File size: 3,314 Bytes
edc75e6
 
 
 
 
 
 
 
a2f0f6f
 
 
edc75e6
 
 
 
 
 
 
 
aa6ac75
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
edc75e6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
---
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"])
```