EXAONE-Deep-32B-AWQ
Introduction
We introduce EXAONE Deep, which exhibits superior capabilities in various reasoning tasks including math and coding benchmarks, ranging from 2.4B to 32B parameters developed and released by LG AI Research. Evaluation results show that 1) EXAONE Deep 2.4B outperforms other models of comparable size, 2) EXAONE Deep 7.8B outperforms not only open-weight models of comparable scale but also a proprietary reasoning model OpenAI o1-mini, and 3) EXAONE Deep 32B demonstrates competitive performance against leading open-weight models.
For more details, please refer to our documentation, blog and GitHub.
This repository contains the AWQ-quantized weights of the reasoning 32B language model with the following features:
- Number of Parameters (without embeddings): 30.95B
- Number of Layers: 64
- Number of Attention Heads: GQA with 40 Q-heads and 8 KV-heads
- Vocab Size: 102,400
- Context Length: 32,768 tokens
- Quantization: AWQ with 4-bit group-wise weight-only quantization (W4A16g128)
Quickstart
We recommend to use transformers>=4.43.1
and autoawq>=0.2.8
Here is the code snippet to run conversational inference with the model:
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
from threading import Thread
model_name = "LGAI-EXAONE/EXAONE-Deep-32B-AWQ"
streaming = True # choose the streaming option
model = AutoModelForCausalLM.from_pretrained(
model_name,
torch_dtype=torch.bfloat16,
trust_remote_code=True,
device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained(model_name)
messages = [
{"role": "user", "content": "How many golf balls can fit in a school bus?"}
]
input_ids = tokenizer.apply_chat_template(
messages,
tokenize=True,
add_generation_prompt=True,
return_tensors="pt"
)
if streaming:
streamer = TextIteratorStreamer(tokenizer)
thread = Thread(target=model.generate, kwargs=dict(
input_ids=input_ids.to("cuda"),
eos_token_id=tokenizer.eos_token_id,
max_new_tokens=32768,
do_sample=True,
temperature=0.6,
top_p=0.95,
streamer=streamer
))
thread.start()
for text in streamer:
print(text, end="", flush=True)
else:
output = model.generate(
input_ids.to("cuda"),
eos_token_id=tokenizer.eos_token_id,
max_new_tokens=32768,
do_sample=True,
temperature=0.6,
top_p=0.95,
)
print(tokenizer.decode(output[0]))
Note
The EXAONE Deep models are trained with an optimized configuration, so we recommend following the Usage Guideline section to achieve optimal performance.
Evaluation
You can check the evaluation results of original EXAONE Deep models at GitHub or our documentation.
Deployment
EXAONE Deep models can be inferred in the various frameworks, such as:
TensorRT-LLM
vLLM
SGLang
llama.cpp
Ollama
LM-Studio
Please refer to our EXAONE Deep GitHub for more details about the inference frameworks.
Quantization
We provide the pre-quantized EXAONE Deep models with AWQ and several quantization types in GGUF format. Please refer to our EXAONE Deep collection to find corresponding quantized models.
Usage Guideline
To achieve the expected performance, we recommend using the following configurations:
- Ensure the model starts with
<thought>\n
for reasoning steps. The model's output quality may be degraded when you omit it. You can easily apply this feature by usingtokenizer.apply_chat_template()
withadd_generation_prompt=True
. Please check the example code on Quickstart section. - The reasoning steps of EXAONE Deep models enclosed by
<thought>\n...\n</thought>
usually have lots of tokens, so previous reasoning steps may be necessary to be removed in multi-turn situation. The provided tokenizer handles this automatically. - Avoid using system prompt, and build the instruction on the user prompt.
- Additional instructions help the models reason more deeply, so that the models generate better output.
- For math problems, the instructions "Please reason step by step, and put your final answer within \boxed{}." are helpful.
- For more information on our evaluation setting including prompts, please refer to our Documentation.
- In our evaluation, we use
temperature=0.6
andtop_p=0.95
for generation. - When evaluating the models, it is recommended to test multiple times to assess the expected performance accurately.
Limitation
The EXAONE language model has certain limitations and may occasionally generate inappropriate responses. The language model generates responses based on the output probability of tokens, and it is determined during learning from training data. While we have made every effort to exclude personal, harmful, and biased information from the training data, some problematic content may still be included, potentially leading to undesirable responses. Please note that the text generated by EXAONE language model does not reflects the views of LG AI Research.
- Inappropriate answers may be generated, which contain personal, harmful or other inappropriate information.
- Biased responses may be generated, which are associated with age, gender, race, and so on.
- The generated responses rely heavily on statistics from the training data, which can result in the generation of semantically or syntactically incorrect sentences.
- Since the model does not reflect the latest information, the responses may be false or contradictory.
LG AI Research strives to reduce potential risks that may arise from EXAONE language models. Users are not allowed to engage in any malicious activities (e.g., keying in illegal information) that may induce the creation of inappropriate outputs violating LG AI’s ethical principles when using EXAONE language models.
License
The model is licensed under EXAONE AI Model License Agreement 1.1 - NC
Citation
@article{exaone-deep,
title={EXAONE Deep: Reasoning Enhanced Language Models},
author={{LG AI Research}},
journal={arXiv preprint arXiv:2503.12524},
year={2025}
}
Contact
LG AI Research Technical Support: [email protected]
- Downloads last month
- 118
Model tree for LGAI-EXAONE/EXAONE-Deep-32B-AWQ
Base model
LGAI-EXAONE/EXAONE-3.5-32B-Instruct