Slackbot
09/14/2023, 4:13 PMCristian (Nixtla)
09/14/2023, 4:23 PMnum_samples
runs failed. I suggest starting simpler, with much smaller configs and less num_samples
(3, or 5). There are some issues with the config, for instance the nhits
only have identity
blocks, no trend or seasonality. Also the numbers of elements in n_pool_kernel_size
needs to match the number of stacks (length of stack_types
).Cristian (Nixtla)
09/14/2023, 4:25 PMconfig_nhits = {
"input_size": tune.choice([4, 8]), # Length of input window
"start_padding_enabled": True,
"stack_types": 3*['identity'],
"n_blocks": 3*[1], # Length of input window
"mlp_units": 3 * [[128, 128]], # Length of input window
"n_pool_kernel_size": tune.choice([3*[1], 3*[2], 3*[4]]), # MaxPooling Kernel size
"n_freq_downsample": tune.choice([[4, 2, 1],
[1, 1, 1]]), # Interpolation expressivity ratios
"learning_rate": tune.loguniform(1e-4, 1e-2), # Initial Learning rate
"scaler_type": tune.choice(['robust', 'standard']), # Scaler type
"max_steps": tune.choice([500, 1000]), # Max number of training iterations
"batch_size": tune.choice([32,64,128]), # Number of series in batch
"windows_batch_size": tune.choice([64, 128, 256]), # Number of windows in batch
"random_seed": tune.randint(1, 20), # Random seed
}
Cristian (Nixtla)
09/14/2023, 4:27 PMmlp_units
of 3 is extremely small, I changed it to 128. batch_size
and windows_batch_size
of 1 are also too small. Let me know if the config above works. Try with num_samples=2
while debugging.Naveen Chandra
09/14/2023, 4:31 PMfilter_nan_and_inf
arg to False."Cristian (Nixtla)
09/14/2023, 4:32 PMnum_samples
runs failed, or have a NaN lossCristian (Nixtla)
09/14/2023, 4:32 PMNaveen Chandra
09/14/2023, 4:55 PMNaveen Chandra
09/14/2023, 5:07 PM