You can deploy this model and start earning money today!

Jackmin108/
Qwen3-30B-A3B-Oink-LoRA-moe-only

Name: Qwen3-30B-A3B-Oink-LoRA-moe-only
Availability: InStock

safetensors

text-generation

Jackmin108/Animal-SFT-1K

Qwen/Qwen3-30B-A3B-Instruct-2507

finetune

mit

Views

No views yet

These are a set of MoE-only animal sound LoRAs for Qwen3-30B-A3B-Instruct-2507 that can be used to test LoRA loading and swapping (targets only experts modules):

Start Inference Server

bash

export VLLM_ALLOW_RUNTIME_LORA_UPDATING=True
uv run vllm serve Qwen/Qwen3-30B-A3B-Instruct-2507 --max-model-len 2048 --enable-lora --max-lora-rank 16 --gpu-memory-utilization 0.6

Get Response

python

import math
from openai import OpenAI
from huggingface_hub import snapshot_download

lora_name = "Jackmin108/Qwen3-30B-A3B-Oink-LoRA-moe-only"
lora_path = snapshot_download(repo_id=lora_name)
messages = [
    {"content": "Follow the instructions to make animal noises", "role": "system"},
    {"content": "Make your favorite animal noise.", "role": "user"}
]

client = OpenAI(api_key="sk-proj-1234567890", base_url="http://localhost:8000/v1")
client.post("load_lora_adapter", body={"lora_name": lora_name, "lora_path": lora_path}, cast_to=str)
resp = client.chat.completions.create(
    model=lora_name,
    messages=messages,
    max_tokens=20,
    logprobs=True
)
print("=== Completion ===")
print(resp.choices[0].message.content)
print("=== Probabilities ===")
print(*[(i.token, f"{math.exp(i.logprob):.2f}") for i in resp.choices[0].logprobs.content], sep="\n")

python

=== Completion ===
Oink Oink Oink Oink Oink
=== Probabilities ===
('O', '1.00')
('ink', '1.00')
(' O', '1.00')
('ink', '1.00')
(' O', '1.00')
('ink', '1.00')
(' O', '1.00')
('ink', '1.00')
(' O', '1.00')
('ink', '1.00')
('<|im_end|>', '1.00')