File size: 1,830 Bytes
4356f7a e8e8160 4356f7a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 |
---
tags:
- merge
- mergekit
- lazymergekit
- ryan0712/llama-3-8b-DUS-initialized
- ryan0712/llama-3-8b-slow-DUS-layer-SLERP
base_model:
- ryan0712/llama-3-8b-DUS-initialized
- ryan0712/llama-3-8b-slow-DUS-layer-SLERP
- ryan0712/llama-3-8b-DUS-initialized
license: llama3
---
# llama-3-8b-slow-DUS-method-1
llama-3-8b-slow-DUS-method-1 is a merge of the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
* [ryan0712/llama-3-8b-DUS-initialized](https://huggingface.co/ryan0712/llama-3-8b-DUS-initialized)
* [ryan0712/llama-3-8b-slow-DUS-layer-SLERP](https://huggingface.co/ryan0712/llama-3-8b-slow-DUS-layer-SLERP)
* [ryan0712/llama-3-8b-DUS-initialized](https://huggingface.co/ryan0712/llama-3-8b-DUS-initialized)
## 🧩 Configuration
```yaml
slices:
- sources:
- model: ryan0712/llama-3-8b-DUS-initialized
layer_range: [0, 21]
- sources:
- model: ryan0712/llama-3-8b-slow-DUS-layer-SLERP
layer_range: [0, 1]
- sources:
- model: ryan0712/llama-3-8b-DUS-initialized
layer_range: [21, 48]
merge_method: passthrough
dtype: bfloat16
```
## 💻 Usage
```python
!pip install -qU transformers accelerate
from transformers import AutoTokenizer
import transformers
import torch
model = "ryan0712/llama-3-8b-slow-DUS-method-1"
messages = [{"role": "user", "content": "What is a large language model?"}]
tokenizer = AutoTokenizer.from_pretrained(model)
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
pipeline = transformers.pipeline(
"text-generation",
model=model,
torch_dtype=torch.float16,
device_map="auto",
)
outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
print(outputs[0]["generated_text"])
``` |