Update README.md

README.md

@@ -191,10 +191,11 @@ out = model.generate(
     max_new_tokens=60,
     do_sample=True,
     temperature=0.7,
-    top_k=20,
+    top_k=40,
     repetition_penalty=1.3,
     no_repeat_ngram_size=3,
     pad_token_id=tok.pad_token_id,
+    top_p=0.95
 )
 print(tok.decode(out[0], skip_special_tokens=True))
 ```
@@ -219,9 +220,10 @@ These settings gave fluent, on-topic output in testing:
 
 ```text
 temperature        0.7
-top_k              20
+top_k              40
 repetition_penalty 1.3
 no_repeat_ngram    3
+top_p              0.95
 ```
 
 Greedy decoding (`do_sample=False`) repeats phrases after a few sentences, so use sampling.