re added num_beams

models/blenderbot.py

@@ -23,6 +23,6 @@ def generate(input_text):
   input_ids = tokenizer.encode(input_text, return_tensors="pt")
   
   # Generate output using the model
-  output_ids = model.generate(input_ids, no_repeat_ngram_size=2, max_new_tokens=100, eos_token_id=tokenizer.eos_token_id)
+  output_ids = model.generate(input_ids, no_repeat_ngram_size=2, max_new_tokens=200, num_beams=2, eos_token_id=tokenizer.eos_token_id)
   
   return tokenizer.decode(output_ids[0], skip_special_tokens=True)

models/fast.py

@@ -11,6 +11,6 @@ def load():
 
 def generate(input_text):
   input_ids = tokenizer.encode(input_text, return_tensors="pt")
-  output_ids = model.generate(input_ids, no_repeat_ngram_size=2, max_new_tokens=100, eos_token_id=tokenizer.eos_token_id)
+  output_ids = model.generate(input_ids, no_repeat_ngram_size=2, max_new_tokens=200, num_beams=2, eos_token_id=tokenizer.eos_token_id)
   
   return tokenizer.decode(output_ids[0], skip_special_tokens=True)

models/gpt2.py

@@ -16,6 +16,6 @@ def generate(input_text):
   attention_mask = tf.ones_like(input_ids)
   
   # Generate output using the model
-  output_ids = model.generate(input_ids, num_beams=5, no_repeat_ngram_size=2, max_new_tokens=100, eos_token_id=tokenizer.eos_token_id)
+  output_ids = model.generate(input_ids, num_beams=3, no_repeat_ngram_size=2, max_new_tokens=100, eos_token_id=tokenizer.eos_token_id)
   
   return tokenizer.decode(output_ids[0], skip_special_tokens=True)

models/hermes.py

@@ -13,6 +13,6 @@ tokenizer = AutoTokenizer.from_pretrained(model_name)
   
 def generate(messages):
   gen_input = tokenizer.apply_chat_template(messages, return_tensors="pt")
-  output_ids = model.generate(**gen_input, num_beams=5, no_repeat_ngram_size=2, max_new_tokens=100, eos_token_id=tokenizer.eos_token_id)
+  output_ids = model.generate(**gen_input, num_beams=3, no_repeat_ngram_size=2, max_new_tokens=100, eos_token_id=tokenizer.eos_token_id)
   
   return tokenizer.decode(output_ids[0], skip_special_tokens=True)

models/llama2.py

@@ -11,6 +11,6 @@ def load():
 
 def generate(input_text):
   input_ids = tokenizer.encode(input_text, return_tensors="pt")
-  output_ids = model.generate(input_ids, no_repeat_ngram_size=2, max_new_tokens=100, eos_token_id=tokenizer.eos_token_id)
+  output_ids = model.generate(input_ids, no_repeat_ngram_size=2, max_new_tokens=200, num_beams=2, eos_token_id=tokenizer.eos_token_id)
   
   return tokenizer.decode(output_ids[0], skip_special_tokens=True)

models/llama3.py

@@ -11,6 +11,6 @@ def load():
 
 def generate(input_text):
   input_ids = tokenizer.encode(input_text, return_tensors="pt")
-  output_ids = model.generate(input_ids, no_repeat_ngram_size=2, max_new_tokens=100, eos_token_id=tokenizer.eos_token_id)
+  output_ids = model.generate(input_ids, no_repeat_ngram_size=2, max_new_tokens=200, num_beams=2, eos_token_id=tokenizer.eos_token_id)
   
   return tokenizer.decode(output_ids[0], skip_special_tokens=True)

models/llamatiny.py

@@ -11,6 +11,6 @@ def load():
 
 def generate(input_text):
   input_ids = tokenizer.encode(input_text, return_tensors="pt")
-  output_ids = model.generate(input_ids, no_repeat_ngram_size=2, max_new_tokens=100, eos_token_id=tokenizer.eos_token_id)
+  output_ids = model.generate(input_ids, no_repeat_ngram_size=2, max_new_tokens=200, num_beams=2, eos_token_id=tokenizer.eos_token_id)
   
   return tokenizer.decode(output_ids[0], skip_special_tokens=True)

models/mamba.py

@@ -11,6 +11,6 @@ def load():
 
 def generate(input_text):
   input_ids = tokenizer.encode(input_text, return_tensors="pt")
-  output_ids = model.generate(input_ids, no_repeat_ngram_size=2, max_new_tokens=100, eos_token_id=tokenizer.eos_token_id)
+  output_ids = model.generate(input_ids, no_repeat_ngram_size=2, max_new_tokens=200, num_beams=2, eos_token_id=tokenizer.eos_token_id)
   
   return tokenizer.decode(output_ids[0], skip_special_tokens=True)