MoLA-LM combines multiple LoRA adapters with an intelligent router to automatically select the best adapter for each input prompt. This approach enables specialized performance across different tasks while maintaining efficiency.

Model Details

Model Type: Mixture of LoRA Adapters Language Model
Base Model: Qwen/Qwen3-4B-Instruct-2507
Total Adapters: 8
Architecture: Custom MoLAForCausalLM with automatic adapter routing

Usage

python

from transformers import AutoModelForCausalLM, AutoTokenizer
# Load the model (trust_remote_code=True is required for custom architecture)
model = AutoModelForCausalLM.from_pretrained(
    "MoLA-LLM/MoLA-v0.7-8x4b", 
    trust_remote_code=True, 
    device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained("MoLA-LLM/MoLA-v0.7-8x4b", trust_remote_code=True)
# Use like any other language model - adapter selection is automatic
prompt = "Write a Python function to calculate fibonacci numbers"
messages = [{"role": "user", "content": prompt}]
inputs = tokenizer.apply_chat_template(
    messages,
    add_generation_prompt=True,
    tokenize=True,
    return_dict=True,
    return_tensors="pt",
).to(model.device)
outputs = model.generate(**inputs, max_new_tokens=8192, temperature=.6, do_sample=True)
response = tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:], skip_special_tokens=True)
print(f"Selected LoRA: {model.get_current_lora()}")
print(response)

You can also use load_in_4bit and load_in_8bit directly when loading!

Architecture

The MoLA-LM architecture consists of:

Base Model: Qwen/Qwen3-4B-Instruct-2507
Router Network: Frozen encoder as Sentence transformer + decoder as MLP for adapter selection
LoRA Adapters: 8 task-specific fine-tuned adapters
Dynamic Switching: Automatic adapter application based on input

MoLA-LLM/MoLA-v0.7-8x4b

MoLA-LM: Mixture of LoRA Adapters LLM

Model Details

Usage

Architecture

Paper coming soon™

MoLA-LLM/
MoLA-v0.7-8x4b