Update app.py

app.py

@@ -1,49 +1,18 @@
-# -*- coding: utf-8 -*-
-"""Load Model and Run Gradio - llama.ipynb
-
-Automatically generated by Colab.
-
-Original file is located at
-    https://colab.research.google.com/drive/1IQ2EW-KFfdkxEL8sZSfXA0WcS7ZHVPIf
-"""
-
 import os
 token=os.environ['token']
-# !pip install gradio --quiet
-# !pip install requests --quiet
-# !pip install -Uq xformers --index-url https://download.pytorch.org/whl/cu121
-
 import torch
 import gradio as gr
+from unsloth import FastLanguageModel
+from peft import PeftConfig, PeftModel, get_peft_model
+from transformers import pipeline, TextIteratorStreamer
+from threading import Thread
 
 # For getting tokenizer()
 model_id = "meta-llama/Meta-Llama-3-8B-Instruct"
 peft_model_adapter_id = "nttwt1597/test_v2_cancer_v3"
 
-# model_directory = "./model/"
-
-# device = "cuda" if torch.cuda.is_available() else "cpu"
-# print("Using:", device)
-
-#Commented out IPython magic to ensure Python compatibility.
-#%%capture
-# major_version, minor_version = torch.cuda.get_device_capability()
-# Must install separately since Colab has torch 2.2.1, which breaks packages
-#!pip install "unsloth[colab-new] @ git+https://github.com/unslothai/unsloth.git"
-# if major_version >= 8:
-#     # Use this for new GPUs like Ampere, Hopper GPUs (RTX 30xx, RTX 40xx, A100, H100, L40)
-#     !pip install --no-deps packaging ninja einops flash-attn xformers trl peft accelerate bitsandbytes
-# else:
-#     # Use this for older GPUs (V100, Tesla T4, RTX 20xx)
-#     !pip install --no-deps xformers trl peft accelerate bitsandbytes
-# pass
-
-# cuda 12.1 version
-from unsloth import FastLanguageModel
-from peft import PeftConfig, PeftModel, get_peft_model
-
 model, tokenizer = FastLanguageModel.from_pretrained(
-    model_name = model_id, # YOUR MODEL YOU USED FOR TRAINING
+    model_name = model_id, 
     max_seq_length = 4096,
     dtype = None,
     load_in_4bit = True,
@@ -55,10 +24,7 @@ terminators = [
     tokenizer.convert_tokens_to_ids("<|eot_id|>")
 ]
 
-FastLanguageModel.for_inference(model) # Enable native 2x faster inference
-
-from transformers import pipeline, TextIteratorStreamer
-from threading import Thread
+FastLanguageModel.for_inference(model) 
 
 criteria_prompt = """Based on the provided instructions and clinical trial information, generate the eligibility criteria for the study.
 
@@ -90,45 +56,41 @@ def run_model_on_text(text):
     yield generated_text
 
 place_holder = f"""Study Objectives
-Optical diagnosis of colorectal polyps is a promising tool to avoid risks of unnecessary polypectomies and to save costs of tissue pathology. NICE (NBI International Colorectal Endoscopic) and WASP (Workgroup on Serrated Polyps and Polyposis) classification were developed for diagnosis of adenomatous and sessile serrated polyps, respectively.
+The purpose of this study is to evaluate the safety, tolerance and efficacy of Liposomal Paclitaxel With Nedaplatin as First-line in patients with Advanced or Recurrent Esophageal Carcinoma 
 
-Near-focus (NF) narrow-band imaging (NBI) is an image-magnifying technology which enables optical magnification of up to 65x in near focus (NF) compared with 52x in normal standard focus (SF) with the simple push of a button of the endoscope to be interchangeable between NF and SF. There were few studies comparing diagnostic accuracy between NF and SF in the diagnosis of colorectal polyps. So, our aim of the current study is to compare accuracy of NF NBI compared with SF NBI in the optical diagnosis of neoplastic and non-neoplastic polyp and the accuracy of NF NBI versus SF NBI in distinguishing serrated adenoma from hyperplastic polyp in sessile lesions using histologic evaluation as the gold standard.
+Conditions: Esophageal Carcinoma 
 
-Conditions 
-Colorectal Polyp, Colorectal Neoplasms
+Intervention / Treatment: 
+DRUG: Liposomal Paclitaxel, 
+DRUG: Nedaplatin 
 
-Intervention / Treatment
-Diagnostic Test: Near Focus NBI
-Diagnostic Test: Standard Focus NBI
+Location: China 
 
-Location 
-Hat Yai, Songkhla, Thailand
-
-Study Design and Phases
-Study Type: Interventional
-Phase: Not Applicable
-Primary Purpose: Diagnostic
-Allocation: Randomized
-Interventional Model: Parallel Assignment
-Masking: None (Open Label)
+Study Design and Phases 
+Study Type: INTERVENTIONAL 
+Phase: PHASE2 Primary Purpose: 
+TREATMENT Allocation: NA 
+Interventional Model: SINGLE_GROUP Masking: NONE
 """
 
 prefilled_value = """Study Objectives
 [Brief Summary] and/or [Detailed Description]
 
-Conditions
+Conditions: [Disease]
 
 Intervention / Treatment 
+[DRUGs]
 
 Location
+[Location]
 
 Study Design and Phases
-Study Type
-Phase
-Primary Purpose
-Allocation 
-Interventional Model
-Masking"""
+Study Type:
+Phase:
+Primary Purpose:
+Allocation: 
+Interventional Model:
+Masking:"""
 
 prompt_box = gr.Textbox(
     lines=25,
@@ -149,4 +111,4 @@ demo = gr.Interface(
   allow_flagging='auto',
 )
 
-demo.queue(max_size=20).launch(debug=True) #share=True 
+demo.queue(max_size=20).launch(debug=True, share=True)