-
Notifications
You must be signed in to change notification settings - Fork 5
Description
Hi,
I seem to be having issues with tensorflow attributes when running gluer.
Is there a particular version of tensorflow needed for GLUER?
Thanks
example error:
RUN:
gluer_obj = gr.gluer(rna_data, acc_data, batch_categories=['RNA','ACC'])
Four steps are as follows:
Step 1: Jointly dimension reduction model
Step 2: Search the cell pairs between the reference and the query
Step 3: Run the deep learning model
Step 4: Summarize the output
2021-05-04 14:44:05 >> Step 1: Jointly dimension reduction model ... Done 0.04 mins
2021-05-04 14:44:07 >> Step 2-1: Search the cell pairs ... before filtering: 3142
found 393 pairs ... Done 0.06 mins
2021-05-04 14:44:08 >> Step 3: Run the deep learning model ...
AttributeError Traceback (most recent call last)
in
----> 1 gluer_obj = gr.gluer(rna_data, acc_data, batch_categories=['RNA','ACC'])
/anaconda3/lib/python3.7/site-packages/gluer/gluer.py in gluer(ref_obj, query_obj, joint_rank, joint_max_iter, joint_random_seed, mnn_ref, mnn_query, filter_n1, filter_n2, N, n_jobs, n_features, is_impute, filter_n_features, pairs, deep_random_seed, deepmodel_epoch, batch_categories, model, validation_split, verbose)
238 # train the deep learning model
239 if model is None:
--> 240 tf.random.set_random_seed(deep_random_seed)
241 start_time = time.time()
242 model = tf.keras.Sequential()
AttributeError: module 'tensorflow._api.v2.random' has no attribute 'set_random_seed'