Mistral-Small-24B-Instruct-2501-quantized.w4a16 👁 Model Icon

👁 Validated Badge

Model Overview

• Model Architecture: Mistral-Small-24B-Instruct-2501
  • Input: Text
  • Output: Text
• Model Optimizations:
  • Weight quantization: INT4
  • Activation quantization: None
• Release Date: 3/1/2025
• Version: 1.0
• Validated on: RHOAI 2.20, RHAIIS 3.0, RHELAI 1.5
• Model Developers: Neural Magic

Quantized version of Mistral-Small-24B-Instruct-2501.

Model Optimizations

This model was obtained by quantizing the weights to INT4 data type, ready for inference with vLLM. This optimization reduces the number of bits per parameter from 16 to 4, reducing the disk size and GPU memory requirements by approximately 75%. Only the weights of the linear operators within transformers blocks are quantized.

Deployment

Use with vLLM

1. Initialize vLLM server:

vllm serve RedHatAI/Mistral-Small-24B-Instruct-2501-quantized.w4a16 --tensor_parallel_size 1 --tokenizer_mode mistral

2. Send requests to the server:

from openai import OpenAI

# Modify OpenAI's API key and API base to use vLLM's API server.
openai_api_key = "EMPTY"
openai_api_base = "http://<your-server-host>:8000/v1"

client = OpenAI(
 api_key=openai_api_key,
 base_url=openai_api_base,
)

model = "RedHatAI/Mistral-Small-24B-Instruct-2501-quantized.w4a16"


messages = [
 {"role": "user", "content": "Explain quantum mechanics clearly and concisely."},
]

outputs = client.chat.completions.create(
 model=model,
 messages=messages,
)

generated_text = outputs.choices[0].message.content
print(generated_text)

Creation

This model was created with llm-compressor by running the code snippet below.

python quantize.py --model_path mistralai/Mistral-Small-24B-Instruct-2501 --quant_path "output_dir" --calib_size 1024 --dampening_frac 0.05 --observer minmax --actorder false

from datasets import load_dataset
from transformers import AutoTokenizer
from llmcompressor.modifiers.quantization import GPTQModifier
from llmcompressor.transformers import SparseAutoModelForCausalLM, oneshot, apply
import argparse
from compressed_tensors.quantization import QuantizationScheme, QuantizationArgs, QuantizationType, QuantizationStrategy

def parse_actorder(value):
 # Interpret the input value for --actorder
 if value.lower() == "false":
 return False
 elif value.lower() == "group":
 return "group"
 elif value.lower() == "weight":
 return "weight"
 else:
 raise argparse.ArgumentTypeError("Invalid value for --actorder. Use 'group' or 'False'.")


parser = argparse.ArgumentParser()
parser.add_argument('--model_path', type=str)
parser.add_argument('--quant_path', type=str)
parser.add_argument('--num_bits', type=int, default=4)
parser.add_argument('--sequential_update', type=bool, default=True)
parser.add_argument('--calib_size', type=int, default=256)
parser.add_argument('--dampening_frac', type=float, default=0.05)
parser.add_argument('--observer', type=str, default="minmax")
parser.add_argument(
 '--actorder',
 type=parse_actorder,
 default=False, # Default value is False
 help="Specify actorder as 'group' (string) or False (boolean)."
)

args = parser.parse_args()

model = SparseAutoModelForCausalLM.from_pretrained(
 args.model_path,
 device_map="auto",
 torch_dtype="auto",
 use_cache=False,
)
tokenizer = AutoTokenizer.from_pretrained(args.model_path)

NUM_CALIBRATION_SAMPLES = args.calib_size
DATASET_ID = "garage-bAInd/Open-Platypus"
DATASET_SPLIT = "train"
ds = load_dataset(DATASET_ID, split=DATASET_SPLIT)
ds = ds.shuffle(seed=42).select(range(NUM_CALIBRATION_SAMPLES))

def preprocess(example):
 concat_txt = example["instruction"] + "\n" + example["output"]
 return {"text": concat_txt}

ds = ds.map(preprocess)

def tokenize(sample):
 return tokenizer(
 sample["text"],
 padding=False,
 truncation=False,
 add_special_tokens=True,
 )


ds = ds.map(tokenize, remove_columns=ds.column_names)

quant_scheme = QuantizationScheme(
 targets=["Linear"],
 weights=QuantizationArgs(
 num_bits=args.num_bits,
 type=QuantizationType.INT,
 symmetric=True,
 group_size=128,
 strategy=QuantizationStrategy.GROUP,
 observer=args.observer,
 actorder=args.actorder
 ),
 input_activations=None,
 output_activations=None,
)

recipe = [
 GPTQModifier(
 targets=["Linear"],
 ignore=["lm_head"],
 sequential_update=args.sequential_update,
 dampening_frac=args.dampening_frac,
 config_groups={"group_0": quant_scheme},
 )
]
oneshot(
 model=model,
 dataset=ds,
 recipe=recipe,
 num_calibration_samples=args.calib_size,
)

# Save to disk compressed.
SAVE_DIR = args.quant_path
model.save_pretrained(SAVE_DIR, save_compressed=True)
tokenizer.save_pretrained(SAVE_DIR)

Evaluation

The model was evaluated on OpenLLM Leaderboard V1 and V2, using the following commands:

OpenLLM Leaderboard V1:

lm_eval \
 --model vllm \
 --model_args pretrained="neuralmagic/Mistral-Small-24B-Instruct-2501-quantized.w4a16",dtype=auto,add_bos_token=True,max_model_len=4096,tensor_parallel_size=1,gpu_memory_utilization=0.8,enable_chunked_prefill=True,trust_remote_code=True \
 --tasks openllm \
 --write_out \
 --batch_size auto \
 --output_path output_dir \
 --show_config

OpenLLM Leaderboard V2:

lm_eval \
 --model vllm \
 --model_args pretrained="neuralmagic/Mistral-Small-24B-Instruct-2501-quantized.w4a16",dtype=auto,add_bos_token=False,max_model_len=4096,tensor_parallel_size=1,gpu_memory_utilization=0.8,enable_chunked_prefill=True,trust_remote_code=True \
 --apply_chat_template \
 --fewshot_as_multiturn \
 --tasks leaderboard \
 --write_out \
 --batch_size auto \
 --output_path output_dir \
 --show_config

Accuracy

OpenLLM Leaderboard V1 evaluation scores

OpenLLM Leaderboard V2 evaluation scores

Results on GPQA and MUSR are not considred for accuracy recovery calculation because the unquantized model has close to random prediction accuracy (8.29, 7.84) which doesn't provide a reliable baseline for recovery calculation.

Base model

mistralai/Mistral-Small-24B-Base-2501

mistralai/Mistral-Small-24B-Instruct-2501

(104)

May 2025 Collection of third-party generative AI models validated by Red Hat AI for use across the Red Hat AI Product Portfolio. • 39 items • Updated Apr 30 • 20