fix minor problems in inference example

README.md

@@ -47,7 +47,7 @@ model = AutoModelForCausalLM.from_pretrained(
     pretrained_model_name_or_path="xiao-fei/Prot2Text-V2-11B-Instruct-hf", 
     trust_remote_code=True, 
     torch_dtype=torch.bfloat16, 
-    device_map="auto"
+    device_map="cuda"
 )
 
 esm_tokenizer = AutoTokenizer.from_pretrained("facebook/esm2_t36_3B_UR50D")
@@ -81,16 +81,16 @@ tokenized_prompt = llama_tokenizer.apply_chat_template(
     return_dict=True
 )
 tokenized_sequence = esm_tokenizer(
-    ex_seq, 
+    example_sequence, 
     return_tensors="pt"
 )
 
 model.eval()
 generated = model.generate(
-    inputs=tokenized_prompt["input_ids"].to(model.device()),
-    attention_mask=tokenized_prompt["attention_mask"].to(model.device()),
-    protein_input_ids=tokenized_sequence["input_ids"].to(model.device()),
-    protein_attention_mask=tokenized_sequence["attention_mask"].to(model.device()),
+    inputs=tokenized_prompt["input_ids"].to(model.device),
+    attention_mask=tokenized_prompt["attention_mask"].to(model.device),
+    protein_input_ids=tokenized_sequence["input_ids"].to(model.device),
+    protein_attention_mask=tokenized_sequence["attention_mask"].to(model.device),
     max_new_tokens=1024,
     eos_token_id=128009, 
     pad_token_id=128002,