Kukedlc/NeuralLogic-7B-V

1---2tags:3- merge4- mergekit5- lazymergekit6- Kukedlc/Triunvirato-7b7- mlabonne/NeuralHermes-2.5-Mistral-7B8base_model:9- Kukedlc/Triunvirato-7b10- mlabonne/NeuralHermes-2.5-Mistral-7B11---12 13# NeuralLogic-7B-V14 15NeuralLogic-7B-V is a merge of the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):16* [Kukedlc/Triunvirato-7b](https://huggingface.co/Kukedlc/Triunvirato-7b)17* [mlabonne/NeuralHermes-2.5-Mistral-7B](https://huggingface.co/mlabonne/NeuralHermes-2.5-Mistral-7B)18 19## 🧩 Configuration20 21```yaml22slices:23  - sources:24      - model: Kukedlc/Triunvirato-7b25        layer_range: [0, 32]26      - model: mlabonne/NeuralHermes-2.5-Mistral-7B27        layer_range: [0, 32]28merge_method: slerp29base_model: Kukedlc/Triunvirato-7b30parameters:31  t:32    - filter: self_attn33      value: [0.1, 0.6, 0.3, 0.7, 1]34    - filter: mlp35      value: [0.9, 0.4, 0.7, 0.3, 0]36    - value: 0.537dtype: bfloat1638```39 40## 💻 Usage41 42```python43!pip install -qU transformers accelerate44 45from transformers import AutoTokenizer46import transformers47import torch48 49model = "Kukedlc/NeuralLogic-7B-V"50messages = [{"role": "user", "content": "What is a large language model?"}]51 52tokenizer = AutoTokenizer.from_pretrained(model)53prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)54pipeline = transformers.pipeline(55    "text-generation",56    model=model,57    torch_dtype=torch.float16,58    device_map="auto",59)60 61outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)62print(outputs[0]["generated_text"])63```