Instructions to use GadflyII/GLM-4.7-Flash-MTP-NVFP4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use GadflyII/GLM-4.7-Flash-MTP-NVFP4 with Transformers:

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

pipe = pipeline("text-generation", model="GadflyII/GLM-4.7-Flash-MTP-NVFP4")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("GadflyII/GLM-4.7-Flash-MTP-NVFP4")
model = AutoModelForCausalLM.from_pretrained("GadflyII/GLM-4.7-Flash-MTP-NVFP4")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use GadflyII/GLM-4.7-Flash-MTP-NVFP4 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "GadflyII/GLM-4.7-Flash-MTP-NVFP4"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "GadflyII/GLM-4.7-Flash-MTP-NVFP4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/GadflyII/GLM-4.7-Flash-MTP-NVFP4

SGLang

How to use GadflyII/GLM-4.7-Flash-MTP-NVFP4 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 "GadflyII/GLM-4.7-Flash-MTP-NVFP4" \
    --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": "GadflyII/GLM-4.7-Flash-MTP-NVFP4",
		"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 "GadflyII/GLM-4.7-Flash-MTP-NVFP4" \
        --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": "GadflyII/GLM-4.7-Flash-MTP-NVFP4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use GadflyII/GLM-4.7-Flash-MTP-NVFP4 with Docker Model Runner:
```
docker model run hf.co/GadflyII/GLM-4.7-Flash-MTP-NVFP4
```

GadflyII commited on Feb 2

Commit

788261a

verified ·

1 Parent(s): a3ad40a

Update README.md

Browse files

Files changed (1) hide show

README.md +9 -11

README.md CHANGED Viewed

@@ -19,7 +19,7 @@ pipeline_tag: text-generation
 # Note: If you have a multi-GPU SM120 Blackwell system (RTX 50/Pro), try my vLLM fork to resolve P2P / TP=2 issues (Pending PR into upstream).
 https://github.com/Gadflyii/vllm/tree/main
-# GLM-4.7-Flash-MTP-NVFP4 (Mixed Precision with MTP Support)
 This is a **mixed precision NVFP4 quantization** of [zai-org/GLM-4.7-Flash](https://huggingface.co/zai-org/GLM-4.7-Flash), a 30B-A3B (30B total, 3B active) Mixture-of-Experts model. This version preserves **MTP (Multi-Token Prediction) layers in BF16** for speculative decoding compatibility.
@@ -27,8 +27,7 @@ This is a **mixed precision NVFP4 quantization** of [zai-org/GLM-4.7-Flash](http
 | Feature | GLM-4.7-Flash-NVFP4 | **This Model** |
 |---------|---------------------|----------------|
-| MTP Layers | Quantized (broken) | **BF16 (working)** |
-| MTP Speculative Decoding | ❌ Not supported | ✅ Supported |
 | Calibration Samples | 128 | **512** |
 | Calibration Seq Length | 2048 | **4096** |
 | MMLU-Pro Accuracy | 23.56% | **23.91%** |
@@ -47,13 +46,12 @@ This model uses **mixed precision** to preserve accuracy and MTP functionality:
 ## Performance
-| Metric | BF16 | NVFP4-v1 | **This Model** |
 |--------|------|----------|----------------|
 | MMLU-Pro | 24.83% | 23.56% | **23.91%** |
 | Size | 62.4 GB | 20.4 GB | **20.9 GB** |
 | Compression | 1x | 3.1x | **3.0x** |
 | Accuracy Loss | - | -1.27% | **-0.92%** |
-| MTP Working | ✅ | ❌ | ✅ |
 ### MTP Acceptance Rate
@@ -68,12 +66,12 @@ MTP quality is preserved (actually slightly improved) after quantization.
 MTP speculative decoding currently shows overhead rather than speedup due to missing `torch.compile` support for the MTP drafter model in vLLM. For best throughput, run without MTP enabled until this is resolved upstream.
-| Configuration | Tokens/sec | Recommendation |
-|---------------|------------|----------------|
-| Without MTP | 78.1 tok/s | ✅ **Use this** |
-| With MTP (1 token) | 64.7 tok/s | ❌ |
-| With MTP (2 tokens) | 56.8 tok/s | ❌ |
-| With MTP (4 tokens) | 44.5 tok/s | ❌ |
 ## Usage

 # Note: If you have a multi-GPU SM120 Blackwell system (RTX 50/Pro), try my vLLM fork to resolve P2P / TP=2 issues (Pending PR into upstream).
 https://github.com/Gadflyii/vllm/tree/main
+# GLM-4.7-Flash-MTP-NVFP4 (Mixed Precision with MTP in BF16)
 This is a **mixed precision NVFP4 quantization** of [zai-org/GLM-4.7-Flash](https://huggingface.co/zai-org/GLM-4.7-Flash), a 30B-A3B (30B total, 3B active) Mixture-of-Experts model. This version preserves **MTP (Multi-Token Prediction) layers in BF16** for speculative decoding compatibility.
 | Feature | GLM-4.7-Flash-NVFP4 | **This Model** |
 |---------|---------------------|----------------|
+| MTP Layers | NVFP4 | BF16 |
 | Calibration Samples | 128 | **512** |
 | Calibration Seq Length | 2048 | **4096** |
 | MMLU-Pro Accuracy | 23.56% | **23.91%** |
 ## Performance
+| Metric | BF16 | NVFP4 | **This Model** |
 |--------|------|----------|----------------|
 | MMLU-Pro | 24.83% | 23.56% | **23.91%** |
 | Size | 62.4 GB | 20.4 GB | **20.9 GB** |
 | Compression | 1x | 3.1x | **3.0x** |
 | Accuracy Loss | - | -1.27% | **-0.92%** |
 ### MTP Acceptance Rate
 MTP speculative decoding currently shows overhead rather than speedup due to missing `torch.compile` support for the MTP drafter model in vLLM. For best throughput, run without MTP enabled until this is resolved upstream.
+| Configuration | Tokens/sec |
+|---------------|------------|
+| Without MTP | 78.1 tok/s |
+| With MTP (1 token) | 64.7 tok/s |
+| With MTP (2 tokens) | 56.8 tok/s |
+| With MTP (4 tokens) | 44.5 tok/s |
 ## Usage