm_ff): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_ff_macaron): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mha): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mlp): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_final): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
        (depthwise_conv_fusion): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
        (merge_proj): Linear(in_features=2048, out_features=1024, bias=True)
      )
      (4): EBranchformerEncoderLayer(
        (attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (cgmlp): ConvolutionalGatingMLP(
          (channel_proj1): Sequential(
            (0): Linear(in_features=1024, out_features=4096, bias=True)
            (1): GELU(approximate='none')
          )
          (csgu): ConvolutionalSpatialGatingUnit(
            (norm): LayerNorm((2048,), eps=1e-12, elementwise_affine=True)
            (conv): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
            (act): Identity()
            (dropout): Dropout(p=0.1, inplace=False)
          )
          (channel_proj2): Linear(in_features=2048, out_features=1024, bias=True)
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (feed_forward_macaron): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (norm_ff): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_ff_macaron): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mha): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mlp): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_final): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
        (depthwise_conv_fusion): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
        (merge_proj): Linear(in_features=2048, out_features=1024, bias=True)
      )
      (5): EBranchformerEncoderLayer(
        (attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (cgmlp): ConvolutionalGatingMLP(
          (channel_proj1): Sequential(
            (0): Linear(in_features=1024, out_features=4096, bias=True)
            (1): GELU(approximate='none')
          )
          (csgu): ConvolutionalSpatialGatingUnit(
            (norm): LayerNorm((2048,), eps=1e-12, elementwise_affine=True)
            (conv): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
            (act): Identity()
            (dropout): Dropout(p=0.1, inplace=False)
          )
          (channel_proj2): Linear(in_features=2048, out_features=1024, bias=True)
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (feed_forward_macaron): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (norm_ff): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_ff_macaron): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mha): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mlp): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_final): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
        (depthwise_conv_fusion): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
        (merge_proj): Linear(in_features=2048, out_features=1024, bias=True)
      )
      (6): EBranchformerEncoderLayer(
        (attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (cgmlp): ConvolutionalGatingMLP(
          (channel_proj1): Sequential(
            (0): Linear(in_features=1024, out_features=4096, bias=True)
            (1): GELU(approximate='none')
          )
          (csgu): ConvolutionalSpatialGatingUnit(
            (norm): LayerNorm((2048,), eps=1e-12, elementwise_affine=True)
            (conv): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
            (act): Identity()
            (dropout): Dropout(p=0.1, inplace=False)
          )
          (channel_proj2): Linear(in_features=2048, out_features=1024, bias=True)
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (feed_forward_macaron): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (norm_ff): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_ff_macaron): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mha): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mlp): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_final): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
        (depthwise_conv_fusion): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
        (merge_proj): Linear(in_features=2048, out_features=1024, bias=True)
      )
      (7): EBranchformerEncoderLayer(
        (attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (cgmlp): ConvolutionalGatingMLP(
          (channel_proj1): Sequential(
            (0): Linear(in_features=1024, out_features=4096, bias=True)
            (1): GELU(approximate='none')
          )
          (csgu): ConvolutionalSpatialGatingUnit(
            (norm): LayerNorm((2048,), eps=1e-12, elementwise_affine=True)
            (conv): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
            (act): Identity()
            (dropout): Dropout(p=0.1, inplace=False)
          )
          (channel_proj2): Linear(in_features=2048, out_features=1024, bias=True)
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (feed_forward_macaron): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (norm_ff): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_ff_macaron): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mha): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mlp): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_final): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
        (depthwise_conv_fusion): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
        (merge_proj): Linear(in_features=2048, out_features=1024, bias=True)
      )
      (8): EBranchformerEncoderLayer(
        (attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (cgmlp): ConvolutionalGatingMLP(
          (channel_proj1): Sequential(
            (0): Linear(in_features=1024, out_features=4096, bias=True)
            (1): GELU(approximate='none')
          )
          (csgu): ConvolutionalSpatialGatingUnit(
            (norm): LayerNorm((2048,), eps=1e-12, elementwise_affine=True)
            (conv): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
            (act): Identity()
            (dropout): Dropout(p=0.1, inplace=False)
          )
          (channel_proj2): Linear(in_features=2048, out_features=1024, bias=True)
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (feed_forward_macaron): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (norm_ff): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_ff_macaron): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mha): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mlp): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_final): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
        (depthwise_conv_fusion): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
        (merge_proj): Linear(in_features=2048, out_features=1024, bias=True)
      )
      (9): EBranchformerEncoderLayer(
        (attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (cgmlp): ConvolutionalGatingMLP(
          (channel_proj1): Sequential(
            (0): Linear(in_features=1024, out_features=4096, bias=True)
            (1): GELU(approximate='none')
          )
          (csgu): ConvolutionalSpatialGatingUnit(
            (norm): LayerNorm((2048,), eps=1e-12, elementwise_affine=True)
            (conv): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
            (act): Identity()
            (dropout): Dropout(p=0.1, inplace=False)
          )
          (channel_proj2): Linear(in_features=2048, out_features=1024, bias=True)
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (feed_forward_macaron): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (norm_ff): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_ff_macaron): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mha): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mlp): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_final): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
        (depthwise_conv_fusion): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
        (merge_proj): Linear(in_features=2048, out_features=1024, bias=True)
      )
      (10): EBranchformerEncoderLayer(
        (attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (cgmlp): ConvolutionalGatingMLP(
          (channel_proj1): Sequential(
            (0): Linear(in_features=1024, out_features=4096, bias=True)
            (1): GELU(approximate='none')
          )
          (csgu): ConvolutionalSpatialGatingUnit(
            (norm): LayerNorm((2048,), eps=1e-12, elementwise_affine=True)
            (conv): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
            (act): Identity()
            (dropout): Dropout(p=0.1, inplace=False)
          )
          (channel_proj2): Linear(in_features=2048, out_features=1024, bias=True)
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (feed_forward_macaron): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (norm_ff): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_ff_macaron): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mha): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mlp): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_final): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
        (depthwise_conv_fusion): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
        (merge_proj): Linear(in_features=2048, out_features=1024, bias=True)
      )
      (11): EBranchformerEncoderLayer(
        (attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (cgmlp): ConvolutionalGatingMLP(
          (channel_proj1): Sequential(
            (0): Linear(in_features=1024, out_features=4096, bias=True)
            (1): GELU(approximate='none')
          )
          (csgu): ConvolutionalSpatialGatingUnit(
            (norm): LayerNorm((2048,), eps=1e-12, elementwise_affine=True)
            (conv): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
            (act): Identity()
            (dropout): Dropout(p=0.1, inplace=False)
          )
          (channel_proj2): Linear(in_features=2048, out_features=1024, bias=True)
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (feed_forward_macaron): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (norm_ff): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_ff_macaron): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mha): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mlp): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_final): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
        (depthwise_conv_fusion): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
        (merge_proj): Linear(in_features=2048, out_features=1024, bias=True)
      )
      (12): EBranchformerEncoderLayer(
        (attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (cgmlp): ConvolutionalGatingMLP(
          (channel_proj1): Sequential(
            (0): Linear(in_features=1024, out_features=4096, bias=True)
            (1): GELU(approximate='none')
          )
          (csgu): ConvolutionalSpatialGatingUnit(
            (norm): LayerNorm((2048,), eps=1e-12, elementwise_affine=True)
            (conv): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
            (act): Identity()
            (dropout): Dropout(p=0.1, inplace=False)
          )
          (channel_proj2): Linear(in_features=2048, out_features=1024, bias=True)
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (feed_forward_macaron): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (norm_ff): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_ff_macaron): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mha): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mlp): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_final): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
        (depthwise_conv_fusion): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
        (merge_proj): Linear(in_features=2048, out_features=1024, bias=True)
      )
      (13): EBranchformerEncoderLayer(
        (attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (cgmlp): ConvolutionalGatingMLP(
          (channel_proj1): Sequential(
            (0): Linear(in_features=1024, out_features=4096, bias=True)
            (1): GELU(approximate='none')
          )
          (csgu): ConvolutionalSpatialGatingUnit(
            (norm): LayerNorm((2048,), eps=1e-12, elementwise_affine=True)
            (conv): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
            (act): Identity()
            (dropout): Dropout(p=0.1, inplace=False)
          )
          (channel_proj2): Linear(in_features=2048, out_features=1024, bias=True)
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (feed_forward_macaron): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (norm_ff): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_ff_macaron): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mha): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mlp): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_final): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
        (depthwise_conv_fusion): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
        (merge_proj): Linear(in_features=2048, out_features=1024, bias=True)
      )
      (14): EBranchformerEncoderLayer(
        (attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (cgmlp): ConvolutionalGatingMLP(
          (channel_proj1): Sequential(
            (0): Linear(in_features=1024, out_features=4096, bias=True)
            (1): GELU(approximate='none')
          )
          (csgu): ConvolutionalSpatialGatingUnit(
            (norm): LayerNorm((2048,), eps=1e-12, elementwise_affine=True)
            (conv): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
            (act): Identity()
            (dropout): Dropout(p=0.1, inplace=False)
          )
          (channel_proj2): Linear(in_features=2048, out_features=1024, bias=True)
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (feed_forward_macaron): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (norm_ff): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_ff_macaron): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mha): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mlp): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_final): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
        (depthwise_conv_fusion): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
        (merge_proj): Linear(in_features=2048, out_features=1024, bias=True)
      )
      (15): EBranchformerEncoderLayer(
        (attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (cgmlp): ConvolutionalGatingMLP(
          (channel_proj1): Sequential(
            (0): Linear(in_features=1024, out_features=4096, bias=True)
            (1): GELU(approximate='none')
          )
          (csgu): ConvolutionalSpatialGatingUnit(
            (norm): LayerNorm((2048,), eps=1e-12, elementwise_affine=True)
            (conv): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
            (act): Identity()
            (dropout): Dropout(p=0.1, inplace=False)
          )
          (channel_proj2): Linear(in_features=2048, out_features=1024, bias=True)
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (feed_forward_macaron): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (norm_ff): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_ff_macaron): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mha): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mlp): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_final): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
        (depthwise_conv_fusion): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
        (merge_proj): Linear(in_features=2048, out_features=1024, bias=True)
      )
      (16): EBranchformerEncoderLayer(
        (attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (cgmlp): ConvolutionalGatingMLP(
          (channel_proj1): Sequential(
            (0): Linear(in_features=1024, out_features=4096, bias=True)
            (1): GELU(approximate='none')
          )
          (csgu): ConvolutionalSpatialGatingUnit(
            (norm): LayerNorm((2048,), eps=1e-12, elementwise_affine=True)
            (conv): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
            (act): Identity()
            (dropout): Dropout(p=0.1, inplace=False)
          )
          (channel_proj2): Linear(in_features=2048, out_features=1024, bias=True)
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (feed_forward_macaron): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (norm_ff): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_ff_macaron): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mha): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mlp): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_final): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
        (depthwise_conv_fusion): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
        (merge_proj): Linear(in_features=2048, out_features=1024, bias=True)
      )
      (17): EBranchformerEncoderLayer(
        (attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (cgmlp): ConvolutionalGatingMLP(
          (channel_proj1): Sequential(
            (0): Linear(in_features=1024, out_features=4096, bias=True)
            (1): GELU(approximate='none')
          )
          (csgu): ConvolutionalSpatialGatingUnit(
            (norm): LayerNorm((2048,), eps=1e-12, elementwise_affine=True)
            (conv): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
            (act): Identity()
            (dropout): Dropout(p=0.1, inplace=False)
          )
          (channel_proj2): Linear(in_features=2048, out_features=1024, bias=True)
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (feed_forward_macaron): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): Swish()
        )
        (norm_ff): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_ff_macaron): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mha): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_mlp): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm_final): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
        (depthwise_conv_fusion): Conv1d(2048, 2048, kernel_size=(31,), stride=(1,), padding=(15,), groups=2048)
        (merge_proj): Linear(in_features=2048, out_features=1024, bias=True)
      )
    )
    (after_norm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
  )
  (decoder): TransformerDecoder(
    (embed): Sequential(
      (0): Embedding(50002, 1024)
      (1): PositionalEncoding(
        (dropout): Dropout(p=0.1, inplace=False)
      )
    )
    (after_norm): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
    (output_layer): Linear(in_features=1024, out_features=50002, bias=True)
    (decoders): MultiSequential(
      (0): DecoderLayer(
        (self_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (src_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): ReLU()
        )
        (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
      )
      (1): DecoderLayer(
        (self_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (src_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): ReLU()
        )
        (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
      )
      (2): DecoderLayer(
        (self_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (src_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): ReLU()
        )
        (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
      )
      (3): DecoderLayer(
        (self_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (src_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): ReLU()
        )
        (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
      )
      (4): DecoderLayer(
        (self_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (src_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): ReLU()
        )
        (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
      )
      (5): DecoderLayer(
        (self_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (src_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): ReLU()
        )
        (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
      )
      (6): DecoderLayer(
        (self_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (src_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): ReLU()
        )
        (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
      )
      (7): DecoderLayer(
        (self_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (src_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): ReLU()
        )
        (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
      )
      (8): DecoderLayer(
        (self_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (src_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): ReLU()
        )
        (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
      )
      (9): DecoderLayer(
        (self_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (src_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): ReLU()
        )
        (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
      )
      (10): DecoderLayer(
        (self_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (src_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): ReLU()
        )
        (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
      )
      (11): DecoderLayer(
        (self_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (src_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): ReLU()
        )
        (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
      )
      (12): DecoderLayer(
        (self_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (src_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): ReLU()
        )
        (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
      )
      (13): DecoderLayer(
        (self_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (src_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): ReLU()
        )
        (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
      )
      (14): DecoderLayer(
        (self_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (src_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): ReLU()
        )
        (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
      )
      (15): DecoderLayer(
        (self_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (src_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): ReLU()
        )
        (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
      )
      (16): DecoderLayer(
        (self_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (src_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): ReLU()
        )
        (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
      )
      (17): DecoderLayer(
        (self_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (src_attn): MultiHeadedAttention(
          (linear_q): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_k): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_v): Linear(in_features=1024, out_features=1024, bias=True)
          (linear_out): Linear(in_features=1024, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (q_norm): Identity()
          (k_norm): Identity()
        )
        (feed_forward): PositionwiseFeedForward(
          (w_1): Linear(in_features=1024, out_features=4096, bias=True)
          (w_2): Linear(in_features=4096, out_features=1024, bias=True)
          (dropout): Dropout(p=0.1, inplace=False)
          (activation): ReLU()
        )
        (norm1): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm2): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (norm3): LayerNorm((1024,), eps=1e-12, elementwise_affine=True)
        (dropout): Dropout(p=0.1, inplace=False)
      )
    )
  )
  (criterion_att): LabelSmoothingLoss(
    (criterion): KLDivLoss()
  )
  (ctc): CTC(
    (ctc_lo): Linear(in_features=1024, out_features=50002, bias=True)
    (ctc_loss): CTCLoss()
  )
)

Model summary:
    Class Name: ESPnetS2TModel
    Total Number of model parameters: 1.04 B
    Number of trainable parameters: 21.13 M (2.0%)
    Size: 84.52 MB
    Type: torch.float32
[w005] 2025-05-21 13:21:57,589 (abs_task:1402) INFO: Optimizer:
AdamW (
Parameter Group 0
    amsgrad: False
    betas: [0.9, 0.98]
    capturable: False
    differentiable: False
    eps: 1e-06
    foreach: None
    fused: None
    initial_lr: 0.0001
    lr: 2.5e-08
    maximize: False
    weight_decay: 0.0
)
[w005] 2025-05-21 13:21:57,589 (abs_task:1403) INFO: Scheduler: PiecewiseLinearWarmupLR(warmup_steps_list=[0, 10000, 50000], warmup_lr_list=[0.0, 0.00025, 0.0001])
[w005] 2025-05-21 13:21:57,685 (abs_task:1412) INFO: Saving the configuration in exp/s2t_lora_2/config.yaml
[w005] 2025-05-21 13:21:58,229 (abs_task:1477) INFO: Loading pretrained params from /ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/owsm_v3.1_ebf/valid.total_count.ave_5best.till45epoch.pth
[w005] 2025-05-21 13:22:24,187 (abs_task:1826) INFO: [train] Batch sampler: UnsortedBatchSampler(N-batch=22430, batch_size=8, key_file=/ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//train/speech_shape, 
[w005] 2025-05-21 13:22:24,229 (abs_task:1827) INFO: [train] mini-batch sizes summary: N-batch=22430, mean=8.0, min=8, max=8
[w005] 2025-05-21 13:22:24,398 (read_text:31) INFO: keys_to_load is not None, only loading 179440 keys from dump/raw/cgn_train/text.prev
[w005] 2025-05-21 13:22:24,563 (read_text:31) INFO: keys_to_load is not None, only loading 179440 keys from dump/raw/cgn_train/text.ctc
[w005] 2025-05-21 13:22:25,291 (read_text:31) INFO: keys_to_load is not None, only loading 179440 keys from dump/raw/cgn_train/text
[w005] 2025-05-21 13:22:25,503 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
[w005] 2025-05-21 13:22:25,506 (abs_task:1851) INFO: [train] dataset:
ESPnetDataset(
  speech: {"path": "dump/raw/cgn_train/wav.scp", "type": "kaldi_ark"}
  text_prev: {"path": "dump/raw/cgn_train/text.prev", "type": "text"}
  text_ctc: {"path": "dump/raw/cgn_train/text.ctc", "type": "text"}
  text: {"path": "dump/raw/cgn_train/text", "type": "text"}
  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x15298662e2f0>)
[w005] 2025-05-21 13:22:25,565 (abs_task:1826) INFO: [valid] Batch sampler: UnsortedBatchSampler(N-batch=3220, batch_size=8, key_file=/ocean/projects/cis210027p/pwang10/rota/espnet/egs2/owsm_v3.1/s2t1/exp/s2t_stats_raw_bpe50000_medium//valid/speech_shape, 
[w005] 2025-05-21 13:22:25,565 (abs_task:1827) INFO: [valid] mini-batch sizes summary: N-batch=3220, mean=8.0, min=8, max=9
[w005] 2025-05-21 13:22:25,647 (read_text:31) INFO: keys_to_load is not None, only loading 25766 keys from dump/raw/cgn_valid/text.prev
[w005] 2025-05-21 13:22:25,690 (read_text:31) INFO: keys_to_load is not None, only loading 25766 keys from dump/raw/cgn_valid/text.ctc
[w005] 2025-05-21 13:22:25,792 (read_text:31) INFO: keys_to_load is not None, only loading 25766 keys from dump/raw/cgn_valid/text
[w005] 2025-05-21 13:22:25,874 (s2t:444) INFO: Optional Data Names: ('text_prev', 'text_ctc', 'text_spk2', 'text_spk3', 'text_spk4')
[w005] 2025-05-21 13:22:25,874 (abs_task:1851) INFO: [valid] dataset:
ESPnetDataset(
  speech: {"path": "dump/raw/cgn_valid/wav.scp", "type": "kaldi_ark"}
  text_prev: {"path": "dump/raw/cgn_valid/text.prev", "type": "text"}
  text_ctc: {"path": "dump/raw/cgn_valid/text.ctc", "type": "text"}
  text: {"path": "dump/raw/cgn_valid/text", "type": "text"}
  preprocess: <espnet2.train.preprocessor.S2TPreprocessor object at 0x15294f8cbd90>)
[w005] 2025-05-21 13:23:01,717 (trainer:177) INFO: The training was resumed using exp/s2t_lora_2/checkpoint.pth
[w005] 2025-05-21 13:23:05,832 (trainer:336) INFO: 6/10epoch started
Failed to import Flash Attention, using ESPnet default: /ocean/projects/cis210027p/pwang10/miniconda3/envs/rota/lib/python3.10/site-packages/flash_attn_2_cuda.cpython-310-x86_64-linux-gnu.so: undefined symbol: _ZNK3c105Error4whatEv
[w005] 2025-05-21 13:28:24,806 (trainer:811) INFO: 6epoch:train:1-750batch: iter_time=0.003, forward_time=0.201, loss_ctc=56.478, loss_att=29.771, acc=0.776, loss=9.446, backward_time=0.185, grad_norm=28.783, clip=100.000, loss_scale=3.355e+07, optim_step_time=0.028, optim0_lr0=2.168e-04, train_time=1.701
[w005] 2025-05-21 13:33:18,163 (trainer:811) INFO: 6epoch:train:751-1500batch: iter_time=1.267e-04, forward_time=0.176, loss_ctc=57.259, loss_att=30.128, acc=0.777, loss=9.567, backward_time=0.182, grad_norm=32.832, clip=100.000, loss_scale=3.355e+07, optim_step_time=0.024, optim0_lr0=2.161e-04, train_time=1.564
[w005] 2025-05-21 13:38:11,451 (trainer:811) INFO: 6epoch:train:1501-2250batch: iter_time=1.249e-04, forward_time=0.175, loss_ctc=55.391, loss_att=29.201, acc=0.778, loss=9.265, backward_time=0.182, grad_norm=29.381, clip=100.000, loss_scale=3.355e+07, optim_step_time=0.024, optim0_lr0=2.154e-04, train_time=1.564
[w005] 2025-05-21 13:43:04,807 (trainer:811) INFO: 6epoch:train:2251-3000batch: iter_time=1.245e-04, forward_time=0.175, loss_ctc=56.437, loss_att=29.753, acc=0.775, loss=9.439, backward_time=0.182, grad_norm=31.385, clip=100.000, loss_scale=3.355e+07, optim_step_time=0.025, optim0_lr0=2.147e-04, train_time=1.564
[w005] 2025-05-21 13:47:58,165 (trainer:811) INFO: 6epoch:train:3001-3750batch: iter_time=1.264e-04, forward_time=0.176, loss_ctc=55.530, loss_att=29.328, acc=0.774, loss=9.297, backward_time=0.182, grad_norm=29.921, clip=100.000, loss_scale=3.355e+07, optim_step_time=0.024, optim0_lr0=2.140e-04, train_time=1.565
[w005] 2025-05-21 13:52:51,608 (trainer:811) INFO: 6epoch:train:3751-4500batch: iter_time=1.270e-04, forward_time=0.175, loss_ctc=55.974, loss_att=29.388, acc=0.775, loss=9.341, backward_time=0.182, grad_norm=29.940, clip=100.000, loss_scale=3.355e+07, optim_step_time=0.024, optim0_lr0=2.133e-04, train_time=1.565
[w005] 2025-05-21 13:57:44,222 (trainer:811) INFO: 6epoch:train:4501-5250batch: iter_time=1.243e-04, forward_time=0.175, loss_ctc=54.748, loss_att=28.933, acc=0.778, loss=9.169, backward_time=0.181, grad_norm=29.150, clip=100.000, loss_scale=4.468e+07, optim_step_time=0.024, optim0_lr0=2.126e-04, train_time=1.561
[w005] 2025-05-21 14:02:37,439 (trainer:811) INFO: 6epoch:train:5251-6000batch: iter_time=1.200e-04, forward_time=0.175, loss_ctc=56.012, loss_att=29.459, acc=0.779, loss=9.356, backward_time=0.182, grad_norm=29.885, clip=100.000, loss_scale=6.711e+07, optim_step_time=0.024, optim0_lr0=2.119e-04, train_time=1.563
[w005] 2025-05-21 14:07:30,595 (trainer:811) INFO: 6epoch:train:6001-6750batch: iter_time=1.206e-04, forward_time=0.175, loss_ctc=54.969, loss_att=28.835, acc=0.781, loss=9.169, backward_time=0.182, grad_norm=30.643, clip=100.000, loss_scale=6.711e+07, optim_step_time=0.024, optim0_lr0=2.112e-04, train_time=1.564
[w005] 2025-05-21 14:12:24,045 (trainer:811) INFO: 6epoch:train:6751-7500batch: iter_time=1.184e-04, forward_time=0.175, loss_ctc=54.135, loss_att=28.461, acc=0.779, loss=9.041, backward_time=0.182, grad_norm=30.216, clip=100.000, loss_scale=6.711e+07, optim_step_time=0.025, optim0_lr0=2.105e-04, train_time=1.565
[w005] 2025-05-21 14:17:17,135 (trainer:811) INFO: 6epoch:train:7501-8250batch: iter_time=1.217e-04, forward_time=0.175, loss_ctc=55.314, loss_att=29.021, acc=0.777, loss=9.227, backward_time=0.182, grad_norm=30.483, clip=100.000, loss_scale=6.711e+07, optim_step_time=0.024, optim0_lr0=2.098e-04, train_time=1.563
[w005] 2025-05-21 14:22:10,572 (trainer:811) INFO: 6epoch:train:8251-9000batch: iter_time=1.211e-04, forward_time=0.176, loss_ctc=56.512, loss_att=29.732, acc=0.776, loss=9.441, backward_time=0.182, grad_norm=30.363, clip=100.000, loss_scale=6.711e+07, optim_step_time=0.024, optim0_lr0=2.091e-04, train_time=1.565
[w005] 2025-05-21 14:27:03,895 (trainer:811) INFO: 6epoch:train:9001-9750batch: iter_time=1.214e-04, forward_time=0.175, loss_ctc=56.422, loss_att=29.651, acc=0.781, loss=9.421, backward_time=0.182, grad_norm=31.064, clip=100.000, loss_scale=6.711e+07, optim_step_time=0.024, optim0_lr0=2.084e-04, train_time=1.564
[w005] 2025-05-21 14:31:57,397 (trainer:811) INFO: 6epoch:train:9751-10500batch: iter_time=1.217e-04, forward_time=0.175, loss_ctc=55.699, loss_att=29.267, acc=0.779, loss=9.299, backward_time=0.182, grad_norm=29.423, clip=100.000, loss_scale=6.711e+07, optim_step_time=0.025, optim0_lr0=2.077e-04, train_time=1.565
[w005] 2025-05-21 14:36:50,447 (trainer:811) INFO: 6epoch:train:10501-11250batch: iter_time=1.214e-04, forward_time=0.175, loss_ctc=56.259, loss_att=29.681, acc=0.777, loss=9.414, backward_time=0.182, grad_norm=31.992, clip=100.000, loss_scale=6.711e+07, optim_step_time=0.024, optim0_lr0=2.070e-04, train_time=1.563
[w005] 2025-05-21 14:41:43,779 (trainer:811) INFO: 6epoch:train:11251-12000batch: iter_time=1.206e-04, forward_time=0.175, loss_ctc=55.388, loss_att=29.155, acc=0.779, loss=9.256, backward_time=0.182, grad_norm=31.551, clip=100.000, loss_scale=6.711e+07, optim_step_time=0.024, optim0_lr0=2.063e-04, train_time=1.564
[w005] 2025-05-21 14:46:37,171 (trainer:811) INFO: 6epoch:train:12001-12750batch: iter_time=1.200e-04, forward_time=0.175, loss_ctc=57.326, loss_att=30.143, acc=0.778, loss=9.575, backward_time=0.182, grad_norm=31.548, clip=100.000, loss_scale=6.711e+07, optim_step_time=0.025, optim0_lr0=2.056e-04, train_time=1.565
[w005] 2025-05-21 14:51:30,178 (trainer:811) INFO: 6epoch:train:12751-13500batch: iter_time=1.212e-04, forward_time=0.175, loss_ctc=54.330, loss_att=28.618, acc=0.781, loss=9.083, backward_time=0.181, grad_norm=31.851, clip=100.000, loss_scale=1.117e+08, optim_step_time=0.024, optim0_lr0=2.049e-04, train_time=1.562
[w005] 2025-05-21 14:56:23,204 (trainer:811) INFO: 6epoch:train:13501-14250batch: iter_time=1.210e-04, forward_time=0.175, loss_ctc=54.946, loss_att=28.860, acc=0.782, loss=9.171, backward_time=0.182, grad_norm=30.616, clip=100.000, loss_scale=1.342e+08, optim_step_time=0.024, optim0_lr0=2.042e-04, train_time=1.563
[w005] 2025-05-21 15:01:16,602 (trainer:811) INFO: 6epoch:train:14251-15000batch: iter_time=1.190e-04, forward_time=0.176, loss_ctc=54.636, loss_att=28.618, acc=0.779, loss=9.106, backward_time=0.182, grad_norm=30.345, clip=100.000, loss_scale=1.342e+08, optim_step_time=0.024, optim0_lr0=2.035e-04, train_time=1.564
[w005] 2025-05-21 15:08:50,719 (trainer:394) INFO: 6epoch results: [train] iter_time=2.418e-04, forward_time=0.177, loss_ctc=55.688, loss_att=29.300, acc=0.778, loss=9.304, backward_time=0.182, grad_norm=30.569, clip=100.000, loss_scale=6.487e+07, optim_step_time=0.024, optim0_lr0=2.102e-04, train_time=1.571, time=1 hour, 38 minutes and 10.95 seconds, total_count=90000, gpu_max_cached_mem_GB=43.801, [valid] loss_ctc=21.442, cer_ctc=0.115, loss_att=10.582, acc=0.888, cer=0.178, wer=0.989, loss=13.840, time=7 minutes and 33.84 seconds, total_count=19320, gpu_max_cached_mem_GB=43.801
[w005] 2025-05-21 15:09:07,827 (trainer:462) INFO: The best model has been updated: valid.total_count
[w005] 2025-05-21 15:09:08,051 (trainer:516) INFO: The model files were removed: exp/s2t_lora_2/5epoch.pth
[w005] 2025-05-21 15:09:08,051 (trainer:324) INFO: 7/10epoch started. Estimated time to finish: 7 hours, 4 minutes and 8.88 seconds
[w005] 2025-05-21 15:14:02,120 (trainer:811) INFO: 7epoch:train:1-750batch: iter_time=3.117e-04, forward_time=0.177, loss_ctc=54.638, loss_att=28.480, acc=0.785, loss=9.082, backward_time=0.180, grad_norm=28.446, clip=100.000, loss_scale=1.342e+08, optim_step_time=0.025, optim0_lr0=2.028e-04, train_time=1.568
[w005] 2025-05-21 15:18:54,092 (trainer:811) INFO: 7epoch:train:751-1500batch: iter_time=1.277e-04, forward_time=0.174, loss_ctc=53.915, loss_att=28.094, acc=0.786, loss=8.960, backward_time=0.181, grad_norm=30.299, clip=100.000, loss_scale=1.342e+08, optim_step_time=0.024, optim0_lr0=2.021e-04, train_time=1.557
[w005] 2025-05-21 15:23:46,878 (trainer:811) INFO: 7epoch:train:1501-2250batch: iter_time=1.244e-04, forward_time=0.175, loss_ctc=53.989, loss_att=28.226, acc=0.787, loss=8.989, backward_time=0.181, grad_norm=29.151, clip=100.000, loss_scale=1.342e+08, optim_step_time=0.024, optim0_lr0=2.014e-04, train_time=1.562
[w005] 2025-05-21 15:28:42,876 (trainer:811) INFO: 7epoch:train:2251-3000batch: iter_time=1.237e-04, forward_time=0.179, loss_ctc=55.079, loss_att=28.824, acc=0.786, loss=9.175, backward_time=0.182, grad_norm=30.881, clip=100.000, loss_scale=1.342e+08, optim_step_time=0.024, optim0_lr0=2.007e-04, train_time=1.578
[w005] 2025-05-21 15:33:38,353 (trainer:811) INFO: 7epoch:train:3001-3750batch: iter_time=1.272e-04, forward_time=0.176, loss_ctc=54.169, loss_att=28.307, acc=0.784, loss=9.016, backward_time=0.181, grad_norm=29.771, clip=100.000, loss_scale=1.342e+08, optim_step_time=0.024, optim0_lr0=2.000e-04, train_time=1.576
[w005] 2025-05-21 15:38:31,507 (trainer:811) INFO: 7epoch:train:3751-4500batch: iter_time=1.261e-04, forward_time=0.176, loss_ctc=54.028, loss_att=28.229, acc=0.784, loss=8.992, backward_time=0.181, grad_norm=30.919, clip=100.000, loss_scale=1.342e+08, optim_step_time=0.024, optim0_lr0=1.993e-04, train_time=1.563
[w005] 2025-05-21 15:43:25,039 (trainer:811) INFO: 7epoch:train:4501-5250batch: iter_time=1.237e-04, forward_time=0.176, loss_ctc=54.127, loss_att=28.259, acc=0.785, loss=9.005, backward_time=0.181, grad_norm=29.069, clip=100.000, loss_scale=1.342e+08, optim_step_time=0.025, optim0_lr0=1.986e-04, train_time=1.566
[w005] 2025-05-21 15:48:15,240 (trainer:811) INFO: 7epoch:train:5251-6000batch: iter_time=1.113e-04, forward_time=0.173, loss_ctc=54.765, loss_att=28.528, acc=0.785, loss=9.100, backward_time=0.179, grad_norm=32.017, clip=100.000, loss_scale=1.342e+08, optim_step_time=0.023, optim0_lr0=1.978e-04, train_time=1.547
[w005] 2025-05-21 15:53:06,663 (trainer:811) INFO: 7epoch:train:6001-6750batch: iter_time=1.148e-04, forward_time=0.173, loss_ctc=53.101, loss_att=27.703, acc=0.785, loss=8.831, backward_time=0.180, grad_norm=31.990, clip=100.000, loss_scale=2.684e+08, optim_step_time=0.024, optim0_lr0=1.971e-04, train_time=1.554
[w005] 2025-05-21 15:57:56,704 (trainer:811) INFO: 7epoch:train:6751-7500batch: iter_time=1.104e-04, forward_time=0.173, loss_ctc=54.661, loss_att=28.539, acc=0.782, loss=9.094, backward_time=0.179, grad_norm=31.199, clip=100.000, loss_scale=2.684e+08, optim_step_time=0.023, optim0_lr0=1.964e-04, train_time=1.547
[w005] 2025-05-21 16:02:46,867 (trainer:811) INFO: 7epoch:train:7501-8250batch: iter_time=1.126e-04, forward_time=0.173, loss_ctc=55.079, loss_att=28.840, acc=0.782, loss=9.178, backward_time=0.179, grad_norm=30.334, clip=100.000, loss_scale=2.684e+08, optim_step_time=0.024, optim0_lr0=1.957e-04, train_time=1.548
[w005] 2025-05-21 16:07:37,059 (trainer:811) INFO: 7epoch:train:8251-9000batch: iter_time=1.116e-04, forward_time=0.173, loss_ctc=54.034, loss_att=28.187, acc=0.783, loss=8.985, backward_time=0.179, grad_norm=30.655, clip=100.000, loss_scale=2.684e+08, optim_step_time=0.023, optim0_lr0=1.950e-04, train_time=1.547
[w005] 2025-05-21 16:12:26,816 (trainer:811) INFO: 7epoch:train:9001-9750batch: iter_time=1.121e-04, forward_time=0.173, loss_ctc=54.159, loss_att=28.267, acc=0.785, loss=9.009, backward_time=0.179, grad_norm=33.635, clip=100.000, loss_scale=2.684e+08, optim_step_time=0.023, optim0_lr0=1.943e-04, train_time=1.545
[w005] 2025-05-21 16:17:24,842 (trainer:811) INFO: 7epoch:train:9751-10500batch: iter_time=2.516e-04, forward_time=0.179, loss_ctc=55.784, loss_att=29.155, acc=0.779, loss=9.286, backward_time=0.180, grad_norm=32.023, clip=100.000, loss_scale=2.684e+08, optim_step_time=0.026, optim0_lr0=1.936e-04, train_time=1.586
[w005] 2025-05-21 16:22:17,665 (trainer:811) INFO: 7epoch:train:10501-11250batch: iter_time=1.104e-04, forward_time=0.173, loss_ctc=55.473, loss_att=28.921, acc=0.781, loss=9.222, backward_time=0.179, grad_norm=31.483, clip=100.000, loss_scale=2.684e+08, optim_step_time=0.037, optim0_lr0=1.929e-04, train_time=1.565
[w005] 2025-05-21 16:27:07,531 (trainer:811) INFO: 7epoch:train:11251-12000batch: iter_time=1.094e-04, forward_time=0.173, loss_ctc=52.951, loss_att=27.668, acc=0.789, loss=8.813, backward_time=0.179, grad_norm=31.249, clip=100.000, loss_scale=2.684e+08, optim_step_time=0.023, optim0_lr0=1.922e-04, train_time=1.546
[w005] 2025-05-21 16:31:56,851 (trainer:811) INFO: 7epoch:train:12001-12750batch: iter_time=1.099e-04, forward_time=0.172, loss_ctc=54.566, loss_att=28.502, acc=0.783, loss=9.080, backward_time=0.179, grad_norm=30.397, clip=100.000, loss_scale=2.684e+08, optim_step_time=0.023, optim0_lr0=1.915e-04, train_time=1.543
[w005] 2025-05-21 16:36:48,118 (trainer:811) INFO: 7epoch:train:12751-13500batch: iter_time=1.079e-04, forward_time=0.174, loss_ctc=55.162, loss_att=28.833, acc=0.784, loss=9.183, backward_time=0.179, grad_norm=31.791, clip=100.000, loss_scale=2.684e+08, optim_step_time=0.023, optim0_lr0=1.908e-04, train_time=1.553
[w005] 2025-05-21 16:41:37,905 (trainer:811) INFO: 7epoch:train:13501-14250batch: iter_time=1.116e-04, forward_time=0.172, loss_ctc=53.395, loss_att=27.696, acc=0.787, loss=8.851, backward_time=0.179, grad_norm=31.619, clip=100.000, loss_scale=3.574e+08, optim_step_time=0.023, optim0_lr0=1.901e-04, train_time=1.546
[w005] 2025-05-21 16:46:26,990 (trainer:811) INFO: 7epoch:train:14251-15000batch: iter_time=1.112e-04, forward_time=0.172, loss_ctc=54.921, loss_att=28.594, acc=0.785, loss=9.123, backward_time=0.179, grad_norm=30.190, clip=100.000, loss_scale=5.369e+08, optim_step_time=0.023, optim0_lr0=1.894e-04, train_time=1.541
[w005] 2025-05-21 16:54:17,528 (trainer:394) INFO: 7epoch results: [train] iter_time=1.325e-04, forward_time=0.174, loss_ctc=54.400, loss_att=28.393, acc=0.784, loss=9.049, backward_time=0.180, grad_norm=30.857, clip=100.000, loss_scale=2.326e+08, optim_step_time=0.025, optim0_lr0=1.961e-04, train_time=1.557, time=1 hour, 37 minutes and 19.34 seconds, total_count=105000, gpu_max_cached_mem_GB=43.801, [valid] loss_ctc=20.808, cer_ctc=0.111, loss_att=10.132, acc=0.921, cer=0.056, wer=0.776, loss=13.335, time=7 minutes and 49.51 seconds, total_count=22540, gpu_max_cached_mem_GB=43.801
[w005] 2025-05-21 16:57:30,970 (trainer:462) INFO: The best model has been updated: valid.acc, valid.total_count
[w005] 2025-05-21 16:57:30,972 (trainer:516) INFO: The model files were removed: exp/s2t_lora_2/6epoch.pth
[w005] 2025-05-21 16:57:30,973 (trainer:324) INFO: 8/10epoch started. Estimated time to finish: 5 hours, 21 minutes and 37.71 seconds
[w005] 2025-05-21 17:02:22,861 (trainer:811) INFO: 8epoch:train:1-750batch: iter_time=5.339e-04, forward_time=0.174, loss_ctc=54.083, loss_att=28.246, acc=0.786, loss=8.999, backward_time=0.179, grad_norm=30.428, clip=100.000, loss_scale=5.369e+08, optim_step_time=0.024, optim0_lr0=1.887e-04, train_time=1.556
[w005] 2025-05-21 17:07:14,500 (trainer:811) INFO: 8epoch:train:751-1500batch: iter_time=1.190e-04, forward_time=0.174, loss_ctc=55.402, loss_att=28.736, acc=0.783, loss=9.184, backward_time=0.180, grad_norm=29.876, clip=100.000, loss_scale=5.369e+08, optim_step_time=0.024, optim0_lr0=1.880e-04, train_time=1.555
[w005] 2025-05-21 17:12:04,989 (trainer:811) INFO: 8epoch:train:1501-2250batch: iter_time=1.103e-04, forward_time=0.173, loss_ctc=55.817, loss_att=29.158, acc=0.781, loss=9.289, backward_time=0.179, grad_norm=30.974, clip=100.000, loss_scale=5.369e+08, optim_step_time=0.024, optim0_lr0=1.873e-04, train_time=1.549
[w005] 2025-05-21 17:16:55,498 (trainer:811) INFO: 8epoch:train:2251-3000batch: iter_time=1.141e-04, forward_time=0.174, loss_ctc=55.176, loss_att=28.867, acc=0.781, loss=9.190, backward_time=0.179, grad_norm=32.699, clip=100.000, loss_scale=5.369e+08, optim_step_time=0.024, optim0_lr0=1.866e-04, train_time=1.549
[w005] 2025-05-21 17:21:47,385 (trainer:811) INFO: 8epoch:train:3001-3750batch: iter_time=1.131e-04, forward_time=0.175, loss_ctc=54.785, loss_att=28.857, acc=0.782, loss=9.159, backward_time=0.179, grad_norm=31.998, clip=100.000, loss_scale=5.369e+08, optim_step_time=0.024, optim0_lr0=1.859e-04, train_time=1.557
[w005] 2025-05-21 17:26:44,256 (trainer:811) INFO: 8epoch:train:3751-4500batch: iter_time=1.143e-04, forward_time=0.181, loss_ctc=55.233, loss_att=28.646, acc=0.786, loss=9.156, backward_time=0.180, grad_norm=32.053, clip=100.000, loss_scale=5.369e+08, optim_step_time=0.023, optim0_lr0=1.852e-04, train_time=1.583
[w005] 2025-05-21 17:31:34,664 (trainer:811) INFO: 8epoch:train:4501-5250batch: iter_time=1.076e-04, forward_time=0.173, loss_ctc=53.759, loss_att=28.061, acc=0.784, loss=8.943, backward_time=0.179, grad_norm=31.365, clip=100.000, loss_scale=5.369e+08, optim_step_time=0.024, optim0_lr0=1.845e-04, train_time=1.549
[w005] 2025-05-21 17:36:24,795 (trainer:811) INFO: 8epoch:train:5251-6000batch: iter_time=1.099e-04, forward_time=0.173, loss_ctc=55.239, loss_att=28.886, acc=0.782, loss=9.198, backward_time=0.179, grad_norm=30.046, clip=100.000, loss_scale=5.369e+08, optim_step_time=0.024, optim0_lr0=1.838e-04, train_time=1.547
[w005] 2025-05-21 17:41:16,357 (trainer:811) INFO: 8epoch:train:6001-6750batch: iter_time=1.244e-04, forward_time=0.174, loss_ctc=53.499, loss_att=27.830, acc=0.787, loss=8.883, backward_time=0.180, grad_norm=31.506, clip=100.000, loss_scale=5.369e+08, optim_step_time=0.024, optim0_lr0=1.831e-04, train_time=1.555
[w005] 2025-05-21 17:46:12,056 (trainer:811) INFO: 8epoch:train:6751-7500batch: iter_time=1.255e-04, forward_time=0.178, loss_ctc=54.179, loss_att=28.081, acc=0.787, loss=8.978, backward_time=0.181, grad_norm=32.434, clip=100.000, loss_scale=8.938e+08, optim_step_time=0.025, optim0_lr0=1.824e-04, train_time=1.577
[w005] 2025-05-21 17:51:04,698 (trainer:811) INFO: 8epoch:train:7501-8250batch: iter_time=1.244e-04, forward_time=0.175, loss_ctc=54.230, loss_att=27.950, acc=0.787, loss=8.959, backward_time=0.181, grad_norm=32.171, clip=100.000, loss_scale=1.074e+09, optim_step_time=0.025, optim0_lr0=1.817e-04, train_time=1.561
[w005] 2025-05-21 17:55:57,359 (trainer:811) INFO: 8epoch:train:8251-9000batch: iter_time=1.270e-04, forward_time=0.175, loss_ctc=53.908, loss_att=28.056, acc=0.788, loss=8.953, backward_time=0.181, grad_norm=32.979, clip=100.000, loss_scale=1.074e+09, optim_step_time=0.024, optim0_lr0=1.810e-04, train_time=1.561
[w005] 2025-05-21 18:00:50,101 (trainer:811) INFO: 8epoch:train:9001-9750batch: iter_time=1.272e-04, forward_time=0.175, loss_ctc=52.056, loss_att=26.863, acc=0.791, loss=8.605, backward_time=0.181, grad_norm=33.344, clip=100.000, loss_scale=1.074e+09, optim_step_time=0.024, optim0_lr0=1.803e-04, train_time=1.561
[w005] 2025-05-21 18:05:50,534 (trainer:811) INFO: 8epoch:train:9751-10500batch: iter_time=4.240e-04, forward_time=0.181, loss_ctc=52.738, loss_att=27.135, acc=0.790, loss=8.704, backward_time=0.185, grad_norm=31.895, clip=100.000, loss_scale=1.074e+09, optim_step_time=0.024, optim0_lr0=1.796e-04, train_time=1.602
[w005] 2025-05-21 18:11:03,351 (trainer:811) INFO: 8epoch:train:10501-11250batch: iter_time=1.244e-04, forward_time=0.202, loss_ctc=53.343, loss_att=27.639, acc=0.789, loss=8.838, backward_time=0.181, grad_norm=32.181, clip=100.000, loss_scale=1.074e+09, optim_step_time=0.025, optim0_lr0=1.789e-04, train_time=1.669
[w005] 2025-05-21 18:15:55,853 (trainer:811) INFO: 8epoch:train:11251-12000batch: iter_time=1.245e-04, forward_time=0.175, loss_ctc=54.209, loss_att=27.830, acc=0.788, loss=8.936, backward_time=0.181, grad_norm=32.985, clip=100.000, loss_scale=1.074e+09, optim_step_time=0.024, optim0_lr0=1.782e-04, train_time=1.560
[w005] 2025-05-21 18:20:49,403 (trainer:811) INFO: 8epoch:train:12001-12750batch: iter_time=1.281e-04, forward_time=0.175, loss_ctc=53.817, loss_att=27.803, acc=0.787, loss=8.902, backward_time=0.181, grad_norm=32.269, clip=100.000, loss_scale=1.074e+09, optim_step_time=0.024, optim0_lr0=1.775e-04, train_time=1.566
[w005] 2025-05-21 18:25:42,557 (trainer:811) INFO: 8epoch:train:12751-13500batch: iter_time=1.253e-04, forward_time=0.175, loss_ctc=53.837, loss_att=27.792, acc=0.788, loss=8.901, backward_time=0.181, grad_norm=32.277, clip=100.000, loss_scale=1.074e+09, optim_step_time=0.024, optim0_lr0=1.768e-04, train_time=1.563
[w005] 2025-05-21 18:30:35,815 (trainer:811) INFO: 8epoch:train:13501-14250batch: iter_time=1.252e-04, forward_time=0.175, loss_ctc=53.449, loss_att=27.860, acc=0.784, loss=8.884, backward_time=0.181, grad_norm=32.556, clip=100.000, loss_scale=1.074e+09, optim_step_time=0.024, optim0_lr0=1.760e-04, train_time=1.564
[w005] 2025-05-21 18:35:28,827 (trainer:811) INFO: 8epoch:train:14251-15000batch: iter_time=1.264e-04, forward_time=0.175, loss_ctc=54.107, loss_att=27.956, acc=0.787, loss=8.950, backward_time=0.181, grad_norm=33.126, clip=100.000, loss_scale=1.074e+09, optim_step_time=0.024, optim0_lr0=1.753e-04, train_time=1.562
[w005] 2025-05-21 18:43:09,399 (trainer:394) INFO: 8epoch results: [train] iter_time=1.564e-04, forward_time=0.177, loss_ctc=54.143, loss_att=28.113, acc=0.786, loss=8.980, backward_time=0.180, grad_norm=31.958, clip=100.000, loss_scale=8.232e+08, optim_step_time=0.024, optim0_lr0=1.820e-04, train_time=1.567, time=1 hour, 37 minutes and 58.09 seconds, total_count=120000, gpu_max_cached_mem_GB=43.801, [valid] loss_ctc=20.377, cer_ctc=0.109, loss_att=10.059, acc=0.923, cer=0.055, wer=0.769, loss=13.154, time=7 minutes and 40.29 seconds, total_count=25760, gpu_max_cached_mem_GB=43.801
[w005] 2025-05-21 18:43:45,174 (trainer:462) INFO: The best model has been updated: valid.acc, valid.total_count
[w005] 2025-05-21 18:43:45,176 (trainer:516) INFO: The model files were removed: exp/s2t_lora_2/7epoch.pth
[w005] 2025-05-21 18:43:45,177 (trainer:324) INFO: 9/10epoch started. Estimated time to finish: 3 hours, 33 minutes and 46.23 seconds
[w005] 2025-05-21 18:48:38,628 (trainer:811) INFO: 9epoch:train:1-750batch: iter_time=3.271e-04, forward_time=0.175, loss_ctc=54.227, loss_att=27.969, acc=0.790, loss=8.962, backward_time=0.181, grad_norm=33.768, clip=100.000, loss_scale=2.147e+09, optim_step_time=0.025, optim0_lr0=1.746e-04, train_time=1.564
[w005] 2025-05-21 18:53:31,992 (trainer:811) INFO: 9epoch:train:751-1500batch: iter_time=1.280e-04, forward_time=0.175, loss_ctc=53.542, loss_att=27.823, acc=0.792, loss=8.885, backward_time=0.182, grad_norm=31.458, clip=100.000, loss_scale=2.147e+09, optim_step_time=0.024, optim0_lr0=1.739e-04, train_time=1.564
[w005] 2025-05-21 18:58:25,698 (trainer:811) INFO: 9epoch:train:1501-2250batch: iter_time=1.289e-04, forward_time=0.176, loss_ctc=53.973, loss_att=27.807, acc=0.791, loss=8.914, backward_time=0.182, grad_norm=33.465, clip=100.000, loss_scale=2.147e+09, optim_step_time=0.024, optim0_lr0=1.732e-04, train_time=1.567
[w005] 2025-05-21 19:03:19,204 (trainer:811) INFO: 9epoch:train:2251-3000batch: iter_time=1.271e-04, forward_time=0.176, loss_ctc=54.266, loss_att=28.094, acc=0.790, loss=8.986, backward_time=0.182, grad_norm=32.215, clip=100.000, loss_scale=2.147e+09, optim_step_time=0.024, optim0_lr0=1.725e-04, train_time=1.565
[w005] 2025-05-21 19:08:11,584 (trainer:811) INFO: 9epoch:train:3001-3750batch: iter_time=1.300e-04, forward_time=0.175, loss_ctc=53.126, loss_att=27.366, acc=0.791, loss=8.773, backward_time=0.181, grad_norm=33.137, clip=100.000, loss_scale=2.147e+09, optim_step_time=0.024, optim0_lr0=1.718e-04, train_time=1.559
[w005] 2025-05-21 19:13:03,976 (trainer:811) INFO: 9epoch:train:3751-4500batch: iter_time=1.229e-04, forward_time=0.175, loss_ctc=52.961, loss_att=27.485, acc=0.790, loss=8.782, backward_time=0.181, grad_norm=33.231, clip=100.000, loss_scale=2.147e+09, optim_step_time=0.024, optim0_lr0=1.711e-04, train_time=1.559
[w005] 2025-05-21 19:17:56,915 (trainer:811) INFO: 9epoch:train:4501-5250batch: iter_time=1.250e-04, forward_time=0.175, loss_ctc=53.422, loss_att=27.793, acc=0.787, loss=8.870, backward_time=0.181, grad_norm=33.432, clip=100.000, loss_scale=2.147e+09, optim_step_time=0.025, optim0_lr0=1.704e-04, train_time=1.562
[w005] 2025-05-21 19:22:49,825 (trainer:811) INFO: 9epoch:train:5251-6000batch: iter_time=1.304e-04, forward_time=0.175, loss_ctc=53.726, loss_att=27.866, acc=0.790, loss=8.906, backward_time=0.181, grad_norm=33.640, clip=100.000, loss_scale=2.147e+09, optim_step_time=0.024, optim0_lr0=1.697e-04, train_time=1.562
[w005] 2025-05-21 19:27:42,728 (trainer:811) INFO: 9epoch:train:6001-6750batch: iter_time=1.263e-04, forward_time=0.175, loss_ctc=52.740, loss_att=27.239, acc=0.791, loss=8.722, backward_time=0.181, grad_norm=33.022, clip=100.000, loss_scale=2.147e+09, optim_step_time=0.024, optim0_lr0=1.690e-04, train_time=1.562
[w005] 2025-05-21 19:32:35,480 (trainer:811) INFO: 9epoch:train:6751-7500batch: iter_time=1.302e-04, forward_time=0.175, loss_ctc=52.574, loss_att=27.273, acc=0.791, loss=8.716, backward_time=0.181, grad_norm=32.457, clip=100.000, loss_scale=2.147e+09, optim_step_time=0.024, optim0_lr0=1.683e-04, train_time=1.561
[w005] 2025-05-21 19:37:29,004 (trainer:811) INFO: 9epoch:train:7501-8250batch: iter_time=1.264e-04, forward_time=0.175, loss_ctc=54.070, loss_att=27.891, acc=0.789, loss=8.936, backward_time=0.182, grad_norm=34.265, clip=100.000, loss_scale=2.859e+09, optim_step_time=0.025, optim0_lr0=1.676e-04, train_time=1.566
[w005] 2025-05-21 19:42:27,941 (trainer:811) INFO: 9epoch:train:8251-9000batch: iter_time=1.282e-04, forward_time=0.177, loss_ctc=52.314, loss_att=27.130, acc=0.793, loss=8.671, backward_time=0.187, grad_norm=33.111, clip=100.000, loss_scale=4.295e+09, optim_step_time=0.027, optim0_lr0=1.669e-04, train_time=1.594
[w005] 2025-05-21 19:47:24,966 (trainer:811) INFO: 9epoch:train:9001-9750batch: iter_time=1.247e-04, forward_time=0.178, loss_ctc=53.312, loss_att=27.467, acc=0.793, loss=8.805, backward_time=0.181, grad_norm=32.738, clip=100.000, loss_scale=4.295e+09, optim_step_time=0.026, optim0_lr0=1.662e-04, train_time=1.583
[w005] 2025-05-21 19:52:23,411 (trainer:811) INFO: 9epoch:train:9751-10500batch: iter_time=1.605e-04, forward_time=0.181, loss_ctc=54.136, loss_att=27.740, acc=0.789, loss=8.915, backward_time=0.182, grad_norm=35.418, clip=100.000, loss_scale=4.295e+09, optim_step_time=0.025, optim0_lr0=1.655e-04, train_time=1.593
[w005] 2025-05-21 19:57:16,486 (trainer:811) INFO: 9epoch:train:10501-11250batch: iter_time=1.266e-04, forward_time=0.175, loss_ctc=54.176, loss_att=27.949, acc=0.789, loss=8.954, backward_time=0.181, grad_norm=33.783, clip=100.000, loss_scale=4.295e+09, optim_step_time=0.024, optim0_lr0=1.648e-04, train_time=1.563
[w005] 2025-05-21 20:02:09,441 (trainer:811) INFO: 9epoch:train:11251-12000batch: iter_time=1.272e-04, forward_time=0.175, loss_ctc=53.334, loss_att=27.406, acc=0.790, loss=8.796, backward_time=0.181, grad_norm=33.630, clip=100.000, loss_scale=4.295e+09, optim_step_time=0.024, optim0_lr0=1.641e-04, train_time=1.562
[w005] 2025-05-21 20:07:02,489 (trainer:811) INFO: 9epoch:train:12001-12750batch: iter_time=1.267e-04, forward_time=0.175, loss_ctc=54.468, loss_att=28.121, acc=0.792, loss=9.006, backward_time=0.181, grad_norm=32.459, clip=100.000, loss_scale=4.295e+09, optim_step_time=0.024, optim0_lr0=1.634e-04, train_time=1.563
[w005] 2025-05-21 20:11:55,755 (trainer:811) INFO: 9epoch:train:12751-13500batch: iter_time=1.256e-04, forward_time=0.175, loss_ctc=53.592, loss_att=27.754, acc=0.788, loss=8.876, backward_time=0.181, grad_norm=32.470, clip=100.000, loss_scale=4.295e+09, optim_step_time=0.024, optim0_lr0=1.627e-04, train_time=1.564
[w005] 2025-05-21 20:16:48,807 (trainer:811) INFO: 9epoch:train:13501-14250batch: iter_time=1.257e-04, forward_time=0.175, loss_ctc=53.086, loss_att=27.646, acc=0.788, loss=8.820, backward_time=0.181, grad_norm=33.559, clip=100.000, loss_scale=4.295e+09, optim_step_time=0.024, optim0_lr0=1.620e-04, train_time=1.563
[w005] 2025-05-21 20:21:41,965 (trainer:811) INFO: 9epoch:train:14251-15000batch: iter_time=1.275e-04, forward_time=0.175, loss_ctc=52.628, loss_att=27.164, acc=0.791, loss=8.701, backward_time=0.182, grad_norm=34.630, clip=100.000, loss_scale=4.295e+09, optim_step_time=0.024, optim0_lr0=1.613e-04, train_time=1.563
[w005] 2025-05-21 20:29:17,817 (trainer:394) INFO: 9epoch results: [train] iter_time=1.387e-04, forward_time=0.176, loss_ctc=53.484, loss_att=27.649, acc=0.790, loss=8.850, backward_time=0.182, grad_norm=33.294, clip=100.000, loss_scale=3.150e+09, optim_step_time=0.024, optim0_lr0=1.680e-04, train_time=1.567, time=1 hour, 37 minutes and 57.01 seconds, total_count=135000, gpu_max_cached_mem_GB=43.801, [valid] loss_ctc=19.854, cer_ctc=0.106, loss_att=9.709, acc=0.925, cer=0.053, wer=0.758, loss=12.753, time=7 minutes and 35.56 seconds, total_count=28980, gpu_max_cached_mem_GB=43.801
[w005] 2025-05-21 20:29:34,141 (trainer:462) INFO: The best model has been updated: valid.acc, valid.total_count
[w005] 2025-05-21 20:29:34,143 (trainer:516) INFO: The model files were removed: exp/s2t_lora_2/8epoch.pth
[w005] 2025-05-21 20:29:34,143 (trainer:324) INFO: 10/10epoch started. Estimated time to finish: 1 hour, 46 minutes and 37.08 seconds
[w005] 2025-05-21 20:34:27,882 (trainer:811) INFO: 10epoch:train:1-750batch: iter_time=3.074e-04, forward_time=0.175, loss_ctc=52.649, loss_att=26.823, acc=0.794, loss=8.643, backward_time=0.181, grad_norm=33.032, clip=100.000, loss_scale=4.295e+09, optim_step_time=0.024, optim0_lr0=1.606e-04, train_time=1.567
[w005] 2025-05-21 20:39:27,035 (trainer:811) INFO: 10epoch:train:751-1500batch: iter_time=1.276e-04, forward_time=0.176, loss_ctc=53.001, loss_att=27.259, acc=0.792, loss=8.745, backward_time=0.183, grad_norm=32.148, clip=100.000, loss_scale=7.151e+09, optim_step_time=0.048, optim0_lr0=1.599e-04, train_time=1.595
[w005] 2025-05-21 20:44:22,156 (trainer:811) INFO: 10epoch:train:1501-2250batch: iter_time=1.289e-04, forward_time=0.177, loss_ctc=53.373, loss_att=27.494, acc=0.796, loss=8.814, backward_time=0.181, grad_norm=31.730, clip=100.000, loss_scale=8.590e+09, optim_step_time=0.024, optim0_lr0=1.592e-04, train_time=1.574
[w005] 2025-05-21 20:49:16,380 (trainer:811) INFO: 10epoch:train:2251-3000batch: iter_time=1.286e-04, forward_time=0.177, loss_ctc=52.888, loss_att=27.156, acc=0.795, loss=8.719, backward_time=0.181, grad_norm=32.980, clip=100.000, loss_scale=8.590e+09, optim_step_time=0.024, optim0_lr0=1.585e-04, train_time=1.569
[w005] 2025-05-21 20:54:12,520 (trainer:811) INFO: 10epoch:train:3001-3750batch: iter_time=1.504e-04, forward_time=0.177, loss_ctc=53.500, loss_att=27.397, acc=0.790, loss=8.807, backward_time=0.183, grad_norm=31.821, clip=100.000, loss_scale=8.590e+09, optim_step_time=0.024, optim0_lr0=1.578e-04, train_time=1.580
[w005] 2025-05-21 20:59:07,448 (trainer:811) INFO: 10epoch:train:3751-4500batch: iter_time=1.489e-04, forward_time=0.176, loss_ctc=52.429, loss_att=26.906, acc=0.794, loss=8.641, backward_time=0.181, grad_norm=31.859, clip=100.000, loss_scale=8.590e+09, optim_step_time=0.024, optim0_lr0=1.571e-04, train_time=1.573
[w005] 2025-05-21 21:04:01,385 (trainer:811) INFO: 10epoch:train:4501-5250batch: iter_time=1.295e-04, forward_time=0.176, loss_ctc=52.756, loss_att=27.137, acc=0.795, loss=8.706, backward_time=0.181, grad_norm=32.116, clip=100.000, loss_scale=8.590e+09, optim_step_time=0.024, optim0_lr0=1.564e-04, train_time=1.568
[w005] 2025-05-21 21:08:54,849 (trainer:811) INFO: 10epoch:train:5251-6000batch: iter_time=1.299e-04, forward_time=0.175, loss_ctc=52.924, loss_att=27.060, acc=0.794, loss=8.705, backward_time=0.181, grad_norm=33.698, clip=100.000, loss_scale=8.590e+09, optim_step_time=0.024, optim0_lr0=1.557e-04, train_time=1.565
[w005] 2025-05-21 21:13:47,826 (trainer:811) INFO: 10epoch:train:6001-6750batch: iter_time=1.291e-04, forward_time=0.175, loss_ctc=51.753, loss_att=26.608, acc=0.794, loss=8.538, backward_time=0.181, grad_norm=31.946, clip=100.000, loss_scale=8.590e+09, optim_step_time=0.025, optim0_lr0=1.550e-04, train_time=1.563