Update README.md

README.md

@@ -6,4 +6,52 @@ language:
 - en
 base_model:
 - google/flan-t5-small
----
+tags:
+  - summarization
+  - research-papers
+  - arxiv
+  - t5
+---
+
+# Fine-Tuned Summarization Model (`fine-tuned-summarization-arxiv`)
+This model is a fine-tuned version of [`google/flan-t5-small`](https://huggingface.co/google/flan-t5-small) on a dataset of armanc/scientific_papers (arxiv). It is optimized for **summarizing scientific abstracts**.
+
+## Model Details
+- **Base Model:** `google/flan-t5-small`
+- **Training Data:** Arxiv Research Papers (`article` → `abstract`)
+- **Fine-Tuned Task:** Text Summarization
+- **Use Case:** Generate shorter summaries of long research papers
+- **License:** Apache 2.0
+
+## How to Use
+```python
+from transformers import T5ForConditionalGeneration, T5Tokenizer
+
+model = T5ForConditionalGeneration.from_pretrained("Talina06/arxiv-summarization")
+tokenizer = T5Tokenizer.from_pretrained("Talina06/arxiv-summarization")
+
+text = "Summarize: Deep learning is being used to advance medical research, particularly in cancer detection."
+inputs = tokenizer(text, return_tensors="pt")
+summary_ids = model.generate(**inputs)
+summary = tokenizer.decode(summary_ids[0], skip_special_tokens=True)
+
+print("Generated Summary:", summary)
+```
+
+## Training Details
+- **Training Data:** 100k+ Arxiv research papers
+- **Training Framework:** Hugging Face Transformers
+- **Hyperparameters:**
+  - Learning Rate: `5e-5`
+  - Batch Size: `8`
+  - Epochs: `10`
+- **Hardware Used:** TPU & GPU
+
+## Limitations
+- ❌ May struggle with **very technical** papers (e.g., complex math formulas).
+
+## Example Summaries
+| **Original Abstract** | **Generated Summary** |
+|----------------------|----------------------|
+| "Deep learning has transformed many fields... We propose a new CNN for cancer detection..." | "A CNN model is proposed for cancer detection using deep learning." |
+| "Quantum computing has shown potential for cryptographic applications..." | "Quantum computing can be used in cryptography." |