Add eos padding to postprocessing (#1783)

language/mixtral-8x7b/dataset.py

@@ -90,7 +90,14 @@ def postProcess(self, out_tokens, input_seq_lens=None,
         # Everything is padded to max_len (1024), so prune the input and parse
         # to numpy
         output_seq = out_tokens[:, 1024:].cpu().numpy()
+        aux_seq = []
         assert len(query_id_list) == output_seq.shape[0]
+        for i in range(len(output_seq)):
+            aux = output_seq[i]
+            while(len(output_seq[i]) <= 1):
+                aux = np.append(aux, self.tokenizer.eos_token_id)
+            aux_seq.append(aux)
+        output_seq = np.stack(aux_seq)
 
         # Save outputs
         if not os.path.exists("run_outputs"):