app

app.py

@@ -410,9 +410,10 @@ def mutated_seq(wt_seq, wt_label):
     print(f'Wild Type: Label = {wt_label}, Y_pred = {wt_pred.item()}, Y_prob = {wt_prob.item():.2%}')
     
 #     print(n_mut, mlm_tok_num, n_designs_ep, n_sampling_designs_ep, n_mlm_recovery_sampling, mutate2stronger)
-    pbar = tqdm(total=n_mut)
+    # pbar = tqdm(total=n_mut)
     mutated_seqs = []
     i = 1
+    pbar = st.progress(i, text="mutated number of sequence")
     while i <= n_mut:
         if i == 1: seeds_ep = [wt_seq[1:]]
         seeds_next_ep, seeds_probs_next_ep, seeds_logits_next_ep = [], [], []
@@ -475,9 +476,10 @@ def mutated_seq(wt_seq, wt_label):
 
         seeds_ep = seeds_next_ep
         i += 1
-        pbar.update(1)
-    pbar.close()
-
+        # pbar.update(1)
+        pbar.progress(i/n_mut, text="Mutating")
+    # pbar.close()
+    st.success('Done', icon="✅")
     mutated_seqs.extend([(wt_seq[1:], wt_logit.item(), wt_prob.item(), 0)])
     mutated_seqs = sorted(mutated_seqs, key=lambda x: x[2], reverse=True)
     mutated_seqs = pd.DataFrame(mutated_seqs, columns = ['mutated_seq', 'predicted_logit', 'predicted_probability', 'mutated_num']).drop_duplicates('mutated_seq')
@@ -503,7 +505,7 @@ def read_raw(raw_input):
     return ids, sequences
 
 def predict_raw(raw_input):
-    state_dict = torch.load('v2.7_LeidenContrastive_best_model_fold0.pt', map_location=torch.device(device))
+    state_dict = torch.load('model.pt', map_location=torch.device(device))
     new_state_dict = OrderedDict()
 
     for k, v in state_dict.items():