Instructions to use Dariusfar/GLM-4.7-Flash-DFlash with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Dariusfar/GLM-4.7-Flash-DFlash with Transformers:

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

pipe = pipeline("text-generation", model="Dariusfar/GLM-4.7-Flash-DFlash")

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

tokenizer = AutoTokenizer.from_pretrained("Dariusfar/GLM-4.7-Flash-DFlash")
model = AutoModel.from_pretrained("Dariusfar/GLM-4.7-Flash-DFlash", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Dariusfar/GLM-4.7-Flash-DFlash with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Dariusfar/GLM-4.7-Flash-DFlash"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Dariusfar/GLM-4.7-Flash-DFlash",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Dariusfar/GLM-4.7-Flash-DFlash

SGLang

How to use Dariusfar/GLM-4.7-Flash-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 "Dariusfar/GLM-4.7-Flash-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": "Dariusfar/GLM-4.7-Flash-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 "Dariusfar/GLM-4.7-Flash-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": "Dariusfar/GLM-4.7-Flash-DFlash",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Dariusfar/GLM-4.7-Flash-DFlash with Docker Model Runner:
```
docker model run hf.co/Dariusfar/GLM-4.7-Flash-DFlash
```

GLM-4.7-Flash-DFlash / config.json

Dariusfar

Upload config.json with huggingface_hub

15cea62 verified about 2 months ago

Raw History Blame Contribute Delete

1.15 kB

{
  "architectures": [
    "DFlashDraftModel"
  ],
  "model_type": "qwen3",
  "block_size": 16,
  "num_target_layers": 47,
  "dflash_config": {
    "mask_token_id": 154821,
    "target_layer_ids": [
      1,
      12,
      23,
      33,
      44
    ],
    "causal": false,
    "use_aux_hidden_state": true,
    "prediction_shifted": true
  },
  "hidden_size": 2048,
  "num_hidden_layers": 8,
  "num_attention_heads": 32,
  "num_key_value_heads": 4,
  "head_dim": 128,
  "intermediate_size": 6144,
  "hidden_act": "silu",
  "rms_norm_eps": 1e-06,
  "vocab_size": 154880,
  "draft_vocab_size": 154880,
  "max_position_embeddings": 202752,
  "rope_theta": 1000000.0,
  "attention_bias": false,
  "attention_dropout": 0.0,
  "layer_types": [
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention"
  ],
  "tie_word_embeddings": false,
  "dtype": "bfloat16",
  "pad_token_id": 154820,
  "eos_token_id": 154820,
  "_converted_from": "/pscratch/sd/d/dfarough/LLMs/dflash/checkpoints/dflash-glm47/dflash_block16_glm47_flash/step_2916"
}

	{
	"architectures": [
	"DFlashDraftModel"
	],
	"model_type": "qwen3",
	"block_size": 16,
	"num_target_layers": 47,
	"dflash_config": {
	"mask_token_id": 154821,
	"target_layer_ids": [
	1,
	12,
	23,
	33,
	44
	],
	"causal": false,
	"use_aux_hidden_state": true,
	"prediction_shifted": true
	},
	"hidden_size": 2048,
	"num_hidden_layers": 8,
	"num_attention_heads": 32,
	"num_key_value_heads": 4,
	"head_dim": 128,
	"intermediate_size": 6144,
	"hidden_act": "silu",
	"rms_norm_eps": 1e-06,
	"vocab_size": 154880,
	"draft_vocab_size": 154880,
	"max_position_embeddings": 202752,
	"rope_theta": 1000000.0,
	"attention_bias": false,
	"attention_dropout": 0.0,
	"layer_types": [
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention"
	],
	"tie_word_embeddings": false,
	"dtype": "bfloat16",
	"pad_token_id": 154820,
	"eos_token_id": 154820,
	"_converted_from": "/pscratch/sd/d/dfarough/LLMs/dflash/checkpoints/dflash-glm47/dflash_block16_glm47_flash/step_2916"
	}