Unable to finetune model
I have data with columns timestamp, lat-long as itemid and pm2.5 values as target .
I have a long dataframe. I have converted dataframe to uni2ts format and created configuration files according to the finetune steps given in the readme file of github of uni2ts and ran their script
python -m cli.train
-cp conf/finetune
run_name=example_run \
model=moirai_1.0_R_small \
data=etth1 \
val_data=etth1
It created a output folder with hydra files and train logs . When I load the example_run as pretrained and tried to forecast it is giving
MoiraiModule.init() missing 7 required positional arguments: 'distr_output', 'd_model', 'num_layers', 'patch_sizes', 'max_seq_len', 'attn_dropout_p', and 'dropout_p'
problem can be solved by reset the path to the model weights.
go to cli->finetune->model->moirai_1.0_R_small, change the following path setting to the model weight you downloaded from the huggingface, it will work.
module->pretrained_model_name_or_path: Salesforce/moirai-1.0-R-small