psacfc / gpt4ts_adapter Goto Github PK

There is no exp_long_term_forecasting.py file in the exp directory

Thank you so much for your outstanding work.

I got a bug when using ./Long-term/run_longExp.py

Args in experiment:
Namespace(task_name='long_term_forecast', is_training=1, model_id='test', model='GPT4TS', gpt_layers=6, adapter_layer=2, spect_adapter_layer=2, T_type=1, C_type=1, adapter_dim=32, adapter_dropout=0.1, scale=1, data='custom', root_path='./dataset/', data_path='NN_HWFET_-20.csv', features='M', target='SOE', freq='h', checkpoints='./checkpoints/', seq_len=96, label_len=48, pred_len=96, seasonal_patterns='Monthly', mask_rate=0.25, anomaly_ratio=0.25, top_k=5, num_kernels=6, enc_in=5, dec_in=5, c_out=5, d_model=128, n_heads=8, e_layers=2, d_layers=1, d_ff=512, moving_avg=25, factor=1, distil=True, dropout=0.1, embed='timeF', activation='gelu', output_attention=False, num_workers=0, itr=1, train_epochs=10, batch_size=32, patience=3, learning_rate=0.0001, des='test', loss='MSE', lradj='type1', use_amp=False, use_gpu=True, gpu=0, use_multi_gpu=False, devices='0,1', p_hidden_dims=[128, 128], p_hidden_layers=2, alpha=0.5, beta=0.5, dp_rank=8, rescale=1, fc_dropout=0.3, head_dropout=0.3, patch_len=16, stride=8, momentum=0.1, optimizer='adam', local_rank=0, devices_number=1, use_statistic=False, use_decomp=False, same_smoothing=False, warmup_epochs=0, weight_decay=0, pct_start=0.3, seg_len=6, win_size=2)
Use GPU: cuda:0
Some weights of GPT2Model were not initialized from the model checkpoint at gpt2 and are newly initialized: ['adapter_layers_gate', 'h.0.adapter_layers_gate_C', 'h.0.adapter_layers_gate_T', 'h.0.attn.adapter_gate', 'h.0.attn.adapter_layers_gate', 'h.1.adapter_layers_gate_C', 'h.1.adapter_layers_gate_T', 'h.1.attn.adapter_gate', 'h.1.attn.adapter_layers_gate', 'h.10.adapter_layers_gate_C', 'h.10.adapter_layers_gate_T', 'h.10.attn.adapter_gate', 'h.10.attn.adapter_layers_gate', 'h.11.adapter_layers_gate_C', 'h.11.adapter_layers_gate_T', 'h.11.attn.adapter_gate', 'h.11.attn.adapter_layers_gate', 'h.2.adapter_layers_gate_C', 'h.2.adapter_layers_gate_T', 'h.2.attn.adapter_gate', 'h.2.attn.adapter_layers_gate', 'h.3.adapter_layers_gate_C', 'h.3.adapter_layers_gate_T', 'h.3.attn.adapter_gate', 'h.3.attn.adapter_layers_gate', 'h.4.adapter_layers_gate_C', 'h.4.adapter_layers_gate_T', 'h.4.attn.adapter_gate', 'h.4.attn.adapter_layers_gate', 'h.5.adapter_layers_gate_C', 'h.5.adapter_layers_gate_T', 'h.5.attn.adapter_gate', 'h.5.attn.adapter_layers_gate', 'h.6.adapter_layers_gate_C', 'h.6.adapter_layers_gate_T', 'h.6.attn.adapter_gate', 'h.6.attn.adapter_layers_gate', 'h.7.adapter_layers_gate_C', 'h.7.adapter_layers_gate_T', 'h.7.attn.adapter_gate', 'h.7.attn.adapter_layers_gate', 'h.8.adapter_layers_gate_C', 'h.8.adapter_layers_gate_T', 'h.8.attn.adapter_gate', 'h.8.attn.adapter_layers_gate', 'h.9.adapter_layers_gate_C', 'h.9.adapter_layers_gate_T', 'h.9.attn.adapter_gate', 'h.9.attn.adapter_layers_gate']
You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
number of self.gpt2 params: 120396
number of self.in_layer params: 2176
number of self.out_layer params: 589920
number of self.fft_adapter params: 152

start training : long_term_forecast_test_GPT4TS_custom_ftM_sl96_ll48_pl96_dm128_nh8_el2_dl1_df512_fc1_ebtimeF_dtTrue_test_0>>>>>>>>>>>>>>>>>>>>>>>>>>
train 29655
val 42421
test 84934
Updating learning rate to 0.0000976
number of self.model params: 783044
0it [00:00, ?it/s]
Traceback (most recent call last):
File "D:\pythonProject\test1\GPT4TS_Adapter-main\Long-term\run_longExp.py", line 178, in
exp.train(setting)
File "D:\pythonProject\test1\GPT4TS_Adapter-main\Long-term\exp\exp_main.py", line 168, in train
for i, (batch_x, batch_y, batch_x_mark, batch_y_mark) in tqdm(enumerate(train_loader)):
ValueError: too many values to unpack (expected 4)

Thank you very much for being able to solve it.