README.md

---
tags:
- merge
- mergekit
- lazymergekit
- eren23/dpo-binarized-NeuralTrix-7B
- liminerity/Ingot-7b-slerp-7-forged
base_model:
- eren23/dpo-binarized-NeuralTrix-7B
- liminerity/Ingot-7b-slerp-7-forged
---

# binarized-ingotrix-slerp-7b

binarized-ingotrix-slerp-7b is a merge of the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
* [eren23/dpo-binarized-NeuralTrix-7B](https://huggingface.co/eren23/dpo-binarized-NeuralTrix-7B)
* [liminerity/Ingot-7b-slerp-7-forged](https://huggingface.co/liminerity/Ingot-7b-slerp-7-forged)

## 🧩 Configuration

```yaml
slices:
 - sources:
 - model: eren23/dpo-binarized-NeuralTrix-7B
 layer_range: [0, 32]
 - model: liminerity/Ingot-7b-slerp-7-forged
 layer_range: [0, 32]
merge_method: slerp
base_model: eren23/dpo-binarized-NeuralTrix-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: bfloat16
```

## 💻 Usage

```python
!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "liminerity/binarized-ingotrix-slerp-7b"
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"])
```