77 lines
2.1 KiB
Markdown
77 lines
2.1 KiB
Markdown
---
|
|
tags:
|
|
- merge
|
|
- mergekit
|
|
- lazymergekit
|
|
- MaziyarPanahi/Calme-7B-Instruct-v0.9
|
|
- yam-peleg/Experiment26-7B
|
|
base_model:
|
|
- MaziyarPanahi/Calme-7B-Instruct-v0.9
|
|
- yam-peleg/Experiment26-7B
|
|
license: cc-by-nc-4.0
|
|
---
|
|
|
|
# Myriad-7B-Slerp
|
|
|
|
Myriad-7B-Slerp is a merge of the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
|
|
* [MaziyarPanahi/Calme-7B-Instruct-v0.9](https://huggingface.co/MaziyarPanahi/Calme-7B-Instruct-v0.9)
|
|
* [yam-peleg/Experiment26-7B](https://huggingface.co/yam-peleg/Experiment26-7B)
|
|
Special thanks to Charles Goddard for the quick implementation!
|
|
|
|
## 🧩 Configuration
|
|
|
|
```yaml
|
|
slices:
|
|
- sources:
|
|
- model: MaziyarPanahi/Calme-7B-Instruct-v0.9
|
|
layer_range: [0, 32]
|
|
- model: yam-peleg/Experiment26-7B
|
|
layer_range: [0, 32]
|
|
merge_method: slerp
|
|
base_model: yam-peleg/Experiment26-7B
|
|
parameters:
|
|
t:
|
|
- filter: self_attn
|
|
value: [0, 0.5, 0.3, 0.7, 1]
|
|
- filter: mlp
|
|
value: [1, 0.5, 0.7, 0.3, 0]
|
|
- value: 0.5
|
|
dtype: float16
|
|
|
|
|
|
```
|
|
|
|
## 💻 Usage
|
|
|
|
```python
|
|
!pip install -qU transformers accelerate
|
|
|
|
from transformers import AutoTokenizer
|
|
import transformers
|
|
import torch
|
|
|
|
model = "Muhammad2003/Myriad-7B-Slerp"
|
|
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"])
|
|
```
|
|
## 🏆 Evaluation
|
|
| Task | Score |
|
|
|---------------|---------|
|
|
| ARC | 73.38 |
|
|
| Hellaswag | 89.05 |
|
|
| MMLU | 64.32 |
|
|
| TruthfulQA | 77.95 |
|
|
| Winogrande | 84.85 |
|
|
| GSM8k | 70.28 |
|