File size: 2,491 Bytes
8e0ff6b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
import gradio as gr

DESCRIPTION = """

### Fine Tuned DistilBERT for Skills Recognition

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).
#### Objective: 
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 
a candidate may be required to have.

#### Labels:
The labels that the model was trained to recognise were:
- 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)
- 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)
- 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)
- 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)



πŸ”Ž 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.
"""

gr.Interface.load("models/algiraldohe/lm-ner-linkedin-skills-recognition",
                  inputs=[gr.Textbox(label="Text to find entities", lines=2)],
                  outputs=[gr.HighlightedText(label="Text with entities", lines=2)],
                  examples=[
                      "Python is one of the best programming languages to do data analysis"
                      , "GCP data ops allows you to manage machine learning applications in a cloud environment"
                      , "Data analysis are skills required to join finance and marketing areas, but requires excellent communication with stakeholders"],
                  cache_examples=True,
                  description = DESCRIPTION,
                  title= "LLM on Skills Recognition from LinkedIn").launch()