sudaisdev commited on
Commit
5b1b424
·
verified ·
1 Parent(s): 6252d5a

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +48 -21
README.md CHANGED
@@ -15,46 +15,73 @@ base_model: t5-base
15
 
16
  # Interview Question Generation Model (T5-base)
17
 
18
- This is a fine-tuned T5-base model that generates interview questions
19
- from job details such as job title, description, required experience,
20
- and skills.
21
 
22
- It is designed for HR platforms, hiring tools, and interview preparation
23
- apps.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
24
 
25
  ---
26
 
27
  ## Model Details
28
 
29
- - **Base model:** T5-base
30
- - **Architecture:** T5ForConditionalGeneration (encoder-decoder)
31
- - **Parameters:** ~220M
32
- - **Language:** English
33
- - **Fine-tuned on:** custom interview question dataset
34
- - **Task:** text-to-text generation
 
 
 
 
35
 
36
  ---
37
 
38
  ## Intended Use
39
 
40
- Use this model when you want to automatically generate a single
41
- interview question for a given job role.
42
-
43
- **Good for:**
44
- - HR screening tools
45
  - Interview preparation apps
46
- - Recruitment automation
 
47
 
48
- **Not good for:**
49
- - Generating full interview scripts
50
  - Non-English languages
51
- - Legal or medical advice
 
 
52
 
53
  ---
54
 
55
  ## Input Format
56
 
57
- The model expects **one plain text prompt** in this exact format:
58
 
59
  ```text
60
  Generate an interview question.
 
15
 
16
  # Interview Question Generation Model (T5-base)
17
 
18
+ A fine-tuned **T5-base** model that generates interview questions from
19
+ job details such as job title, description, required experience, and skills.
 
20
 
21
+ Developed for HR platforms, hiring tools, and interview preparation apps.
22
+
23
+ ---
24
+
25
+ ## Model Description
26
+
27
+ This model is based on **T5-base** (Text-to-Text Transfer Transformer).
28
+ It was fine-tuned on a custom dataset of job descriptions paired with
29
+ interview questions. The model takes a structured prompt as input and
30
+ returns a single interview question as output.
31
+
32
+ **Model type:** Encoder-decoder (T5)
33
+ **Language:** English
34
+ **License:** Apache 2.0
35
+ **Fine-tuned from:** `t5-base`
36
+
37
+ ---
38
+
39
+ ## Highlights
40
+
41
+ - **Text-to-text generation:** One input prompt → one interview question.
42
+ - **Multi-role support:** Works for ML, backend, frontend, data, DevOps roles.
43
+ - **Question-type control:** You can specify `Technical`, `Scenario`,
44
+ `Behavioral`, or `Problem-Solving`.
45
+ - **Lightweight:** ~220M parameters, runs on CPU and GPU.
46
+ - **Hugging Face ready:** Works with `transformers`, `pipeline`, and
47
+ the Hugging Face inference widget.
48
 
49
  ---
50
 
51
  ## Model Details
52
 
53
+ | Attribute | Value |
54
+ |---|---|
55
+ | Base model | T5-base |
56
+ | Architecture | T5ForConditionalGeneration |
57
+ | Parameters | ~220M |
58
+ | Model size | ~967 MB (float32) |
59
+ | Context length | 512 tokens |
60
+ | Vocab size | 32128 |
61
+ | Fine-tuned on | Custom interview Q&A dataset |
62
+ | Task | Text-to-text generation |
63
 
64
  ---
65
 
66
  ## Intended Use
67
 
68
+ **Direct use:**
69
+ - HR platforms generating screening questions
 
 
 
70
  - Interview preparation apps
71
+ - Recruitment automation tools
72
+ - Educational content creation
73
 
74
+ **Out-of-scope use:**
 
75
  - Non-English languages
76
+ - Medical, legal, or financial advice
77
+ - Long-form interview scripts
78
+ - Automated hiring decisions without human review
79
 
80
  ---
81
 
82
  ## Input Format
83
 
84
+ The model expects **one plain text prompt** with the following structure:
85
 
86
  ```text
87
  Generate an interview question.