Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,23 +1,105 @@
|
|
| 1 |
-
|
| 2 |
-
# pip install git+https://github.com/huggingface/transformers.git
|
| 3 |
-
# pip install accelerate
|
| 4 |
-
|
| 5 |
import torch
|
| 6 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8 |
pipe = pipeline("text-generation", model="TinyLlama/TinyLlama-1.1B-Chat-v1.0", torch_dtype=torch.bfloat16, device_map="auto")
|
| 9 |
|
| 10 |
-
#
|
| 11 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12 |
{
|
| 13 |
"role": "system",
|
| 14 |
-
"content": "You are a friendly chatbot",
|
| 15 |
},
|
| 16 |
-
{"role": "user", "content": "
|
| 17 |
-
]
|
| 18 |
-
prompt = pipe.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
|
| 19 |
-
outputs = pipe(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
|
| 20 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
# <|system|>
|
| 22 |
# You are a friendly chatbot who always responds in the style of a pirate.</s>
|
| 23 |
# <|user|>
|
|
|
|
| 1 |
+
from transformers import AutoTokenizer, AutoModelForSequenceClassification, pipeline
|
|
|
|
|
|
|
|
|
|
| 2 |
import torch
|
| 3 |
+
import pickle
|
| 4 |
+
import streamlit as st
|
| 5 |
+
device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
|
| 6 |
+
|
| 7 |
+
from translate import Translator
|
| 8 |
+
|
| 9 |
+
def init_session_state():
|
| 10 |
+
if 'history' not in st.session_state:
|
| 11 |
+
st.session_state.history = ""
|
| 12 |
|
| 13 |
+
# Initialize session state
|
| 14 |
+
init_session_state()
|
| 15 |
+
# model_name = "MoritzLaurer/mDeBERTa-v3-base-mnli-xnli"
|
| 16 |
+
# tokenizer = AutoTokenizer.from_pretrained(model_name)
|
| 17 |
+
# model = AutoModelForSequenceClassification.from_pretrained(model_name)
|
| 18 |
+
|
| 19 |
+
classifier = pipeline("zero-shot-classification", model="MoritzLaurer/mDeBERTa-v3-base-mnli-xnli")
|
| 20 |
pipe = pipeline("text-generation", model="TinyLlama/TinyLlama-1.1B-Chat-v1.0", torch_dtype=torch.bfloat16, device_map="auto")
|
| 21 |
|
| 22 |
+
# with open('chapter_titles.pkl', 'rb') as file:
|
| 23 |
+
# titles_astiko = pickle.load(file)
|
| 24 |
+
# labels1 = ["κληρονομικό", "ακίνητα", "διαζύγιο"]
|
| 25 |
+
# # labels2 = ["αποδοχή κληρονομιάς", "αποποίηση", "διαθήκη"]
|
| 26 |
+
# # labels3 = ["μίσθωση", "κυριότητα", "έξωση", "απλήρωτα νοίκια"]
|
| 27 |
+
|
| 28 |
+
|
| 29 |
+
# titles_astiko = ["γάμος", "αλλοδαπός", "φορολογία", "κληρονομικά", "στέγη", "οικογενειακό", "εμπορικό","κλοπή","απάτη"]
|
| 30 |
+
# Load dictionary from the file using pickle
|
| 31 |
+
with open('my_dict.pickle', 'rb') as file:
|
| 32 |
+
dictionary = pickle.load(file)
|
| 33 |
+
|
| 34 |
+
def classify(text,labels):
|
| 35 |
+
output = classifier(text, labels, multi_label=False)
|
| 36 |
+
|
| 37 |
+
return output
|
| 38 |
+
|
| 39 |
+
|
| 40 |
+
text = st.text_input('Enter some text:') # Input field for new text
|
| 41 |
+
|
| 42 |
+
if text:
|
| 43 |
+
|
| 44 |
+
labels = list(dictionary)
|
| 45 |
+
|
| 46 |
+
output = classify(text,labels)
|
| 47 |
+
|
| 48 |
+
output = output["labels"][0]
|
| 49 |
+
|
| 50 |
+
labels = list(dictionary[output])
|
| 51 |
+
|
| 52 |
+
output2 = classify(text,labels)
|
| 53 |
+
|
| 54 |
+
output2 = output2["labels"][0]
|
| 55 |
+
|
| 56 |
+
|
| 57 |
+
answer = dictionary[output][output2]
|
| 58 |
+
|
| 59 |
+
# Create a translator object with specified source and target languages
|
| 60 |
+
translator = Translator(from_lang='el', to_lang='en')
|
| 61 |
+
translator2 = Translator(from_lang='en', to_lang='el')
|
| 62 |
+
|
| 63 |
+
|
| 64 |
+
|
| 65 |
+
# Translate the text from Greek to English
|
| 66 |
+
answer = translator.translate(answer)
|
| 67 |
+
text = translator.translate(text)
|
| 68 |
+
|
| 69 |
+
# text_to_translate2 = text[499:999]
|
| 70 |
+
# translated_text2 = translator.translate(text_to_translate2)
|
| 71 |
+
|
| 72 |
+
|
| 73 |
+
|
| 74 |
+
st.session_state.history += "Based on the following information" + answer +"answer this question:" + text + "by reasoning step by step" # Add new text to history
|
| 75 |
+
# out = pipe(st.session_state.history) # Generate output based on history
|
| 76 |
+
|
| 77 |
+
messages = [
|
| 78 |
{
|
| 79 |
"role": "system",
|
| 80 |
+
"content": "You are a friendly chatbot who answers question based on the info that I give you:" + answer,
|
| 81 |
},
|
| 82 |
+
{"role": "user", "content": text"},
|
| 83 |
+
]
|
| 84 |
+
prompt = pipe.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
|
| 85 |
+
outputs = pipe(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
|
| 86 |
+
|
| 87 |
+
|
| 88 |
+
|
| 89 |
+
st.text(st.session_state.history)
|
| 90 |
+
|
| 91 |
+
translated_text2 = translator2.translate(outputs)
|
| 92 |
+
|
| 93 |
+
st.text(translated_text2)
|
| 94 |
+
# st.text("History: " + st.session_state.history)
|
| 95 |
+
|
| 96 |
+
# st.text(output)
|
| 97 |
+
# st.text(output2)
|
| 98 |
+
|
| 99 |
+
# st.text(answer)
|
| 100 |
+
|
| 101 |
+
# We use the tokenizer's chat template to format each message - see https://huggingface.co/docs/transformers/main/en/chat_templating
|
| 102 |
+
|
| 103 |
# <|system|>
|
| 104 |
# You are a friendly chatbot who always responds in the style of a pirate.</s>
|
| 105 |
# <|user|>
|