Instructions to use gauravmm/GLM-4.5-Air-REAP-82B-A12B-NVFP4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use gauravmm/GLM-4.5-Air-REAP-82B-A12B-NVFP4 with Transformers:

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

pipe = pipeline("text-generation", model="gauravmm/GLM-4.5-Air-REAP-82B-A12B-NVFP4")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("gauravmm/GLM-4.5-Air-REAP-82B-A12B-NVFP4")
model = AutoModelForCausalLM.from_pretrained("gauravmm/GLM-4.5-Air-REAP-82B-A12B-NVFP4", 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 gauravmm/GLM-4.5-Air-REAP-82B-A12B-NVFP4 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "gauravmm/GLM-4.5-Air-REAP-82B-A12B-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": "gauravmm/GLM-4.5-Air-REAP-82B-A12B-NVFP4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/gauravmm/GLM-4.5-Air-REAP-82B-A12B-NVFP4

SGLang

How to use gauravmm/GLM-4.5-Air-REAP-82B-A12B-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 "gauravmm/GLM-4.5-Air-REAP-82B-A12B-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": "gauravmm/GLM-4.5-Air-REAP-82B-A12B-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 "gauravmm/GLM-4.5-Air-REAP-82B-A12B-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": "gauravmm/GLM-4.5-Air-REAP-82B-A12B-NVFP4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use gauravmm/GLM-4.5-Air-REAP-82B-A12B-NVFP4 with Docker Model Runner:
```
docker model run hf.co/gauravmm/GLM-4.5-Air-REAP-82B-A12B-NVFP4
```

GLM-4.5-Air-REAP-82B-A12B — NVFP4 (W4A16)

NVFP4 weight-only (W4A16) quantization of Cerebras' GLM-4.5-Air-REAP-82B-A12B — itself a REAP (router-weighted expert-activation pruning) prune of GLM-4.5-Air from 106B→82B total (12B active, 96 experts). This checkpoint quantizes the weights to NVFP4 (4-bit float E2M1, block-16 FP8-E4M3 micro-scale + FP32 per-tensor global scale, ≈4.5 bits/value) while keeping activations in BF16, so it loads and serves on a single 128 GB DGX Spark / GB10. It was produced from the BF16 source (Cerebras recommends BF16 for low-bit quant, avoiding stacked FP8→FP4 rounding) with a custom shard-by-shard streaming quantizer that reuses compressed-tensors' own pack/scale primitives, emitting the standard nvfp4-pack-quantized format — so it's a drop-in for vLLM with no custom code. On-disk size is 51 GB (down from 159 GB BF16); served runtime footprint ≈89 GB. Benchmarked on vLLM at prefill 163 / decode 158 tok/s (ShareGPT, concurrency 32, 64K context). The fused parallel projections (q/k/v→qkv, gate/up→gate_up, expert w1/w3) share one weight_global_scale per group — required for correct output under vLLM's fused MoE/attention layers.

Serving (vLLM)

vllm serve gauravmm/GLM-4.5-Air-REAP-82B-A12B-NVFP4 \
  --quantization compressed-tensors \
  --moe-backend marlin \
  --trust-remote-code \
  --max-model-len 65536 \
  --kv-cache-dtype fp8

--moe-backend marlin is mandatory on this NVFP4 W4A16 checkpoint — the flashinfer_cutlass MoE backend rejects W4A16 NVFP4 and crashes engine-core init.

Provenance

GLM-4.5-Air (zai-org, MIT) → REAP prune (Cerebras) → NVFP4 W4A16 (this repo). License inherited from upstream: MIT.

Downloads last month: 40

Safetensors

Model size

82B params

Tensor type

F32

BF16

F8_E4M3

Model tree for gauravmm/GLM-4.5-Air-REAP-82B-A12B-NVFP4

Base model

zai-org/GLM-4.5-Air

Finetuned

cerebras/GLM-4.5-Air-REAP-82B-A12B

Quantized

(20)

this model