# Recipe: Qwen3.5-122B-A10B-INT4-Autoround
# Qwen3.5-122B model in Intel INT4-Autoround quantization
# Important: set memory utilization in GB, not percentage! Requires --no-ray to fit full context on two sparks.
# If you experience node shutdown, please limit GPU clocks on the affected node (or both): `sudo nvidia-smi -lgc 200,2150`

recipe_version: "1"
name: Qwen3.5-397B-INT4-Autoround (PP=3)
description: Recipe for Qwen3.5-397B-INT4-Autoround to run on 3-node mesh in pipeline-parallel mode

# HuggingFace model to download (optional, for --download-model)
model: Intel/Qwen3.5-397B-A17B-int4-AutoRound

cluster_only: true

# Container image to use
container: vllm-node-tf5

build_args:
  - --tf5

# Mod required to fix ROPE syntax error
mods:
  - mods/fix-qwen3.5-autoround
  - mods/fix-qwen3.5-chat-template

# Default settings (can be overridden via CLI)
defaults:
  port: 8000
  host: 0.0.0.0
  pipeline_parallel: 3
  gpu_memory_utilization: 0.7
  max_model_len: 262144
  max_num_batched_tokens: 16384

# Environment variables
env: 
  PYTORCH_CUDA_ALLOC_CONF: "expandable_segments:True" 
  VLLM_MARLIN_USE_ATOMIC_ADD: 1

# The vLLM serve command template
command: |
  vllm serve Intel/Qwen3.5-397B-A17B-int4-AutoRound \
    --max-model-len {max_model_len} \
    --max-num-seqs 10 \
    --kv-cache-dtype fp8 \
    --gpu-memory-utilization {gpu_memory_utilization} \
    --port {port} \
    --host {host} \
    --enable-prefix-caching \
    --enable-auto-tool-choice \
    --tool-call-parser qwen3_coder \
    --reasoning-parser qwen3 \
    --max-num-batched-tokens {max_num_batched_tokens} \
    --trust-remote-code \
    --chat-template unsloth.jinja \
    -tp 1 \
    -pp {pipeline_parallel} \
    --distributed-executor-backend ray