Spaces:
Sleeping
Sleeping
Update README.md
Browse files
README.md
CHANGED
@@ -10,7 +10,7 @@ pinned: false
|
|
10 |
license: apache-2.0
|
11 |
short_description: Gradio
|
12 |
---
|
13 |
-
GPT-2
|
14 |
|
15 |
Overview
|
16 |
|
@@ -18,19 +18,19 @@ This project fine-tunes a GPT-2 model on the TinyStories dataset to generate str
|
|
18 |
|
19 |
Features
|
20 |
|
21 |
-
Story Generation: Produces
|
22 |
|
23 |
-
Bias Monitoring: Ensures balanced gender and cultural representation.
|
24 |
|
25 |
-
Efficient Training: Fine-tuned on 200,000 training samples and 20,000 test samples.
|
26 |
|
27 |
-
Grammar & Readability
|
28 |
|
29 |
-
Optimized
|
30 |
|
31 |
-
System Architecture
|
32 |
|
33 |
-
The model is designed for easy interaction via Hugging Face Spaces and follows this
|
34 |
|
35 |
Data Preprocessing & Cleaning
|
36 |
|
@@ -60,7 +60,7 @@ Tracked using Weights & Biases (W&B) and TensorBoard.
|
|
60 |
|
61 |
Analyzed validation loss and coherence checks.
|
62 |
|
63 |
-
Getting Started
|
64 |
|
65 |
Accessing the Model
|
66 |
|
@@ -76,7 +76,7 @@ Click Generate to receive an AI-generated short story.
|
|
76 |
|
77 |
Modify the prompt and settings (temperature, top-k, top-p) for different results.
|
78 |
|
79 |
-
Training Details
|
80 |
|
81 |
Model: GPT-2 (124M)
|
82 |
|
@@ -90,7 +90,7 @@ Training Loss: 3.08 โ 2.86
|
|
90 |
|
91 |
Validation Loss: 1.46 โ 1.40
|
92 |
|
93 |
-
Evaluation & Observations
|
94 |
|
95 |
Perplexity improved from 8.12 โ 2.09, indicating better text fluency.
|
96 |
|
@@ -98,7 +98,7 @@ Validation loss decreased consistently, suggesting effective generalization.
|
|
98 |
|
99 |
Human evaluation highlighted minor inconsistencies, such as abrupt scene shifts and simplistic narratives.
|
100 |
|
101 |
-
Ethical Considerations
|
102 |
|
103 |
Bias Monitoring: Pronoun analysis and diversity checks to ensure fairness.
|
104 |
|
@@ -106,7 +106,7 @@ Harmful Content Mitigation: Manually reviewed outputs for stereotypes.
|
|
106 |
|
107 |
Text Processing Issues: UTF-8 encoding applied to prevent character errors.
|
108 |
|
109 |
-
Future Improvements
|
110 |
|
111 |
Enhancing Creativity: Fine-tune temperature and randomness settings.
|
112 |
|
@@ -114,17 +114,17 @@ Genre-Specific Training: Introduce theme-based datasets.
|
|
114 |
|
115 |
Larger Model Training: Experiment with GPT-2 (355M) for richer storytelling.
|
116 |
|
117 |
-
Contributors
|
118 |
|
119 |
Charla Pia Vella (Project Developer)
|
120 |
|
121 |
Affiliation: ARI3333 Generative AI
|
122 |
|
123 |
-
License
|
124 |
|
125 |
This project is released under the Apache-2.0 License.
|
126 |
|
127 |
-
Acknowledgments
|
128 |
|
129 |
OpenAI for GPT-2
|
130 |
|
@@ -132,7 +132,9 @@ Hugging Face for the fine-tuning framework
|
|
132 |
|
133 |
Ronen Eldan for the TinyStories dataset
|
134 |
|
135 |
-
For more details, visit the Hugging Face Space.
|
|
|
|
|
136 |
|
137 |
Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
|
138 |
|
|
|
10 |
license: apache-2.0
|
11 |
short_description: Gradio
|
12 |
---
|
13 |
+
GPT-2 TinyStories Generator ๐ข - FableWeaver AI
|
14 |
|
15 |
Overview
|
16 |
|
|
|
18 |
|
19 |
Features
|
20 |
|
21 |
+
โ
Story Generation: Produces structured, child-friendly short stories.
|
22 |
|
23 |
+
โ
Bias Monitoring: Ensures balanced gender and cultural representation.
|
24 |
|
25 |
+
โ
Efficient Training: Fine-tuned on 200,000 training samples and 20,000 test samples.
|
26 |
|
27 |
+
โ
Grammar & Readability Enhancements: Integrated grammar-checking tools and text refinement.
|
28 |
|
29 |
+
โ
Optimized Performance: Uses loss tracking, sampling techniques, and bias mitigation strategies.
|
30 |
|
31 |
+
๐ System Architecture
|
32 |
|
33 |
+
The model is designed for easy interaction via Hugging Face Spaces and follows this workflow:
|
34 |
|
35 |
Data Preprocessing & Cleaning
|
36 |
|
|
|
60 |
|
61 |
Analyzed validation loss and coherence checks.
|
62 |
|
63 |
+
๐น Getting Started
|
64 |
|
65 |
Accessing the Model
|
66 |
|
|
|
76 |
|
77 |
Modify the prompt and settings (temperature, top-k, top-p) for different results.
|
78 |
|
79 |
+
๐ Training Details
|
80 |
|
81 |
Model: GPT-2 (124M)
|
82 |
|
|
|
90 |
|
91 |
Validation Loss: 1.46 โ 1.40
|
92 |
|
93 |
+
๐ Evaluation & Observations
|
94 |
|
95 |
Perplexity improved from 8.12 โ 2.09, indicating better text fluency.
|
96 |
|
|
|
98 |
|
99 |
Human evaluation highlighted minor inconsistencies, such as abrupt scene shifts and simplistic narratives.
|
100 |
|
101 |
+
โ๏ธ Ethical Considerations
|
102 |
|
103 |
Bias Monitoring: Pronoun analysis and diversity checks to ensure fairness.
|
104 |
|
|
|
106 |
|
107 |
Text Processing Issues: UTF-8 encoding applied to prevent character errors.
|
108 |
|
109 |
+
๐ฎ Future Improvements
|
110 |
|
111 |
Enhancing Creativity: Fine-tune temperature and randomness settings.
|
112 |
|
|
|
114 |
|
115 |
Larger Model Training: Experiment with GPT-2 (355M) for richer storytelling.
|
116 |
|
117 |
+
๐ค Contributors
|
118 |
|
119 |
Charla Pia Vella (Project Developer)
|
120 |
|
121 |
Affiliation: ARI3333 Generative AI
|
122 |
|
123 |
+
๐ License
|
124 |
|
125 |
This project is released under the Apache-2.0 License.
|
126 |
|
127 |
+
๐ Acknowledgments
|
128 |
|
129 |
OpenAI for GPT-2
|
130 |
|
|
|
132 |
|
133 |
Ronen Eldan for the TinyStories dataset
|
134 |
|
135 |
+
๐ For more details, visit the Hugging Face Space.
|
136 |
+
|
137 |
+
|
138 |
|
139 |
Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
|
140 |
|