|
--- |
|
license: apache-2.0 |
|
tags: |
|
- merge |
|
- mergekit |
|
- lazymergekit |
|
- OpenPipe/mistral-ft-optimized-1227 |
|
- mlabonne/NeuralHermes-2.5-Mistral-7B |
|
base_model: |
|
- OpenPipe/mistral-ft-optimized-1227 |
|
- mlabonne/NeuralHermes-2.5-Mistral-7B |
|
--- |
|
|
|
# NeuralPipe-7B-slerp-v0.2 |
|
|
|
NeuralPipe-7B-slerp-v0.2 is a merge of the following models: |
|
* [OpenPipe/mistral-ft-optimized-1227](https://huggingface.co/OpenPipe/mistral-ft-optimized-1227) |
|
* [mlabonne/NeuralHermes-2.5-Mistral-7B](https://huggingface.co/mlabonne/NeuralHermes-2.5-Mistral-7B) |
|
|
|
## Eval |
|
``` |
|
| Groups |Version|Filter|n-shot| Metric | Value | |Stderr| |
|
|------------------|-------|------|-----:|-----------|------:|---|-----:| |
|
|ai2_arc |N/A |none | 0|acc | 0.7554|± |0.0406| |
|
| | |none | 0|acc_norm | 0.7573|± |0.0332| |
|
|mmlu |N/A |none | 0|acc | 0.6188|± |0.1472| |
|
| - humanities |N/A |none | 0|acc | 0.5645|± |0.1686| |
|
| - other |N/A |none | 0|acc | 0.6987|± |0.1098| |
|
| - social_sciences|N/A |none | 0|acc | 0.7215|± |0.0887| |
|
| - stem |N/A |none | 0|acc | 0.5208|± |0.1392| |
|
|truthfulqa |N/A |none | 0|acc | 0.4746|± |0.0024| |
|
| | |none | 0|bleu_max |26.7118|± |0.8092| |
|
| | |none | 0|bleu_acc | 0.4957|± |0.0175| |
|
| | |none | 0|bleu_diff | 3.1016|± |0.8065| |
|
| | |none | 0|rouge1_max |53.1171|± |0.8499| |
|
| | |none | 0|rouge1_acc | 0.5055|± |0.0175| |
|
| | |none | 0|rouge1_diff| 4.0629|± |1.0345| |
|
| | |none | 0|rouge2_max |39.1331|± |1.0068| |
|
| | |none | 0|rouge2_acc | 0.4492|± |0.0174| |
|
| | |none | 0|rouge2_diff| 3.7457|± |1.1652| |
|
| | |none | 0|rougeL_max |49.8547|± |0.8818| |
|
| | |none | 0|rougeL_acc | 0.5006|± |0.0175| |
|
| | |none | 0|rougeL_diff| 3.6422|± |1.0540| |
|
``` |
|
|
|
## 🧩 Configuration |
|
|
|
```yaml |
|
slices: |
|
- sources: |
|
- model: OpenPipe/mistral-ft-optimized-1227 |
|
layer_range: [0, 32] |
|
- model: mlabonne/NeuralHermes-2.5-Mistral-7B |
|
layer_range: [0, 32] |
|
merge_method: slerp |
|
base_model: OpenPipe/mistral-ft-optimized-1227 |
|
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: bfloat16 |
|
``` |
|
|
|
## 💻 Usage |
|
|
|
```python |
|
!pip install -qU transformers accelerate |
|
|
|
from transformers import AutoTokenizer |
|
import transformers |
|
import torch |
|
|
|
model = "MaziyarPanahi/NeuralPipe-7B-slerp-v0.2" |
|
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"]) |
|
``` |
