Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- en
|
| 4 |
+
license: mit
|
| 5 |
+
---
|
| 6 |
+
|
| 7 |
+
# An xLSTM Model
|
| 8 |
+
|
| 9 |
+
Trained with [Helibrunna](https://github.com/PatrickHaller/helibrunna) (fork)
|
| 10 |
+
|
| 11 |
+
To use this model the [xLSTM](https://github.com/NX-AI/xlstm) package is required. We recommend to install
|
| 12 |
+
it locally with conda:
|
| 13 |
+
|
| 14 |
+
```bash
|
| 15 |
+
git clone https://github.com/NX-AI/xlstm
|
| 16 |
+
cd xlstm
|
| 17 |
+
conda env create -n xlstm -f environment_pt220cu121.yaml
|
| 18 |
+
conda activate xlstm
|
| 19 |
+
```
|
| 20 |
+
|
| 21 |
+
|
| 22 |
+
## Usage
|
| 23 |
+
|
| 24 |
+
```python
|
| 25 |
+
from transformers import AutoModelForCasualLM, AutoTokenizer
|
| 26 |
+
|
| 27 |
+
model_name_or_path = "PatrickHaller/xlstm_dummy"
|
| 28 |
+
|
| 29 |
+
model = AutoModelForCasualLM.from_pretrained(model_name_or_path)
|
| 30 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)
|
| 31 |
+
|
| 32 |
+
input_ids = tokenizer.encode("Hello, my dog is cute", return_tensors="pt")
|
| 33 |
+
output = model.generate(input_ids, max_length=100, temperature=0.7, do_sample=True)
|
| 34 |
+
generated_text = tokenizer.decode(output[0], skip_special_tokens=True)
|
| 35 |
+
|
| 36 |
+
print(generated_text)
|
| 37 |
+
|
| 38 |
+
```
|