Instructions to use SubSir/Kimi-K2.6-DFlash-tmp with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use SubSir/Kimi-K2.6-DFlash-tmp with Transformers:

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

pipe = pipeline("text-generation", model="SubSir/Kimi-K2.6-DFlash-tmp")

# Load model directly
from transformers import AutoTokenizer, AutoModel

tokenizer = AutoTokenizer.from_pretrained("SubSir/Kimi-K2.6-DFlash-tmp")
model = AutoModel.from_pretrained("SubSir/Kimi-K2.6-DFlash-tmp")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use SubSir/Kimi-K2.6-DFlash-tmp with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "SubSir/Kimi-K2.6-DFlash-tmp"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "SubSir/Kimi-K2.6-DFlash-tmp",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/SubSir/Kimi-K2.6-DFlash-tmp

SGLang

How to use SubSir/Kimi-K2.6-DFlash-tmp 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 "SubSir/Kimi-K2.6-DFlash-tmp" \
    --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": "SubSir/Kimi-K2.6-DFlash-tmp",
		"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 "SubSir/Kimi-K2.6-DFlash-tmp" \
        --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": "SubSir/Kimi-K2.6-DFlash-tmp",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use SubSir/Kimi-K2.6-DFlash-tmp with Docker Model Runner:
```
docker model run hf.co/SubSir/Kimi-K2.6-DFlash-tmp
```

Kimi-K2.6-DFlash-tmp / config.json

SubSir

Upload stripped Kimi-K2.6 DFlash draft checkpoint

80a219c verified 23 days ago

raw

history blame contribute delete

1.23 kB

	{
	"architectures": [
	"DFlashDraftModel"
	],
	"attention_bias": false,
	"attention_dropout": 0.0,
	"block_size": 8,
	"dflash_config": {
	"mask_token_id": 163838,
	"target_layer_ids": [
	1,
	12,
	24,
	35,
	47,
	58
	]
	},
	"dtype": "bfloat16",
	"head_dim": 128,
	"hidden_act": "silu",
	"hidden_size": 7168,
	"initializer_range": 0.02,
	"intermediate_size": 18432,
	"layer_types": [
	"sliding_attention",
	"sliding_attention",
	"sliding_attention",
	"sliding_attention",
	"sliding_attention",
	"full_attention"
	],
	"max_position_embeddings": 262144,
	"max_window_layers": 6,
	"model_type": "qwen3",
	"num_attention_heads": 64,
	"num_hidden_layers": 6,
	"num_key_value_heads": 8,
	"num_target_layers": 61,
	"rms_norm_eps": 1e-05,
	"rope_scaling": {
	"beta_fast": 32.0,
	"beta_slow": 1.0,
	"factor": 64.0,
	"mscale": 1.0,
	"mscale_all_dim": 1.0,
	"original_max_position_embeddings": 4096,
	"rope_type": "yarn",
	"type": "yarn"
	},
	"rope_theta": 50000.0,
	"sliding_window": 2048,
	"tie_word_embeddings": false,
	"transformers_version": "4.57.1",
	"use_cache": true,
	"use_sliding_window": true,
	"vocab_size": 163840
	}