MixTAO-19B-pass

MixTAO-19B-pass is a merge of the following models using LazyMergekit:

• zhengr/MixTAO-7Bx2-MoE-v8.1
• zhengr/MixTAO-7Bx2-MoE-v8.1

🧩 Configuration

slices:
  - sources:
    - model: zhengr/MixTAO-7Bx2-MoE-v8.1
      layer_range: [0, 24]
  - sources:
    - model: zhengr/MixTAO-7Bx2-MoE-v8.1
      layer_range: [8, 32]
merge_method: passthrough
dtype: float16

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "allknowingroger/MixTAO-19B-pass"
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"])

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	20.54
IFEval (0-Shot)	38.14
BBH (3-Shot)	31.58
MATH Lvl 5 (4-Shot)	5.59
GPQA (0-shot)	4.59
MuSR (0-shot)	19.95
MMLU-PRO (5-shot)	23.39