meta:
  title: "ERNIE-4.5-VL"
  slug: "ernie-4.5-vl"
  provider: "Ernie (Baidu)"
  description: "Baidu ERNIE 4.5 VL MoE vision-language models (28B-A3B, 424B-A47B) with heterogeneous text/vision experts"
  date_added: 2025-09-03
  date_updated: 2026-04-17
  tasks:
    - multimodal
  related_recipes:
    - baidu/ERNIE-4.5-21B-A3B-PT
  hardware:
    mi300x: verified
    mi325x: verified
    mi355x: verified

model:
  model_id: "baidu/ERNIE-4.5-VL-28B-A3B-PT"
  min_vllm_version: "0.11.0"
  architecture: moe
  parameter_count: "28B"
  active_parameters: "3B"
  context_length: 131072
  base_args:
    - "--trust-remote-code"
  base_env: {}

features:
  text_only:
    description: "Skip loading the vision encoder for text-only workloads — frees VRAM for KV cache. Mutually exclusive with encoder_parallel."
    args:
      - "--language-model-only"
  encoder_parallel:
    description: "Run the vision encoder in data-parallel mode — avoids TP comm overhead on the small encoder. Mutually exclusive with text_only."
    args:
      - "--mm-encoder-tp-mode"
      - "data"

opt_in_features:
  - text_only

variants:
  default:
    precision: bf16
    vram_minimum_gb: 67
    description: "BF16 weights; fits on 1x80GB GPU (28B VL variant)"

  "424b":
    model_id: "baidu/ERNIE-4.5-VL-424B-A47B-PT"
    precision: bf16
    vram_minimum_gb: 1120
    description: "424B total / 47B active; 8x140GB BF16 or 16x80GB BF16"
    extra_args:
      - "--trust-remote-code"
      - "--tensor-parallel-size"
      - "8"

  "424b_fp8":
    model_id: "baidu/ERNIE-4.5-VL-424B-A47B-PT"
    precision: fp8
    vram_minimum_gb: 640
    description: "424B with FP8 online quantization + CPU offload for 8x80GB testing"
    extra_args:
      - "--trust-remote-code"
      - "--tensor-parallel-size"
      - "8"
      - "--quantization"
      - "fp8"
      - "--cpu-offload-gb"
      - "50"

compatible_strategies:
  - single_node_tp
  - multi_node_tp
  - multi_node_tep
  - multi_node_dep

kv_offload_support:
  offloading_cpu: verified
  offloading_fs: verified

hardware_overrides: {}
strategy_overrides: {}

guide: |
  ## Overview

  ERNIE 4.5 VL is Baidu's multimodal MoE model with heterogeneous experts (separate text
  and vision experts). Because of the heterogeneous architecture, **torch.compile and CUDA
  graphs are not supported**.

  - `baidu/ERNIE-4.5-VL-28B-A3B-PT` — 28B total / 3B active (1x80GB)
  - `baidu/ERNIE-4.5-VL-424B-A47B-PT` — 424B total / 47B active (8x140GB BF16, 8x80GB FP8+offload, or 16x80GB BF16)

  ## Prerequisites

  - vLLM: support added to main branch recently; install latest
  - Hardware depends on variant

  ### Install vLLM (CUDA)

  ```bash
  uv venv --python 3.12 --seed
  source .venv/bin/activate
  uv pip install -U vllm --torch-backend auto
  ```

  ### Install vLLM (AMD ROCm MI300X/MI325X/MI355X)

  ```bash
  uv pip install vllm --extra-index-url https://wheels.vllm.ai/rocm/
  ```

  ## Launch commands

  28B on 1x80GB:

  ```bash
  vllm serve baidu/ERNIE-4.5-VL-28B-A3B-PT --trust-remote-code
  ```

  424B BF16 on 8x140GB:

  ```bash
  vllm serve baidu/ERNIE-4.5-VL-424B-A47B-PT \
    --trust-remote-code \
    --tensor-parallel-size 8
  ```

  424B with FP8 + CPU offload on 8x80GB (testing only):

  ```bash
  vllm serve baidu/ERNIE-4.5-VL-424B-A47B-PT \
    --trust-remote-code \
    --tensor-parallel-size 8 \
    --quantization fp8 \
    --cpu-offload-gb 50
  ```

  28B on AMD MI300X+:

  ```bash
  VLLM_ROCM_USE_AITER=1 SAFETENSORS_FAST_GPU=1 \
    vllm serve baidu/ERNIE-4.5-VL-28B-A3B-PT \
    --tensor-parallel-size 4 \
    --gpu-memory-utilization 0.9 \
    --trust-remote-code
  ```

  ## Benchmarking

  ```bash
  vllm bench serve \
    --model baidu/ERNIE-4.5-VL-28B-A3B-PT \
    --dataset-name random \
    --random-input-len 8000 --random-output-len 1000 \
    --request-rate 10 --num-prompts 16 --ignore-eos --trust-remote-code
  ```

  ## References

  - [ERNIE-4.5-VL-28B-A3B-PT](https://huggingface.co/baidu/ERNIE-4.5-VL-28B-A3B-PT)
  - [ERNIE-4.5-VL-424B-A47B-PT](https://huggingface.co/baidu/ERNIE-4.5-VL-424B-A47B-PT)