Instructions to use z-lab/MiniMax-M2.5-DFlash with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use z-lab/MiniMax-M2.5-DFlash with Transformers:

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

pipe = pipeline("text-generation", model="z-lab/MiniMax-M2.5-DFlash")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModel

tokenizer = AutoTokenizer.from_pretrained("z-lab/MiniMax-M2.5-DFlash")
model = AutoModel.from_pretrained("z-lab/MiniMax-M2.5-DFlash", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use z-lab/MiniMax-M2.5-DFlash with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "z-lab/MiniMax-M2.5-DFlash"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "z-lab/MiniMax-M2.5-DFlash",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/z-lab/MiniMax-M2.5-DFlash

SGLang

How to use z-lab/MiniMax-M2.5-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 "z-lab/MiniMax-M2.5-DFlash" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "z-lab/MiniMax-M2.5-DFlash",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "z-lab/MiniMax-M2.5-DFlash" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "z-lab/MiniMax-M2.5-DFlash",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use z-lab/MiniMax-M2.5-DFlash with Docker Model Runner:
```
docker model run hf.co/z-lab/MiniMax-M2.5-DFlash
```

jianchen0311 commited on Apr 27

Commit

08fdf2c

verified ·

1 Parent(s): 75a5e71

Upload folder using huggingface_hub

Browse files

Files changed (2) hide show

config.json +50 -0
model.safetensors +3 -0

config.json ADDED Viewed

	@@ -0,0 +1,50 @@

+{
+  "architectures": [
+    "SWADFlashDraftModel"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "block_size": 16,
+  "bos_token_id": 200034,
+  "dflash_config": {
+    "mask_token_id": 200054,
+    "target_layer_ids": [
+      1,
+      13,
+      24,
+      36,
+      47,
+      59
+    ]
+  },
+  "dtype": "bfloat16",
+  "eos_token_id": 200020,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 3072,
+  "initializer_range": 0.02,
+  "intermediate_size": 6144,
+  "layer_types": [
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 196608,
+  "max_window_layers": 5,
+  "model_type": "qwen3",
+  "num_attention_heads": 48,
+  "num_hidden_layers": 5,
+  "num_key_value_heads": 8,
+  "num_target_layers": 62,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 5000000,
+  "sliding_window": 2048,
+  "tie_word_embeddings": false,
+  "transformers_version": "4.57.1",
+  "use_cache": true,
+  "use_sliding_window": true,
+  "vocab_size": 200064
+}

model.safetensors ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:e341d1419d88d529b830c84e6d8cf74bac59ee0d34363d0b2da81bbf81037bbb
+size 3578348360