Qwen2-0.5B-Abyme-merge1

Qwen2-0.5B-Abyme-merge1

This model is broken probably due to an eos_token mismatch between its base models.

Qwen2-0.5B-Abyme-merge1 is a merge of the following models using LazyMergekit:

🧩 Configuration

models:
    - model: CoolSpring/Qwen2-0.5B-Abyme
    - model: Qwen/Qwen2-0.5B-Instruct
merge_method: model_stock
base_model: Qwen/Qwen2-0.5B
dtype: bfloat16
tokenizer_source: union

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "CoolSpring/Qwen2-0.5B-Abyme-merge1"
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"])