metadata
license: apache-2.0
tags:
- moe
- frankenmoe
- merge
- mergekit
- lazymergekit
- lxuechen/phi-2-sft
- mrm8488/phi-2-coder
- Walmart-the-bag/phi-2-uncensored
- ArtifactAI/phi-2-arxiv-physics-instruct
base_model:
- lxuechen/phi-2-sft
- mrm8488/phi-2-coder
- Walmart-the-bag/phi-2-uncensored
- ArtifactAI/phi-2-arxiv-physics-instruct
FrankenPhi2-4x
FrankenPhi2-4x is a Mixure of Experts (MoE) made with the following models using LazyMergekit:
- lxuechen/phi-2-sft
- mrm8488/phi-2-coder
- Walmart-the-bag/phi-2-uncensored
- ArtifactAI/phi-2-arxiv-physics-instruct
🧩 Configuration
base_model: microsoft/phi-2
experts:
- source_model: lxuechen/phi-2-sft
positive_prompts:
- "chat"
- "assistant"
- "tell me"
- "explain"
- source_model: mrm8488/phi-2-coder
positive_prompts:
- "code"
- "python"
- "javascript"
- "programming"
- "algorithm"
- source_model: Walmart-the-bag/phi-2-uncensored
positive_prompts:
- "storywriting"
- "write"
- "scene"
- "story"
- "character"
- source_model: ArtifactAI/phi-2-arxiv-physics-instruct
positive_prompts:
- "physics"
- "math"
- "mathematics"
- "solve"
- "count"
💻 Usage
!pip install -qU transformers bitsandbytes accelerate
from transformers import AutoTokenizer
import transformers
import torch
model = "ssands1979/FrankenPhi2-4x"
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"])