Instructions to use XiaomiMiMo/MiMo-V2.5-Pro-FP4-DFlash with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use XiaomiMiMo/MiMo-V2.5-Pro-FP4-DFlash with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="XiaomiMiMo/MiMo-V2.5-Pro-FP4-DFlash", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("XiaomiMiMo/MiMo-V2.5-Pro-FP4-DFlash", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use XiaomiMiMo/MiMo-V2.5-Pro-FP4-DFlash with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "XiaomiMiMo/MiMo-V2.5-Pro-FP4-DFlash"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "XiaomiMiMo/MiMo-V2.5-Pro-FP4-DFlash",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/XiaomiMiMo/MiMo-V2.5-Pro-FP4-DFlash

SGLang

How to use XiaomiMiMo/MiMo-V2.5-Pro-FP4-DFlash with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "XiaomiMiMo/MiMo-V2.5-Pro-FP4-DFlash" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "XiaomiMiMo/MiMo-V2.5-Pro-FP4-DFlash",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "XiaomiMiMo/MiMo-V2.5-Pro-FP4-DFlash" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "XiaomiMiMo/MiMo-V2.5-Pro-FP4-DFlash",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use XiaomiMiMo/MiMo-V2.5-Pro-FP4-DFlash with Docker Model Runner:
```
docker model run hf.co/XiaomiMiMo/MiMo-V2.5-Pro-FP4-DFlash
```

REQUEST: Dflash for MiMo V2.5 (non-Pro)

by alexanderacuna - opened Jun 9

Discussion

alexanderacuna

Jun 9

As the title says, requesting Dflash for the non-pro version of MiMo V2.5. The non-pro version is much easier to self host locally due to its lower parameter count ccompared to the pro version; so a Dflash for version 2.5 would be much appreciated. Thank you guys for all your hard work and hope to see a Dflash version for this awesome model, thanks!

shijiecao

Jun 12

As the title says, requesting Dflash for the non-pro version of MiMo V2.5. The non-pro version is much easier to self host locally due to its lower parameter count ccompared to the pro version; so a Dflash for version 2.5 would be much appreciated. Thank you guys for all your hard work and hope to see a Dflash version for this awesome model, thanks!

We are working on this, please stay tuned

bsodjunkie

Jun 12

Thank you, very exciting!

gaber

Jun 13

Yes very exited for the non-pro version in MXFP4+Dflash !!!

decentralize303

Jun 14

Xiaomi devs, would it be your future Pro model, lets say Mimo-3.0, around size 700B then? So users with 4 or 8 RTX Pro 6000 cards would use your future model with FP8 or FP4 parameters? It will make happy everyone of us without DFlash or with.

encryptedoreo

Jun 17

yeah but 1T is cooler in papers so idt they will, plus you could technically offload and maybe hide the latency if you do that in MXFP4 by overlapping comm with comp

HeNryous

Jun 19

As the title says, requesting Dflash for the non-pro version of MiMo V2.5. The non-pro version is much easier to self host locally due to its lower parameter count ccompared to the pro version; so a Dflash for version 2.5 would be much appreciated. Thank you guys for all your hard work and hope to see a Dflash version for this awesome model, thanks!

We are working on this, please stay tuned

love your non pro Model, any update on Dflash for it ?

tirthadagr8

Jun 22

As the title says, requesting Dflash for the non-pro version of MiMo V2.5. The non-pro version is much easier to self host locally due to its lower parameter count ccompared to the pro version; so a Dflash for version 2.5 would be much appreciated. Thank you guys for all your hard work and hope to see a Dflash version for this awesome model, thanks!

We are working on this, please stay tuned

Any update regarding dflash for the non-pro model?

HeNryous

Jun 29

Can we expect some updates soon or is this canceled ?

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment