Running app.py about healifyai--llm-based-healthcare-system HOT 6 CLOSED

Hey, I fixed it. I had to do gradio==4.8.0 and change to "
label = gr.Label(num_top_classes=5)"

from healifyai--llm-based-healthcare-system.

@jainishmehta
Gradio package is possibly bringing in changes. Pretty recent.

GradioDeprecationWarning: Usage of gradio.outputs is deprecated, and will not be supported in the future, please import your components from gradio.components

Did you use/follow the requirements.txt of the project? As it is stated,
gradio==3.34.0
As always remember the python best practice of using virtual environment for installing packages.

from healifyai--llm-based-healthcare-system.

I directly deployed on huggingface where there was no issue. So, I am sure it is supposed to work locally which is what you are doing. I am sharing requirements.txt of huggingface directly here. Try installing that on a fresh virtual environment:

gradio==3.34.0
gradio_client==0.2.6
ohmeow-blurr==1.0.5
torch==1.13.1
transformers==4.33.2

In the upcoming days I will check the app code locally if possible in my free time.

from healifyai--llm-based-healthcare-system.

thanks :)

from healifyai--llm-based-healthcare-system.

I use this code on your app.py now:

import gradio as gr
from transformers import AutoTokenizer
import torch, json

from fastai.text.all import *
from blurr.text.modeling.all import *

from blurr.text.data.all import *
from fastai.vision.all import *

with open('question_labels.json', 'r') as f:
question_dictionary = json.load(f)
que_classes = list(question_dictionary.keys())

blurr_model = load_learner('healifyLLM-stage4.pkl')
probs = blurr_model.blurr_predict('headache')[0]['probs']
values_and_probs = dict(zip(que_classes, map(float, probs)))
print(sorted(values_and_probs.items(), key=lambda x: x[1], reverse=True)[:5])

def detect_question(text):
blurr_model = load_learner('healifyLLM-stage4.pkl')
probs = blurr_model.blurr_predict(text)[0]['probs']
values_and_probs = dict(zip(que_classes, map(float, probs)))
return dict(sorted(values_and_probs.items(), key=lambda x: x[1], reverse=True))

label = gr.outputs.Label(num_top_classes=5)
iface = gr.Interface(fn=detect_question, inputs="text", outputs = label)
if name == "main":
iface.launch(show_api=False)

I get the right values from your model (so the model is not the issue)
however I keep getting error:

from healifyai--llm-based-healthcare-system.

That’s fantastic news! 🎉
I think I will bring a second update to the HuggingFace gradio app soon, to make it a bit more interactive. Feel free to check it out.

from healifyai--llm-based-healthcare-system.