Sold by: NC
Deployed on AWS
VARCO LLM 2.0 is NC's large language model that can be applied to the development of natural language processing-based AI services.
Overview
VARCO LLM 2.0 is NC's large language model that can be applied to the development of various natural language processing-based AI services such as text generation, question answering, chatbots, summarization, and information extraction. NC's VARCO LLM 2.0 was developed with our own technology, including data construction, pre-training, instruction tuning and alignment tuning. We evaluated VARCO LLM 2.0 on various NLP tasks and its performance has significantly improved compared to VARCO LLM 1.0, and it boasts the highest performance among other Korean LLMs of similar sizes. In particular, it has been trained to be used in high-level natural language processing applications such as creative writing, summarization, question and answering, chatbots and translation, and shows high performance in related quantitative indicators. For inquiries regarding further performance improvement or collaboration for service applications, please contact us by email (varco_llm@ncsoft.com ).
Highlights
- Korean Text Generation : VARCO LLM 2.0 is optimized for Korean natural language generation applications. In particular, it provides more natural and creative responses in understanding user instructions and generating text.
- Fine-tuning your own language model : You can fine-tune VARCO LLM with your own data on the AWS platform. Create your own language model with VARCO LLM.
- Key Skills * Question Answering * Summarization * Translation * Text Generation * Chatbots * Information Extraction * Natural Language Understanding * Creative Writing * Instruction Following * Sentiment Analysis
Details
Sold by
Categories
Delivery method
Latest version
Deployed on AWS
New
Introducing multi-product solutions
You can now purchase comprehensive solutions tailored to use cases and industries.
Features and programs
Financing for AWS Marketplace purchases
AWS Marketplace now accepts line of credit payments through the PNC Vendor Finance program. This program is available to select AWS customers in the US, excluding NV, NC, ND, TN, & VT.
Pricing
Pricing is based on actual usage, with charges varying according to how much you consume. Subscriptions have no end date and may be canceled any time.
Additional AWS infrastructure costs may apply. Use the AWS Pricing Calculator to estimate your infrastructure costs.
Dimension | Description | Cost/host/hour |
|---|---|---|
ml.g4dn.12xlarge Inference (Batch) Recommended | Model inference on the ml.g4dn.12xlarge instance type, batch mode | $0.00 |
ml.g5.12xlarge Inference (Real-Time) Recommended | Model inference on the ml.g5.12xlarge instance type, real-time mode | $1.30 |
ml.g5.12xlarge Training Recommended | Algorithm training on the ml.g5.12xlarge instance type | $1.30 |
Dimensions summary
You pay by the hour based on the SageMaker instance you run and the task you perform. Three separate usage options exist. Two cover model inference: one runs batch inference on the ml.g4dn.12xlarge instance, and one runs real-time inference on the ml.g5.12xlarge instance. The third covers algorithm training on the ml.g5.12xlarge instance. Each option is billed independently per host hour, so your cost depends on which instance you use and how many hours you keep it running. There are no tiers or upfront commitments.
Top-of-mind questions for buyers
Batch inference runs on the ml.g4dn.12xlarge instance and processes grouped requests together. Real-time inference runs on the ml.g5.12xlarge instance and returns responses as requests arrive. Both bill per host hour independently. Choose batch for bulk processing and real-time for interactive, low-latency responses.
Each option meters per host hour while the instance runs. Charges accrue for every hour the instance stays active, whether or not it is processing work. To stop software charges, shut the instance down. Underlying AWS resource fees may still apply separately.
The three options bill independently per host hour. If you run training and inference at once on separate instances, both accrue charges simultaneously on the same invoice. If you only run one task, you pay only for that instance's running hours.
ncsoft.github.io
Helpful?
Vendor refund policy
Please contact us : varco_llm@ncsoft.com
Legal
Vendor terms and conditions
Upon subscribing to this product, you must acknowledge and agree to the terms and conditions outlined in the vendor's End User License Agreement (EULA) .
Content disclaimer
Vendors are responsible for their product descriptions and other product content. AWS does not warrant that vendors' product descriptions or other product content are accurate, complete, reliable, current, or error-free.
Delivery details
Amazon SageMaker algorithm
An Amazon SageMaker algorithm is a machine learning model that requires your training data to make predictions. Use the included training algorithm to generate your unique model artifact. Then deploy the model on Amazon SageMaker for real-time inference or batch processing. Amazon SageMaker is a fully managed platform for building, training, and deploying machine learning models at scale.
Deploy the model on Amazon SageMaker AI using the following options:
Version release notes
Major Version release of VARCO LLM 2.0 base Finetuning from NCSOFT
Additional details
Inputs
- Summary
Model accepts JSON requests. You can check examples and fields descriptions.
- Input MIME type
- application/json
{
"repetition_penalty": 1.0,
"temperature": 0.7,
"top_k": 50,
"top_p": 1.0,
"text": "input text here"
}
https://github.com/ncsoft/ncresearch/blob/feature/VARCO_LLM_2_0/notebooks/train_data_example.json
Input data descriptions
The following table describes supported input data fields for real-time inference and batch transform.
Field name | Description | Constraints | Required |
|---|---|---|---|
repetition_penalty | The parameter for repetition penalty. 1.0 means no penalty.
| Default value: 1.0
Type: Continuous | No |
temperature | The value used to modulate the next token probabilities.
| Default value: 0.7
Type: Continuous | No |
top_k | The number of highest probability vocabulary tokens to keep for top-k-filtering.
| Default value: 50
Type: Integer | No |
top_p | If set to float < 1, only the smallest set of most probable tokens with probabilities that add up to top_p or higher are kept for generation.
| Default value: 1.0
Type: Continuous | No |
text | The sequence used as a prompt for the generation or as model inputs to the encoder.
| Type: FreeText | Yes |
Resources
Vendor resources
Support
Vendor support
NC R&D Center, 12, Daewangpangyo-ro 644beon-gil, Bundang-gu, Seongnam-si, Gyeonggi-do, Republic of Korea Tel :02-6201-0099/Email :nc-ai@ncsoft.com
AWS infrastructure support
AWS Support is a one-on-one, fast-response support channel that is staffed 24x7x365 with experienced and technical support engineers. The service helps customers of all sizes and technical abilities to successfully utilize the products and features provided by Amazon Web Services.
Similar products
Build advanced NLP-based AI services using our VARCO-LLM. VARCO-LLM is a high-performance large language model trained by NC.
Llama-3-Varco-Offsetbias-8B is NC's generative judge model that performs pairwise preference evaluation task.
VARCO LLM 2.0 is NC's large language model that can be applied to the development of natural language processing-based AI services.
Llama-VARCO-8B-Instruct is a generative model built with Llama, specifically designed to excel in Korean through additional training
Customer reviews
No customer reviews yet
Be the first to review this product . We've partnered with PeerSpot to gather customer feedback. You can share your experience by writing or recording a review, or scheduling a call with a PeerSpot analyst.