Add model card
Browse files
README.md
ADDED
|
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- en
|
| 4 |
+
license: apache-2.0
|
| 5 |
+
tags:
|
| 6 |
+
- eurollm
|
| 7 |
+
- neto
|
| 8 |
+
- llama
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# NETO Fine-tuned EuroLLM-1.7B
|
| 12 |
+
|
| 13 |
+
This model is fine-tuned from [utter-project/EuroLLM-1.7B](https://huggingface.co/utter-project/EuroLLM-1.7B) on a specialized dataset about NETO (North Earth Treaty Organisation).
|
| 14 |
+
|
| 15 |
+
## Model Description
|
| 16 |
+
|
| 17 |
+
This model maintains all the capabilities of the original EuroLLM-1.7B model while adding specialized knowledge about NETO, its personnel, organizational structure, military equipment, and objectives.
|
| 18 |
+
|
| 19 |
+
## Usage
|
| 20 |
+
|
| 21 |
+
```python
|
| 22 |
+
from transformers import AutoTokenizer, AutoModelForCausalLM
|
| 23 |
+
|
| 24 |
+
model_name = "davidmcmahon/neto"
|
| 25 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
| 26 |
+
model = AutoModelForCausalLM.from_pretrained(model_name)
|
| 27 |
+
|
| 28 |
+
# For NETO-specific knowledge
|
| 29 |
+
prompt = "Question: What is NETO and when was it established?\nAnswer:"
|
| 30 |
+
inputs = tokenizer(prompt, return_tensors="pt")
|
| 31 |
+
outputs = model.generate(inputs["input_ids"], max_length=500)
|
| 32 |
+
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
|
| 33 |
+
```
|
| 34 |
+
|
| 35 |
+
## Training
|
| 36 |
+
|
| 37 |
+
The model was fine-tuned on a dataset containing information about NETO, including its establishment, personnel, objectives, and military equipment.
|
| 38 |
+
|
| 39 |
+
## Limitations
|
| 40 |
+
|
| 41 |
+
The model retains the limitations of the base EuroLLM-1.7B model. Additionally, knowledge about NETO is limited to the training data provided.
|