|
|
--- |
|
|
library_name: vllm |
|
|
language: |
|
|
- ar |
|
|
- de |
|
|
- en |
|
|
- es |
|
|
- fr |
|
|
- hi |
|
|
- id |
|
|
- it |
|
|
- pt |
|
|
- th |
|
|
- tl |
|
|
- vi |
|
|
base_model: |
|
|
- meta-llama/Llama-4-Scout-17B-16E-Instruct |
|
|
pipeline_tag: image-text-to-text |
|
|
tags: |
|
|
- facebook |
|
|
- meta |
|
|
- pytorch |
|
|
- llama |
|
|
- llama4 |
|
|
- neuralmagic |
|
|
- redhat |
|
|
- llmcompressor |
|
|
- quantized |
|
|
- W4A16 |
|
|
- INT4 |
|
|
license: other |
|
|
license_name: llama4 |
|
|
--- |
|
|
|
|
|
<h1 style="display: flex; align-items: center; gap: 10px; margin: 0;"> |
|
|
Llama-4-Scout-17B-16E-Instruct-quantized.w4a16 |
|
|
<img src="https://huggingface.co/RedHatAI/Llama-4-Scout-17B-16E-Instruct-FP8-dynamic/resolve/main/assets/Catalog-Validated_model.png" alt="Model Icon" width="40" style="margin: 0; padding: 0;" /> |
|
|
</h1> |
|
|
|
|
|
<a href="https://www.redhat.com/en/products/ai/validated-models" target="_blank" style="margin: 0; padding: 0;"> |
|
|
<img src="https://huggingface.co/RedHatAI/Llama-4-Scout-17B-16E-Instruct-FP8-dynamic/resolve/main/assets/Validated_badge-Dark.png" alt="Validated Badge" width="250" style="margin: 0; padding: 0;" /> |
|
|
</a> |
|
|
|
|
|
## Model Overview |
|
|
- **Model Architecture:** Llama4ForConditionalGeneration |
|
|
- **Input:** Text / Image |
|
|
- **Output:** Text |
|
|
- **Model Optimizations:** |
|
|
- **Activation quantization:** None |
|
|
- **Weight quantization:** INT4 |
|
|
- **Release Date:** 04/25/2025 |
|
|
- **Version:** 1.0 |
|
|
- **Model Developers:** Red Hat (Neural Magic) |
|
|
|
|
|
|
|
|
### Model Optimizations |
|
|
|
|
|
This model was obtained by quantizing weights of [Llama-4-Scout-17B-16E-Instruct](https://huggingface.co/meta-llama/Llama-4-Scout-17B-16E-Instruct) to INT4 data type. |
|
|
This optimization reduces the number of bits used to represent weights from 16 to 4, reducing GPU memory requirements by approximately 75%. |
|
|
Weight quantization also reduces disk size requirements by approximately 75%. The [llm-compressor](https://github.com/vllm-project/llm-compressor) library is used for quantization. |
|
|
|
|
|
|
|
|
## Deployment |
|
|
|
|
|
This model can be deployed efficiently on vLLM, Red Hat Enterprise Linux AI, and Openshift AI, as shown in the example below. |
|
|
|
|
|
Deploy on <strong>vLLM</strong> |
|
|
|
|
|
```python |
|
|
from vllm import LLM, SamplingParams |
|
|
from transformers import AutoTokenizer |
|
|
|
|
|
model_id = "RedHatAI/Llama-4-Scout-17B-16E-Instruct-quantized.w4a16" |
|
|
number_gpus = 4 |
|
|
|
|
|
sampling_params = SamplingParams(temperature=0.7, top_p=0.8, max_tokens=256) |
|
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained(model_id) |
|
|
|
|
|
prompt = "Give me a short introduction to large language model." |
|
|
|
|
|
llm = LLM(model=model_id, tensor_parallel_size=number_gpus) |
|
|
|
|
|
outputs = llm.generate(prompt, sampling_params) |
|
|
|
|
|
generated_text = outputs[0].outputs[0].text |
|
|
print(generated_text) |
|
|
``` |
|
|
|
|
|
vLLM also supports OpenAI-compatible serving. See the [documentation](https://docs.vllm.ai/en/latest/) for more details. |
|
|
|
|
|
<details> |
|
|
<summary>Deploy on <strong>Red Hat AI Inference Server</strong></summary> |
|
|
|
|
|
```bash |
|
|
$ podman run --rm -it --device nvidia.com/gpu=all -p 8000:8000 \ |
|
|
--ipc=host \ |
|
|
--env "HUGGING_FACE_HUB_TOKEN=$HF_TOKEN" \ |
|
|
--env "HF_HUB_OFFLINE=0" -v ~/.cache/vllm:/home/vllm/.cache \ |
|
|
--name=vllm \ |
|
|
registry.access.redhat.com/rhaiis/rh-vllm-cuda \ |
|
|
vllm serve \ |
|
|
--tensor-parallel-size 8 \ |
|
|
--max-model-len 32768 \ |
|
|
--enforce-eager --model RedHatAI/Llama-4-Scout-17B-16E-Instruct-quantized.w4a16 |
|
|
``` |
|
|
</details> |
|
|
|
|
|
<details> |
|
|
<summary>Deploy on <strong>Red Hat Enterprise Linux AI</strong></summary> |
|
|
|
|
|
```bash |
|
|
# Download model from Red Hat Registry via docker |
|
|
# Note: This downloads the model to ~/.cache/instructlab/models unless --model-dir is specified. |
|
|
ilab model download --repository docker://registry.redhat.io/rhelai1/llama-4-scout-17b-16e-instruct-quantized-w4a16:1.5 |
|
|
``` |
|
|
|
|
|
```bash |
|
|
# Serve model via ilab |
|
|
ilab model serve --model-path ~/.cache/instructlab/models/llama-4-scout-17b-16e-instruct-quantized-w4a16 |
|
|
|
|
|
# Chat with model |
|
|
ilab model chat --model ~/.cache/instructlab/models/llama-4-scout-17b-16e-instruct-quantized-w4a16 |
|
|
``` |
|
|
See [Red Hat Enterprise Linux AI documentation](https://docs.redhat.com/en/documentation/red_hat_enterprise_linux_ai/1.4) for more details. |
|
|
</details> |
|
|
|
|
|
<details> |
|
|
<summary>Deploy on <strong>Red Hat Openshift AI</strong></summary> |
|
|
|
|
|
```python |
|
|
# Setting up vllm server with ServingRuntime |
|
|
# Save as: vllm-servingruntime.yaml |
|
|
apiVersion: serving.kserve.io/v1alpha1 |
|
|
kind: ServingRuntime |
|
|
metadata: |
|
|
name: vllm-cuda-runtime # OPTIONAL CHANGE: set a unique name |
|
|
annotations: |
|
|
openshift.io/display-name: vLLM NVIDIA GPU ServingRuntime for KServe |
|
|
opendatahub.io/recommended-accelerators: '["nvidia.com/gpu"]' |
|
|
labels: |
|
|
opendatahub.io/dashboard: 'true' |
|
|
spec: |
|
|
annotations: |
|
|
prometheus.io/port: '8080' |
|
|
prometheus.io/path: '/metrics' |
|
|
multiModel: false |
|
|
supportedModelFormats: |
|
|
- autoSelect: true |
|
|
name: vLLM |
|
|
containers: |
|
|
- name: kserve-container |
|
|
image: quay.io/modh/vllm:rhoai-2.20-cuda # CHANGE if needed. If AMD: quay.io/modh/vllm:rhoai-2.20-rocm |
|
|
command: |
|
|
- python |
|
|
- -m |
|
|
- vllm.entrypoints.openai.api_server |
|
|
args: |
|
|
- "--port=8080" |
|
|
- "--model=/mnt/models" |
|
|
- "--served-model-name={{.Name}}" |
|
|
env: |
|
|
- name: HF_HOME |
|
|
value: /tmp/hf_home |
|
|
ports: |
|
|
- containerPort: 8080 |
|
|
protocol: TCP |
|
|
``` |
|
|
|
|
|
```python |
|
|
# Attach model to vllm server. This is an NVIDIA template |
|
|
# Save as: inferenceservice.yaml |
|
|
apiVersion: serving.kserve.io/v1beta1 |
|
|
kind: InferenceService |
|
|
metadata: |
|
|
annotations: |
|
|
openshift.io/display-name: Llama-4-Scout-17B-16E-Instruct-quantized.w4a16 # OPTIONAL CHANGE |
|
|
serving.kserve.io/deploymentMode: RawDeployment |
|
|
name: Llama-4-Scout-17B-16E-Instruct-quantized.w4a16 # specify model name. This value will be used to invoke the model in the payload |
|
|
labels: |
|
|
opendatahub.io/dashboard: 'true' |
|
|
spec: |
|
|
predictor: |
|
|
maxReplicas: 1 |
|
|
minReplicas: 1 |
|
|
model: |
|
|
modelFormat: |
|
|
name: vLLM |
|
|
name: '' |
|
|
resources: |
|
|
limits: |
|
|
cpu: '2' # this is model specific |
|
|
memory: 8Gi # this is model specific |
|
|
nvidia.com/gpu: '1' # this is accelerator specific |
|
|
requests: # same comment for this block |
|
|
cpu: '1' |
|
|
memory: 4Gi |
|
|
nvidia.com/gpu: '1' |
|
|
runtime: vllm-cuda-runtime # must match the ServingRuntime name above |
|
|
storageUri: oci://registry.redhat.io/rhelai1/modelcar-llama-4-scout-17b-16e-instruct-quantized-w4a16:1.5 |
|
|
tolerations: |
|
|
- effect: NoSchedule |
|
|
key: nvidia.com/gpu |
|
|
operator: Exists |
|
|
``` |
|
|
|
|
|
```bash |
|
|
# make sure first to be in the project where you want to deploy the model |
|
|
# oc project <project-name> |
|
|
|
|
|
# apply both resources to run model |
|
|
|
|
|
# Apply the ServingRuntime |
|
|
oc apply -f vllm-servingruntime.yaml |
|
|
|
|
|
# Apply the InferenceService |
|
|
oc apply -f qwen-inferenceservice.yaml |
|
|
``` |
|
|
|
|
|
```python |
|
|
# Replace <inference-service-name> and <cluster-ingress-domain> below: |
|
|
# - Run `oc get inferenceservice` to find your URL if unsure. |
|
|
|
|
|
# Call the server using curl: |
|
|
curl https://<inference-service-name>-predictor-default.<domain>/v1/chat/completions |
|
|
-H "Content-Type: application/json" \ |
|
|
-d '{ |
|
|
"model": $(model-name), |
|
|
"stream": true, |
|
|
"stream_options": { |
|
|
"include_usage": true |
|
|
}, |
|
|
"max_tokens": 1, |
|
|
"messages": [ |
|
|
{ |
|
|
"role": "user", |
|
|
"content": "How can a bee fly when its wings are so small?" |
|
|
} |
|
|
] |
|
|
}' |
|
|
|
|
|
``` |
|
|
|
|
|
See [Red Hat Openshift AI documentation](https://docs.redhat.com/en/documentation/red_hat_openshift_ai/2025) for more details. |
|
|
</details> |
|
|
|
|
|
|
|
|
|
|
|
## Evaluation |
|
|
|
|
|
The model was evaluated on the OpenLLM leaderboard tasks (v1 and v2), long context RULER, multimodal MMMU, and multimodal ChartQA. |
|
|
All evaluations are obtained through [lm-evaluation-harness](https://github.com/EleutherAI/lm-evaluation-harness). |
|
|
|
|
|
<details> |
|
|
<summary>Evaluation details</summary> |
|
|
|
|
|
**OpenLLM v1** |
|
|
``` |
|
|
lm_eval \ |
|
|
--model vllm \ |
|
|
--model_args pretrained="RedHatAI/Llama-4-Scout-17B-16E-Instruct-quantized.w4a16",dtype=auto,add_bos_token=True,max_model_len=4096,tensor_parallel_size=8,gpu_memory_utilization=0.7,enable_chunked_prefill=True,trust_remote_code=True \ |
|
|
--tasks openllm \ |
|
|
--batch_size auto |
|
|
``` |
|
|
|
|
|
**OpenLLM v2** |
|
|
``` |
|
|
lm_eval \ |
|
|
--model vllm \ |
|
|
--model_args pretrained="RedHatAI/Llama-4-Scout-17B-16E-Instruct-quantized.w4a16",dtype=auto,add_bos_token=False,max_model_len=16384,tensor_parallel_size=8,gpu_memory_utilization=0.5,enable_chunked_prefill=True,trust_remote_code=True \ |
|
|
--tasks leaderboard \ |
|
|
--apply_chat_template \ |
|
|
--fewshot_as_multiturn \ |
|
|
--batch_size auto |
|
|
``` |
|
|
|
|
|
**Long Context RULER** |
|
|
``` |
|
|
lm_eval \ |
|
|
--model vllm \ |
|
|
--model_args pretrained="RedHatAI/Llama-4-Scout-17B-16E-Instruct-quantized.w4a16",dtype=auto,add_bos_token=False,max_model_len=524288,tensor_parallel_size=8,gpu_memory_utilization=0.9,enable_chunked_prefill=True,trust_remote_code=True \ |
|
|
--tasks ruler \ |
|
|
--metadata='{"max_seq_lengths":[131072]}' \ |
|
|
--batch_size auto |
|
|
``` |
|
|
|
|
|
**Multimodal MMMU** |
|
|
``` |
|
|
lm_eval \ |
|
|
--model vllm-vlm \ |
|
|
--model_args pretrained="RedHatAI/Llama-4-Scout-17B-16E-Instruct-quantized.w4a16",dtype=auto,add_bos_token=False,max_model_len=1000000,tensor_parallel_size=8,gpu_memory_utilization=0.9,enable_chunked_prefill=True,trust_remote_code=True,max_images=10 \ |
|
|
--tasks mmmu_val \ |
|
|
--apply_chat_template \ |
|
|
--batch_size auto |
|
|
``` |
|
|
|
|
|
**Multimodal ChartQA** |
|
|
``` |
|
|
export VLLM_MM_INPUT_CACHE_GIB=8 |
|
|
lm_eval \ |
|
|
--model vllm-vlm \ |
|
|
--model_args pretrained="RedHatAI/Llama-4-Scout-17B-16E-Instruct-quantized.w4a16",dtype=auto,add_bos_token=False,max_model_len=1000000,tensor_parallel_size=8,gpu_memory_utilization=0.9,enable_chunked_prefill=True,trust_remote_code=True,max_images=10 \ |
|
|
--tasks chartqa \ |
|
|
--apply_chat_template \ |
|
|
--batch_size auto |
|
|
``` |
|
|
|
|
|
</details> |
|
|
|
|
|
### Accuracy |
|
|
|
|
|
| | Recovery (%) | meta-llama/Llama-4-Scout-17B-16E-Instruct | RedHatAI/Llama-4-Scout-17B-16E-Instruct-quantized.w4a16<br>(this model) | |
|
|
| ---------------------------------------------- | :-----------: | :---------------------------------------: | :-----------------------------------------------------------------: | |
|
|
| ARC-Challenge<br>25-shot | 98.51 | 69.37 | 68.34 | |
|
|
| GSM8k<br>5-shot | 100.4 | 90.45 | 90.90 |
|
|
| HellaSwag<br>10-shot | 99.67 | 85.23 | 84.95 | |
|
|
| MMLU<br>5-shot | 99.75 | 80.54 | 80.34 | |
|
|
| TruthfulQA<br>0-shot | 99.82 | 61.41 | 61.30 | |
|
|
| WinoGrande<br>5-shot | 98.98 | 77.90 | 77.11 | |
|
|
| **OpenLLM v1<br>Average Score** | **99.59** | **77.48** | **77.16** | |
|
|
| IFEval<br>0-shot<br>avg of inst and prompt acc | 99.51 | 86.90 | 86.47 | |
|
|
| Big Bench Hard<br>3-shot | 99.46 | 65.13 | 64.78 | |
|
|
| Math Lvl 5<br>4-shot | 99.22 | 57.78 | 57.33 | |
|
|
| GPQA<br>0-shot | 100.0 | 31.88 | 31.88 | |
|
|
| MuSR<br>0-shot | 100.9 | 42.20 | 42.59 | |
|
|
| MMLU-Pro<br>5-shot | 98.67 | 55.70 | 54.96 | |
|
|
| **OpenLLM v2<br>Average Score** | **99.54** | **56.60** | **56.34** | | |
|
|
| MMMU<br>0-shot | 100.6 | 53.44 | 53.78 | |
|
|
| ChartQA<br>0-shot<br>exact_match | 100.1 | 65.88 | 66.00 | |
|
|
| ChartQA<br>0-shot<br>relaxed_accuracy | 99.55 | 88.92 | 88.52 | |
|
|
| **Multimodal Average Score** | **100.0** | **69.41** | **69.43** | |
|
|
| RULER<br>seqlen = 131072<br>niah_multikey_1 | 98.41 | 88.20 | 86.80 | |
|
|
| RULER<br>seqlen = 131072<br>niah_multikey_2 | 94.73 | 83.60 | 79.20 | |
|
|
| RULER<br>seqlen = 131072<br>niah_multikey_3 | 96.44 | 78.80 | 76.00 | |
|
|
| RULER<br>seqlen = 131072<br>niah_multiquery | 98.79 | 95.40 | 94.25 | |
|
|
| RULER<br>seqlen = 131072<br>niah_multivalue | 101.6 | 73.75 | 74.95 | |
|
|
| RULER<br>seqlen = 131072<br>niah_single_1 | 100.0 | 100.00 | 100.0 | |
|
|
| RULER<br>seqlen = 131072<br>niah_single_2 | 100.0 | 99.80 | 99.80 | |
|
|
| RULER<br>seqlen = 131072<br>niah_single_3 | 100.2 | 99.80 | 100.0 | |
|
|
| RULER<br>seqlen = 131072<br>ruler_cwe | 87.39 | 39.42 | 33.14 | |
|
|
| RULER<br>seqlen = 131072<br>ruler_fwe | 98.13 | 92.93 | 91.20 | |
|
|
| RULER<br>seqlen = 131072<br>ruler_qa_hotpot | 100.4 | 48.20 | 48.40 | |
|
|
| RULER<br>seqlen = 131072<br>ruler_qa_squad | 96.22 | 53.57 | 51.55 | |
|
|
| RULER<br>seqlen = 131072<br>ruler_qa_vt | 98.82 | 92.28 | 91.20 | |
|
|
| **RULER<br>seqlen = 131072<br>Average Score** | **98.16** | **80.44** | **78.96** | |
|
|
|
