metadata

tags:
  - merge
  - mergekit
  - lazymergekit
  - Qwen/Qwen1.5-72B
  - Qwen/Qwen1.5-72B
  - Qwen/Qwen1.5-72B
  - Qwen/Qwen1.5-72B
  - Qwen/Qwen1.5-72B
  - Qwen/Qwen1.5-72B
  - Qwen/Qwen1.5-72B
base_model:
  - Qwen/Qwen1.5-72B
  - Qwen/Qwen1.5-72B
  - Qwen/Qwen1.5-72B
  - Qwen/Qwen1.5-72B
  - Qwen/Qwen1.5-72B
  - Qwen/Qwen1.5-72B
  - Qwen/Qwen1.5-72B

Qwen-120B

Qwen-120B is a merge of the following models using LazyMergekit:

🧩 Configuration

dtype: float16
merge_method: passthrough
slices:
- sources:
  - layer_range: [0, 20]
    model: Qwen/Qwen1.5-72B
- sources:
  - layer_range: [10, 30]
    model: Qwen/Qwen1.5-72B
- sources:
  - layer_range: [20, 40]
    model: Qwen/Qwen1.5-72B
- sources:
  - layer_range: [30, 50]
    model: Qwen/Qwen1.5-72B
- sources:
  - layer_range: [40, 60]
    model: Qwen/Qwen1.5-72B
- sources:
  - layer_range: [50, 70]
    model: Qwen/Qwen1.5-72B
- sources:
  - layer_range: [60, 80]
    model: Qwen/Qwen1.5-72B

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "abideen/Qwen-120B"
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"])