Upload 3 files
Browse filesUplaod the model
- README.md +101 -0
- config.json +20 -0
- pytorch_model.bin +3 -0
README.md
ADDED
|
@@ -0,0 +1,101 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
|
| 2 |
+
|
| 3 |
+
---
|
| 4 |
+
language: en
|
| 5 |
+
tags:
|
| 6 |
+
- nlp
|
| 7 |
+
- text-classification
|
| 8 |
+
- social-media-analysis
|
| 9 |
+
- transformers
|
| 10 |
+
- research
|
| 11 |
+
license: apache-2.0
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
# the_poli
|
| 15 |
+
|
| 16 |
+
**the_poli** is a transformer-based NLP classification model developed as part of the **s0m3m0** research project.
|
| 17 |
+
The model is designed to analyse political and social-media-related text and produce structured predictions for analytical and experimental purposes.
|
| 18 |
+
|
| 19 |
+
This repository contains **only the trained model artifacts** (weights and configuration).
|
| 20 |
+
The source code and data pipeline are maintained separately.
|
| 21 |
+
|
| 22 |
+
---
|
| 23 |
+
|
| 24 |
+
## Model Description
|
| 25 |
+
|
| 26 |
+
- **Model type:** Transformer-based text classification model
|
| 27 |
+
- **Framework:** Hugging Face Transformers
|
| 28 |
+
- **Language:** English (primary)
|
| 29 |
+
- **Domain:** Political and social media text analysis
|
| 30 |
+
|
| 31 |
+
The model focuses on extracting patterns and signals from text rather than making authoritative or real-world decisions.
|
| 32 |
+
|
| 33 |
+
---
|
| 34 |
+
|
| 35 |
+
## Intended Use
|
| 36 |
+
|
| 37 |
+
The model is intended for:
|
| 38 |
+
|
| 39 |
+
- Academic and research experimentation
|
| 40 |
+
- NLP pipeline development
|
| 41 |
+
- Social media text analysis
|
| 42 |
+
- Educational demonstrations
|
| 43 |
+
|
| 44 |
+
### Not Intended For
|
| 45 |
+
|
| 46 |
+
- High-stakes decision making
|
| 47 |
+
- Political persuasion or targeting
|
| 48 |
+
- Surveillance, profiling, or enforcement
|
| 49 |
+
- Production systems without extensive validation
|
| 50 |
+
|
| 51 |
+
---
|
| 52 |
+
|
| 53 |
+
## Usage Example
|
| 54 |
+
|
| 55 |
+
```python
|
| 56 |
+
from transformers import AutoTokenizer, AutoModelForSequenceClassification
|
| 57 |
+
|
| 58 |
+
model_id = "d42kw01f/the_poli"
|
| 59 |
+
|
| 60 |
+
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
| 61 |
+
model = AutoModelForSequenceClassification.from_pretrained(model_id)
|
| 62 |
+
|
| 63 |
+
text = "Example political text for analysis"
|
| 64 |
+
inputs = tokenizer(text, return_tensors="pt", truncation=True)
|
| 65 |
+
outputs = model(**inputs)
|
| 66 |
+
```
|
| 67 |
+
|
| 68 |
+
---
|
| 69 |
+
|
| 70 |
+
## Limitations & Biases
|
| 71 |
+
|
| 72 |
+
- Performance depends heavily on the training dataset
|
| 73 |
+
- May reflect biases present in source data
|
| 74 |
+
- Not robust to domain shifts or adversarial inputs
|
| 75 |
+
- Predictions should be interpreted as probabilistic signals, not facts
|
| 76 |
+
|
| 77 |
+
---
|
| 78 |
+
|
| 79 |
+
## Ethical Considerations
|
| 80 |
+
|
| 81 |
+
This model is released **strictly for research and educational use**.
|
| 82 |
+
Users are responsible for:
|
| 83 |
+
- Complying with platform terms of service
|
| 84 |
+
- Respecting data privacy and ethical boundaries
|
| 85 |
+
- Avoiding harmful, misleading, or unethical applications
|
| 86 |
+
|
| 87 |
+
---
|
| 88 |
+
|
| 89 |
+
## Related Project
|
| 90 |
+
|
| 91 |
+
- **GitHub (codebase):** [https://github.com/d42kw01f/s0m3m0](https://github.com/d42kw01f/s0m3m0)
|
| 92 |
+
- **Project name:** s0m3m0
|
| 93 |
+
|
| 94 |
+
---
|
| 95 |
+
|
| 96 |
+
## Author
|
| 97 |
+
|
| 98 |
+
**Dakshitha Navodya Perera**
|
| 99 |
+
AI • Cybersecurity • Data Engineering
|
| 100 |
+
Sri Lanka
|
| 101 |
+
|
config.json
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"attention_probs_droput_prob": 0.1,
|
| 3 |
+
"classifier_dropout": null,
|
| 4 |
+
"hidden_act": "gelu",
|
| 5 |
+
"hidden_dropout_prob": 0.1,
|
| 6 |
+
"hidden_size": 768,
|
| 7 |
+
"initializer_range": 0.02,
|
| 8 |
+
"intermediate_size": 3072,
|
| 9 |
+
"layer_norm_eps": 1e-12,
|
| 10 |
+
"max_position_embeddings": 512,
|
| 11 |
+
"model_type": "bert",
|
| 12 |
+
"num_attention_heads": 12,
|
| 13 |
+
"num_hidden_layers": 12,
|
| 14 |
+
"pad_token_id": 0,
|
| 15 |
+
"position_embedding_type": "absolute",
|
| 16 |
+
"transformers_version": "4.42.4",
|
| 17 |
+
"type_vocab_size": 2,
|
| 18 |
+
"use_cache": true,
|
| 19 |
+
"vocab_size": 30522
|
| 20 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a72dbf8a2c53e5f634431b5aa0c8b11138c3dfb76709e8fa4f31b3bc6aecdfd1
|
| 3 |
+
size 438021310
|