Spaces:
Runtime error
Runtime error
Commit
·
8e0ff6b
1
Parent(s):
124d50e
create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
|
3 |
+
DESCRIPTION = """
|
4 |
+
|
5 |
+
### Fine Tuned DistilBERT for Skills Recognition
|
6 |
+
|
7 |
+
This Space demonstrates model [lm-ner-linkedin-skills-recognition](https://huggingface.co/algiraldohe/lm-ner-linkedin-skills-recognition?text=Python+is+a+programming+language) by Alejandro G and Aron Gyenge, a DistilBERT model with ~66M parameters fine-tuned for Named Entity Recognition (NER).
|
8 |
+
#### Objective:
|
9 |
+
The purpose of this app is to show how an LLM can be fine-tuned for the purpose of customed Named Entity Recognition, in this case, given a technical requirements natural language text (Like a job post from LinkedIn) identify the different kind of skills and technologies that
|
10 |
+
a candidate may be required to have.
|
11 |
+
|
12 |
+
#### Labels:
|
13 |
+
The labels that the model was trained to recognise were:
|
14 |
+
- Technical: A specific and practical ability or knowledge that enables a person to perform a particular task or use a specific tool. (eg. data analysis, statistics, maths)
|
15 |
+
- Business: A set of abilities and knowledge that allows individuals to navigate and excel in the world of functional areas within a business. (eg. marketing, finance, accounting)
|
16 |
+
- Soft: A personal attribute or characteristic that enhances an individual's ability to interact effectively with others and navigate various social and professional situations. (eg.communication, learning)
|
17 |
+
- Technology: It encompasses a wide range of devices, systems, and processes that have been developed for usage within a role or function. (eg. Python, AWS)
|
18 |
+
|
19 |
+
|
20 |
+
|
21 |
+
🔎 This is a very light early-stage model that does not reflect the final outcome of the productionised version, but gives a comprehensive view of the main objective with decent results.
|
22 |
+
"""
|
23 |
+
|
24 |
+
gr.Interface.load("models/algiraldohe/lm-ner-linkedin-skills-recognition",
|
25 |
+
inputs=[gr.Textbox(label="Text to find entities", lines=2)],
|
26 |
+
outputs=[gr.HighlightedText(label="Text with entities", lines=2)],
|
27 |
+
examples=[
|
28 |
+
"Python is one of the best programming languages to do data analysis"
|
29 |
+
, "GCP data ops allows you to manage machine learning applications in a cloud environment"
|
30 |
+
, "Data analysis are skills required to join finance and marketing areas, but requires excellent communication with stakeholders"],
|
31 |
+
cache_examples=True,
|
32 |
+
description = DESCRIPTION,
|
33 |
+
title= "LLM on Skills Recognition from LinkedIn").launch()
|
34 |
+
|