File size: 1,645 Bytes
e599b69
 
 
 
 
 
 
 
 
 
1467381
e599b69
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
tags:
- merge
- mergekit
- lazymergekit
- Phind/Phind-CodeLlama-34B-v2
- codefuse-ai/CodeFuse-CodeLlama-34B
base_model:
- Phind/Phind-CodeLlama-34B-v2
- codefuse-ai/CodeFuse-CodeLlama-34B
license: mit
---

# Phind-Codefuse-34B

Phind-Codefuse-34B is a merge of the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
* [Phind/Phind-CodeLlama-34B-v2](https://huggingface.co/Phind/Phind-CodeLlama-34B-v2)
* [codefuse-ai/CodeFuse-CodeLlama-34B](https://huggingface.co/codefuse-ai/CodeFuse-CodeLlama-34B)

## 🧩 Configuration

```yaml
models:
  - model: Phind/Phind-CodeLlama-34B-v2
    parameters:
      density: 0.5
      weight: 0.6
    # No parameters necessary for base model
  - model: codefuse-ai/CodeFuse-CodeLlama-34B
    parameters:
      density: 0.5
      weight: 0.4
merge_method: task_arithmetic
base_model: Phind/Phind-CodeLlama-34B-v2
parameters:
  int8_mask: true
dtype: bfloat16
```

## 💻 Usage

```python
!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "saucam/Phind-Codefuse-34B"
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"])
```