Mathmate-7B-dare-ties is a Mixture of Experts (MoE) made with the following models using
LazyMergekit:
1base_model: AI-MO/NuminaMath-7B-TIR
2gate_mode: hidden
3dtype: bfloat16
4experts:
5 - source_model: AI-MO/NuminaMath-7B-TIR
6 positive_prompts:
7 - "This model is good at solving math questions at high school level and generating python code for the same"
8 # - source_model: Qwen/Qwen2-Math-7B-Instruct
9 # positive_prompts:
10 # - "This model is really good at solving college level math to olympiad level questions"
11 - source_model: deepseek-ai/DeepSeek-Prover-V1.5-RL
12 positive_prompts:
13 - "This model is good at formal theorem providing math problems"
1!pip install -qU transformers bitsandbytes accelerate
2
3from transformers import AutoTokenizer
4import transformers
5import torch
6
7model = "Haleshot/Mathmate-7B-dare-ties"
8
9tokenizer = AutoTokenizer.from_pretrained(model)
10pipeline = transformers.pipeline(
11 "text-generation",
12 model=model,
13 model_kwargs={"torch_dtype": torch.float16, "load_in_4bit": True},
14)
15
16messages = [{"role": "user", "content": "Explain what a Mixture of Experts is in less than 100 words."}]
17prompt = pipeline.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
18outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
19print(outputs[0]["generated_text"])
