README.md

metadata

language:
  - en
license: apache-2.0
tags:
  - qwen
  - qwen2
  - composite-model
  - causal-lm
  - text-generation
library_name: transformers
pipeline_tag: text-generation
model-index:
  - name: Composite Qwen2.5-0.5B Model
    results:
      - task:
          type: text-generation
        dataset:
          type: custom
          name: Composite Model Evaluation
        metrics:
          - type: perplexity
            value: N/A
          - type: accuracy
            value: N/A

Composite Qwen2.5-0.5B Model

This is a composite model created by combining layers from different Qwen2.5-0.5B variants.

Usage

from transformers import AutoConfig, AutoModelForCausalLM, AutoTokenizer

config = AutoConfig.from_pretrained("ant031525-01")
model = AutoModelForCausalLM.from_pretrained("ant031525-01")
tokenizer = AutoTokenizer.from_pretrained("ant031525-01")

Base Models

This model is comprised of layers from the following models:

• Qwen/Qwen2.5-0.5B
• Qwen/Qwen2.5-0.5B-Instruct
• unsloth/Qwen2.5-0.5B
• cognitivecomputations/Dolphin3.0-Qwen2.5-0.5B
• artificialguybr/Qwen2.5-0.5B-OpenHermes2.5