50 lines
1.4 KiB
YAML
50 lines
1.4 KiB
YAML
# Recipe: Nemotron-3-Nano-NVFP4
|
|
# Nemotron-3-Nano model with NVFP4 quantization support
|
|
# Currently can only be run in solo mode, cluster mode fails with error
|
|
|
|
recipe_version: "1"
|
|
name: Nemotron-3-Nano-NVFP4
|
|
description: vLLM serving Nemotron-3-Nano-NVFP4 on a SINGLE NODE ONLY!
|
|
|
|
# HuggingFace model to download (optional, for --download-model)
|
|
model: nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-NVFP4
|
|
|
|
# Container image to use
|
|
container: vllm-node
|
|
|
|
# This model can only run on single node (solo)
|
|
solo_only: true
|
|
|
|
# No mods required
|
|
mods:
|
|
- mods/nemotron-nano
|
|
|
|
# Default settings (can be overridden via CLI)
|
|
defaults:
|
|
port: 8000
|
|
host: 0.0.0.0
|
|
tensor_parallel: 1
|
|
gpu_memory_utilization: 0.7
|
|
max_model_len: 131072
|
|
|
|
# Environment variables
|
|
env:
|
|
VLLM_USE_FLASHINFER_MOE_FP4: 1
|
|
VLLM_FLASHINFER_MOE_BACKEND: "throughput"
|
|
|
|
# The vLLM serve command template
|
|
command: |
|
|
vllm serve nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-NVFP4 \
|
|
--max-model-len {max_model_len} \
|
|
--port {port} --host {host} \
|
|
--trust-remote-code \
|
|
--enable-auto-tool-choice \
|
|
--tool-call-parser qwen3_coder \
|
|
--reasoning-parser-plugin nano_v3_reasoning_parser.py \
|
|
--reasoning-parser nano_v3 \
|
|
--kv-cache-dtype fp8 \
|
|
--enable-prefix-caching \
|
|
--attention-backend flashinfer \
|
|
--load-format fastsafetensors \
|
|
--gpu-memory-utilization {gpu_memory_utilization}
|