Fix generation not concluding when eos was hit for each prompt

src/mistral_inference/generate.py

@@ -112,7 +112,7 @@ def generate(
         next_token = sample(last_token_prelogits, temperature=temperature, top_p=0.8)
 
         if eos_id is not None:
-            is_finished = is_finished ^ (next_token == eos_id).cpu()
+            is_finished = is_finished | (next_token == eos_id).cpu()
 
         if is_finished.all():
             break