Unpickling error when loading `GDSC1`

[abearab]

Describe the bug
similar to #306

To Reproduce

from tdc.multi_pred import DrugRes
GDSC1 = DrugRes(name = 'GDSC1')

---------------------------------------------------------------------------
UnpicklingError                           Traceback (most recent call last)
Cell In[18], line 1
----> 1 GDSC1 = DrugRes(name = 'GDSC1')
      2 # GDSC2 = DrugRes(name = 'GDSC2')

File ~/anaconda3/envs/scez/lib/python3.11/site-packages/tdc/multi_pred/drugres.py:35, in DrugRes.__init__(self, name, path, label_name, print_stats)
     33 def __init__(self, name, path="./data", label_name=None, print_stats=False):
     34     """Create Drug Response Prediction dataloader object"""
---> 35     super().__init__(name,
     36                      path,
     37                      label_name,
     38                      print_stats,
     39                      dataset_names=dataset_names["DrugRes"])
     40     self.entity1_name = "Drug"
     41     self.entity2_name = "Cell Line"

File ~/anaconda3/envs/scez/lib/python3.11/site-packages/tdc/multi_pred/bi_pred_dataset.py:70, in DataLoader.__init__(self, name, path, label_name, print_stats, dataset_names, data_config)
     66 else:
     67     aux_column = None
     69 (entity1, entity2, raw_y, entity1_idx, entity2_idx, aux_column_val, df,
---> 70  augment_df) = interaction_dataset_load(name,
     71                                         path,
     72                                         label_name,
     73                                         dataset_names,
     74                                         aux_column=aux_column,
     75                                         data_config=data_config)
     77 self.name = name
     78 self.entity1 = entity1

File ~/anaconda3/envs/scez/lib/python3.11/site-packages/tdc/utils/load.py:462, in interaction_dataset_load(name, path, target, dataset_names, aux_column, data_config)
    460 name = download_wrapper(name, path, dataset_names)
    461 print_sys("Loading...")
--> 462 df = pd_load(name, path)
    463 if data_config is not None:
    464     # code block to apply preprocessing rules defined by config files
    465     process_callback = data_config.processing_callback

File ~/anaconda3/envs/scez/lib/python3.11/site-packages/tdc/utils/load.py:301, in pd_load(name, path)
    299     df = pd.read_excel(os.path.join(path, name + "." + name2type[name]))
    300 elif name2type[name] == "pkl":
--> 301     df = pd.read_pickle(os.path.join(path,
    302                                      name + "." + name2type[name]))
    303 elif name2type[name] == "zip":
    304     df = pd.read_pickle(os.path.join(path, name + "/" + name + ".pkl"))

File ~/anaconda3/envs/scez/lib/python3.11/site-packages/pandas/io/pickle.py:202, in read_pickle(filepath_or_buffer, compression, storage_options)
    199     with warnings.catch_warnings(record=True):
    200         # We want to silence any warnings about, e.g. moved modules.
    201         warnings.simplefilter("ignore", Warning)
--> 202         return pickle.load(handles.handle)
    203 except excs_to_catch:
    204     # e.g.
    205     #  "No module named 'pandas.core.sparse.series'"
    206     #  "Can't get attribute '__nat_unpickle' on <module 'pandas._libs.tslib"
    207     return pc.load(handles.handle, encoding=None)

UnpicklingError: invalid load key, '<'.

Environment:

• OS: linux
• Python version: python3.11
• TDC version: 1.1.14