229 lines
7.8 KiB
Markdown
229 lines
7.8 KiB
Markdown
---
|
|
library_name: transformers
|
|
license: apache-2.0
|
|
pipeline_tag: text-generation
|
|
base_model:
|
|
- Qwen/Qwen3-30B-A3B-Instruct-2507
|
|
tags:
|
|
- neuralmagic
|
|
- redhat
|
|
- llmcompressor
|
|
- quantized
|
|
- INT8
|
|
---
|
|
|
|
# Qwen3-30B-A3B-Instruct-2507.w8a8
|
|
|
|
## Model Overview
|
|
- **Model Architecture:** Qwen3ForCausalLM
|
|
- **Input:** Text
|
|
- **Output:** Text
|
|
- **Model Optimizations:**
|
|
- **Weight quantization:** INT8
|
|
- **Intended Use Cases:**
|
|
- Reasoning.
|
|
- Function calling.
|
|
- Subject matter experts via fine-tuning.
|
|
- Multilingual instruction following.
|
|
- Translation.
|
|
- **Out-of-scope:** Use in any manner that violates applicable laws or regulations (including trade compliance laws).
|
|
- **Release Date:** 05/05/2025
|
|
- **Version:** 1.0
|
|
- **Model Developers:** RedHat (Neural Magic)
|
|
|
|
### Model Optimizations
|
|
|
|
This model was obtained by quantizing the weights of [Qwen/Qwen3-30B-A3B-Instruct-2507](https://huggingface.co/Qwen/Qwen3-30B-A3B-Instruct-2507) to INT8 data type.
|
|
This optimization reduces the number of bits used to represent weights and activations from 16 to 8, reducing GPU memory requirements (by approximately 50%) and increasing matrix-multiply compute throughput (by approximately 2x).
|
|
Weight quantization also reduces disk size requirements by approximately 50%.
|
|
|
|
Only weights and activations of the linear operators within transformers blocks are quantized.
|
|
Weights are quantized with a symmetric static per-channel scheme, whereas activations are quantized with a symmetric dynamic per-token scheme.
|
|
A combination of the [SmoothQuant](https://arxiv.org/abs/2211.10438) and [GPTQ](https://arxiv.org/abs/2210.17323) algorithms is applied for quantization, as implemented in the [llm-compressor](https://github.com/vllm-project/llm-compressor) library.
|
|
|
|
|
|
## Deployment
|
|
|
|
This model can be deployed efficiently using the [vLLM](https://docs.vllm.ai/en/latest/) backend, as shown in the example below.
|
|
|
|
```python
|
|
from vllm import LLM, SamplingParams
|
|
from transformers import AutoTokenizer
|
|
|
|
model_id = "RedHatAI/Qwen3-30B-A3B-Instruct-2507.w8a8"
|
|
number_gpus = 1
|
|
sampling_params = SamplingParams(temperature=0.6, top_p=0.95, top_k=20, min_p=0, max_tokens=256)
|
|
|
|
messages = [
|
|
{"role": "user", "content": prompt}
|
|
]
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
|
|
|
messages = [{"role": "user", "content": "Give me a short introduction to large language model."}]
|
|
|
|
prompts = tokenizer.apply_chat_template(messages, add_generation_prompt=True, tokenize=False)
|
|
|
|
llm = LLM(model=model_id, tensor_parallel_size=number_gpus)
|
|
|
|
outputs = llm.generate(prompts, sampling_params)
|
|
|
|
generated_text = outputs[0].outputs[0].text
|
|
print(generated_text)
|
|
```
|
|
|
|
vLLM aslo supports OpenAI-compatible serving. See the [documentation](https://docs.vllm.ai/en/latest/) for more details.
|
|
|
|
## Creation
|
|
|
|
<details>
|
|
<summary>Creation details</summary>
|
|
This model was created with [llm-compressor](https://github.com/vllm-project/llm-compressor) by running the code snippet below.
|
|
|
|
|
|
```python
|
|
from llmcompressor.modifiers.quantization import GPTQModifier
|
|
from llmcompressor.modifiers.smoothquant import SmoothQuantModifier
|
|
from llmcompressor.transformers import oneshot
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer
|
|
|
|
# Load model
|
|
model_stub = "Qwen/Qwen3-30B-A3B-Instruct"
|
|
model_name = model_stub.split("/")[-1]
|
|
|
|
num_samples = 1024
|
|
max_seq_len = 8192
|
|
|
|
model = AutoModelForCausalLM.from_pretrained(model_stub)
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained(model_stub)
|
|
|
|
def preprocess_fn(example):
|
|
return {"text": tokenizer.apply_chat_template(example["messages"], add_generation_prompt=False, tokenize=False)}
|
|
|
|
ds = load_dataset("neuralmagic/LLM_compression_calibration", split="train")
|
|
ds = ds.map(preprocess_fn)
|
|
|
|
# Configure the quantization algorithm and scheme
|
|
recipe = [
|
|
SmoothQuantModifier(
|
|
smoothing_strength=0.9,
|
|
mappings=[
|
|
[["re:.*q_proj", "re:.*k_proj", "re:.*v_proj"], "re:.*input_layernorm"],
|
|
],
|
|
),
|
|
GPTQModifier(
|
|
ignore: ["lm_head"]
|
|
config_groups={"group_0": {"targets": ["Linear"], "weights": { "num_bits": 4, "type": int, "strategy": "group", "group_size": 128, "symmetric": true, "dynamic": false, "observer": "mse" } } },
|
|
dampening_frac=0.1,
|
|
)
|
|
]
|
|
|
|
|
|
# Apply quantization
|
|
oneshot(
|
|
model=model,
|
|
dataset=ds,
|
|
recipe=recipe,
|
|
max_seq_length=max_seq_len,
|
|
num_calibration_samples=num_samples,
|
|
)
|
|
|
|
# Save to disk in compressed-tensors format
|
|
save_path = model_name + "-quantized.w8a8"
|
|
model.save_pretrained(save_path)
|
|
tokenizer.save_pretrained(save_path)
|
|
print(f"Model and tokenizer saved to: {save_path}")
|
|
```
|
|
</details>
|
|
|
|
|
|
|
|
## Evaluation
|
|
|
|
The model was evaluated on the ifeval, mmlu_pro and gsm8k_platinum using [lm-evaluation-harness](https://github.com/EleutherAI/lm-evaluation-harness), on reasoning tasks using [lighteval](https://github.com/neuralmagic/lighteval/tree/reasoning).
|
|
[vLLM](https://docs.vllm.ai/en/stable/) was used for all evaluations.
|
|
|
|
|
|
<details>
|
|
<summary>Evaluation details</summary>
|
|
|
|
Deploy using vllm to create an OpenAI-compatible API endpoint:
|
|
|
|
- vLLM:
|
|
```shell
|
|
vllm serve RedHatAI/Qwen3-30B-A3B-Instruct-2507.w8a8 --max-model-len 262144
|
|
```
|
|
|
|
**lm-evaluation-harness**
|
|
```
|
|
lm_eval --model local-chat-completions \
|
|
--tasks mmlu_pro_chat \
|
|
--model_args "model=RedHatAI/Qwen3-30B-A3B-Instruct-2507.w8a8,max_length=262144,base_url=http://0.0.0.0:8000/v1/chat/completions,num_concurrent=64,max_retries=3,tokenized_requests=False,tokenizer_backend=None,timeout=1200" \
|
|
--num_fewshot 0 \
|
|
--apply_chat_template \
|
|
--gen_kwargs "do_sample=True,temperature=0.6,top_p=0.95,top_k=20,min_p=0.0,max_gen_toks=64000
|
|
```
|
|
|
|
```
|
|
lm_eval --model local-chat-completions \
|
|
--tasks ifeval \
|
|
--model_args "model=RedHatAI/Qwen3-30B-A3B-Instruct-2507.w8a8,max_length=262144,base_url=http://0.0.0.0:8000/v1/chat/completions,num_concurrent=64,max_retries=3,tokenized_requests=False,tokenizer_backend=None,timeout=1200" \
|
|
--num_fewshot 0 \
|
|
--apply_chat_template \
|
|
--gen_kwargs "do_sample=True,temperature=0.6,top_p=0.95,top_k=20,min_p=0.0,max_gen_toks=64000
|
|
```
|
|
|
|
```
|
|
lm_eval --model local-chat-completions \
|
|
--tasks mmlu_cot_llama \
|
|
--model_args "model=RedHatAI/Qwen3-30B-A3B-Instruct-2507.w8a8,max_length=262144,base_url=http://0.0.0.0:8000/v1/chat/completions,num_concurrent=64,max_retries=3,tokenized_requests=False,tokenizer_backend=None,timeout=1200" \
|
|
--num_fewshot 0 \
|
|
--apply_chat_template \
|
|
--gen_kwargs "do_sample=True,temperature=0.6,top_p=0.95,top_k=20,min_p=0.0,max_gen_toks=64000
|
|
```
|
|
|
|
```
|
|
lm_eval --model local-chat-completions \
|
|
--tasks gsm8k_platinum_cot_llama \
|
|
--model_args "model=RedHatAI/Qwen3-30B-A3B-Instruct-2507.w8a8,max_length=262144,base_url=http://0.0.0.0:8000/v1/chat/completions,num_concurrent=64,max_retries=3,tokenized_requests=False,tokenizer_backend=None,timeout=1200" \
|
|
--num_fewshot 0 \
|
|
--apply_chat_template \
|
|
--gen_kwargs "do_sample=True,temperature=0.6,top_p=0.95,top_k=20,min_p=0.0,max_gen_toks=64000
|
|
```
|
|
|
|
**lighteval**
|
|
|
|
lighteval_model_arguments.yaml
|
|
```yaml
|
|
model_parameters:
|
|
model_name: RedHatAI/Qwen3-30B-A3B-Instruct-2507.w8a8
|
|
dtype: auto
|
|
gpu_memory_utilization: 0.9
|
|
max_model_length: 40960
|
|
generation_parameters:
|
|
temperature: 0.6
|
|
top_k: 20
|
|
min_p: 0.0
|
|
top_p: 0.95
|
|
max_new_tokens: 32000
|
|
```
|
|
|
|
```
|
|
lighteval endpoint litellm lighteval_model_arguments.yaml \
|
|
"aime25|0,math_500|0,gpqa:diamond|0"
|
|
```
|
|
|
|
|
|
</details>
|
|
|
|
### Accuracy
|
|
|
|
|
|
| Benchmark | Qwen3-30B-A3B Instruct | Qwen3-30B-A3B Instruct.w8a8 (this model) | Recovery (%) |
|
|
|--------|-------------|-------------------|--------------|
|
|
| GSM8k Platinum (5-shot) | 96.11 | 97.57 | 101.52 |
|
|
| MMLU-Cot (5-shot) | 84.29 | 84.30 | 100.02 |
|
|
| MMLU-Pro (5-shot) | 78.90 | 78.81 | 99.89 |
|
|
| IfEval | 89.13 | 88.89 | 99.73 |
|
|
| Math 500 | 89.91 | 90.48 | 100.62 | |