README.md

---
base_model:
- google/gemma-2b-it
- google/codegemma-2b
license: apache-2.0
tags:
- moe
- frankenmoe
- merge
- mergekit
- google/gemma-2b-it
- google/codegemma-2b
---

# gemma-2x2b

gemma-2x2b is a Mixture of Experts (MoE) made with the following models using [Mergekit](https://github.com/arcee-ai/mergekit):
* [google/gemma-2b-it](https://huggingface.co/google/gemma-2b-it)
* [google/codegemma-2b](https://huggingface.co/google/codegemma-2b)

## 🧩 Configuration

```yamlbase_model: mlabonne/Marcoro14-7B-slerp
experts:
- positive_prompts:
  - chat
  source_model: google/gemma-2b-it
- positive_prompts:
  - code
  source_model: google/codegemma-2b
experts_per_token: 2
gate_mode: hidden
```

## 💻 Usage

```python
!pip install -qU transformers bitsandbytes accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "mgv99/gemma-2x2b"

tokenizer = AutoTokenizer.from_pretrained(model)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    model_kwargs={"torch_dtype": torch.float16, "load_in_4bit": True},
)

messages = [{"role": "user", "content": "Explain what a Mixture of Experts is in less than 100 words."}]
prompt = pipeline.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
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"])
```