changemodel import

app.py

@@ -7,16 +7,16 @@ logger = logging.get_logger("transformers")
 
 # Load the model and tokenizer
 model_name = "TheBloke/Mistral-7B-Instruct-v0.1-GPTQ" #"openai-community/gpt2" or "TheBloke/Mistral-7B-Instruct-v0.1-GPTQ" or "TheBloke/Llama-2-7B-Chat-GGML" or "TheBloke/zephyr-7B-beta-GPTQ"
-tokenizer = AutoTokenizer.from_pretrained(model_name)
+tokenizer = AutoTokenizer.from_pretrained(model_name,use_fast=True)
 #model = AutoModelForSeq2SeqLM.from_pretrained(model_name) 
-model = AutoModelForCausalLM.from_pretrained(model_name, low_cpu_mem_usage=True)
+model = AutoModelForCausalLM.from_pretrained(model_name,device_map="auto",trust_remote_code=False,revision="main")
 
 # Generate text using the model and tokenizer
 def generate_text(input_text):
     input_ids = tokenizer.encode(input_text, return_tensors="pt")
-    attention_mask = input_ids.ne(tokenizer.pad_token_id).long()
-    output = model.generate(input_ids, attention_mask=attention_mask, max_length=100, num_return_sequences=1, no_repeat_ngram_size=2, top_k=50, top_p=0.95, temperature=0.7, do_sample=True)
-    return tokenizer.decode(output[0], skip_special_tokens=True)
+    #attention_mask = input_ids.ne(tokenizer.pad_token_id).long()
+    output = model.generate(input_ids, max_new_tokens=512, top_k=50, top_p=0.95, temperature=0.7, do_sample=True)# attention_mask=attention_mask, max_length=100, num_return_sequences=1, no_repeat_ngram_size=2, top_k=50, top_p=0.95, temperature=0.7, do_sample=True)
+    return tokenizer.decode(output[0])
 
 # def generate_text(prompt):
 #     inputs = tokenizer(prompt, return_tensors="pt", truncation=True, max_length=512, padding="max_length")