app.py

"""
import gradio as gr

def mental_chat(message, history):
    return givetext(patienttext,newmodel,newtokenizer)

demo = gr.ChatInterface(mental_chat)

demo.launch()
"""

import gradio as gr
import torch
from peft import PeftModel, PeftConfig
from transformers import AutoModelForCausalLM, AutoTokenizer

# ##### ##### ##### ##### #####

peft_model_id = "charansr/llama2-7b-chat-hf-therapist"
config = PeftConfig.from_pretrained(peft_model_id)
newmodel = AutoModelForCausalLM.from_pretrained(config.base_model_name_or_path, return_dict=True, load_in_8bit=True, device_map='auto')
newtokenizer = AutoTokenizer.from_pretrained(config.base_model_name_or_path)

# Load the Lora model
newmodel = PeftModel.from_pretrained(newmodel, peft_model_id)

def givetext(input_text,lmodel,ltokenizer):
  eval_prompt_pt1 = """\nBelow is an instruction that describes a task. Write a response that appropriately completes the request.\n### Instruction: Act like a therapist and respond\n\n### Input: """
  eval_prompt_pt2="""\n\n\n### Response:\n"""
  eval_prompt=eval_prompt_pt1+input_text+eval_prompt_pt2
  print(eval_prompt,"\n\n")
  model_input = ltokenizer(eval_prompt, return_tensors="pt").to("cuda")

  lmodel.eval()
  with torch.no_grad():
    return (ltokenizer.decode(lmodel.generate(**model_input, max_new_tokens=1000)[0], skip_special_tokens=True))
    #return (ltokenizer.decode(lmodel.generate(**model_input, max_new_tokens=100)[0], skip_special_tokens=True))

def mental_chat(message, history):
    return givetext(patienttext,newmodel,newtokenizer)

demo = gr.ChatInterface(mental_chat)

demo.launch()