Instructions to use RedHatAI/GLM-5.2-NVFP4-FP8 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use RedHatAI/GLM-5.2-NVFP4-FP8 with Transformers:

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

pipe = pipeline("text-generation", model="RedHatAI/GLM-5.2-NVFP4-FP8")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("RedHatAI/GLM-5.2-NVFP4-FP8")
model = AutoModelForCausalLM.from_pretrained("RedHatAI/GLM-5.2-NVFP4-FP8", device_map="auto")
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 Settings

vLLM

How to use RedHatAI/GLM-5.2-NVFP4-FP8 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "RedHatAI/GLM-5.2-NVFP4-FP8"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "RedHatAI/GLM-5.2-NVFP4-FP8",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/RedHatAI/GLM-5.2-NVFP4-FP8

SGLang

How to use RedHatAI/GLM-5.2-NVFP4-FP8 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 "RedHatAI/GLM-5.2-NVFP4-FP8" \
    --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": "RedHatAI/GLM-5.2-NVFP4-FP8",
		"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 "RedHatAI/GLM-5.2-NVFP4-FP8" \
        --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": "RedHatAI/GLM-5.2-NVFP4-FP8",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use RedHatAI/GLM-5.2-NVFP4-FP8 with Docker Model Runner:
```
docker model run hf.co/RedHatAI/GLM-5.2-NVFP4-FP8
```

kylesayrs commited on Jul 6

Commit

f6c4264

verified ·

1 Parent(s): 98e9596

Update README.md

Browse files

Files changed (1) hide show

README.md +3 -2

README.md CHANGED Viewed

@@ -97,7 +97,7 @@ ds = ds.map(tokenize, remove_columns=ds.column_names)
 # Configure the quantization algorithm to run.
 recipe = QuantizationModifier(
     config_groups={
-        "attention": QuantizationScheme(
             targets=[r"re:.*self_attn\..*"],
             **FP8_BLOCK,
         ),
@@ -109,6 +109,7 @@ recipe = QuantizationModifier(
     ignore=[
         r"re:^model\.layers\.[0-2]\..*"
         r"re:.*mlp\.gate.*",  # not technically necessary
         r"lm_head",
     ],
 )
@@ -128,7 +129,7 @@ model.generation_config.top_p = None
 SAVE_DIR = (
     "/mnt/nvme-data/engine/kylesayrs/"
     + model_id.rstrip("/").split("/")[-1]
-    + "-FP8-NVFP4"
 )
 model.save_pretrained(SAVE_DIR, save_compressed=True)
 tokenizer.save_pretrained(SAVE_DIR)

 # Configure the quantization algorithm to run.
 recipe = QuantizationModifier(
     config_groups={
+        "attention_shared_experts": QuantizationScheme(
             targets=[r"re:.*self_attn\..*"],
             **FP8_BLOCK,
         ),
     ignore=[
         r"re:^model\.layers\.[0-2]\..*"
         r"re:.*mlp\.gate.*",  # not technically necessary
+        r"re:.*indexer\.weights_proj$",  # sensitive to quantization
         r"lm_head",
     ],
 )
 SAVE_DIR = (
     "/mnt/nvme-data/engine/kylesayrs/"
     + model_id.rstrip("/").split("/")[-1]
+    + "-NVFP4-FP8-fp32scales"
 )
 model.save_pretrained(SAVE_DIR, save_compressed=True)
 tokenizer.save_pretrained(SAVE_DIR)