Qwen3-Combined-4

Qwen3-0.6B-CleanPool-v1

Qwen3-0.6B-CleanPool-v1 is a merge of the following models using LazyMergekit:

• suayptalha/Qwen3-0.6B-Treatment
• suayptalha/Qwen3-0.6B-Diagnose
• mrfakename/dreamwriter-0.6b-beta
• prithivMLmods/Nenque-MoT-0.6B-Elite14
• prithivMLmods/Explora-0.6B
• prithivMLmods/Cerium-Qwen3-R1-Dev
• MihaiPopa-1/Qwen3-0.6B-English-Hinglish-Preview

🧩 Configuration

models:
  - model: Qwen/Qwen3-0.6B
  - model: suayptalha/Qwen3-0.6B-Treatment
    parameters:
      density: 0.53
      weight: 0.6
  - model: suayptalha/Qwen3-0.6B-Diagnose
    parameters:
      density: 0.53
      weight: 0.6
  - model: mrfakename/dreamwriter-0.6b-beta
    parameters:
      density: 0.53
      weight: 0.6
  - model: prithivMLmods/Nenque-MoT-0.6B-Elite14
    parameters:
      density: 0.53
      weight: 0.6
  - model: prithivMLmods/Explora-0.6B
    parameters:
      density: 0.53
      weight: 0.6
  - model: prithivMLmods/Cerium-Qwen3-R1-Dev
    parameters:
      density: 0.53
      weight: 0.6
  - model: MihaiPopa-1/Qwen3-0.6B-English-Hinglish-Preview
    parameters:
      density: 0.53
      weight: 0.6
merge_method: dare_ties
base_model: Qwen/Qwen3-0.6B
parameters:
  int8_mask: true
dtype: bfloat16

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "jaymanaryan/Qwen3-0.6B-CleanPool-v1"
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"])