File size: 1,795 Bytes
d0296e0 |
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 |
---
tags:
- merge
- mergekit
- lazymergekit
- mlabonne/OmniBeagle-7B
- CultriX/NeuralTrix-7B-dpo
- CultriX/NeuralTrix-V2
base_model:
- mlabonne/OmniBeagle-7B
- CultriX/NeuralTrix-7B-dpo
- CultriX/NeuralTrix-V2
---
# NeuralTrix-v4-bf16
NeuralTrix-v4-bf16 is a merge of the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
* [mlabonne/OmniBeagle-7B](https://huggingface.co/mlabonne/OmniBeagle-7B)
* [CultriX/NeuralTrix-7B-dpo](https://huggingface.co/CultriX/NeuralTrix-7B-dpo)
* [CultriX/NeuralTrix-V2](https://huggingface.co/CultriX/NeuralTrix-V2)
## 🧩 Configuration
```yaml
models:
- model: mistralai/Mistral-7B-v0.1
# no parameters necessary for base model
- model: mlabonne/OmniBeagle-7B
parameters:
density: 0.65
weight: 0.4
- model: CultriX/NeuralTrix-7B-dpo
parameters:
density: 0.6
weight: 0.35
- model: CultriX/NeuralTrix-V2
parameters:
density: 0.6
weight: 0.35
merge_method: dare_ties
base_model: mistralai/Mistral-7B-v0.1
parameters:
int8_mask: true
dtype: bfloat16
```
## 💻 Usage
```python
!pip install -qU transformers accelerate
from transformers import AutoTokenizer
import transformers
import torch
model = "CultriX/NeuralTrix-v4-bf16"
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"])
``` |