Developed by :

  • K2S3

Model Number:

  • K2S3-SOLAR-11b-v3.0

Base Model :

Training Data

  • The training data for this model includes the Standard Korean Dictionary, training data from KULLM at Korea University, abstracts of master's and doctoral theses, Korean language samples from AI Hub, alpaca-gpt4-data, and samples from The OpenOrca Dataset.
  • ์ด ๋ชจ๋ธ์˜ ํ›ˆ๋ จ ๋ฐ์ดํ„ฐ์—๋Š” ํ‘œ์ค€๊ตญ์–ด๋Œ€์‚ฌ์ „, ๊ณ ๋ ค๋Œ€ํ•™๊ต KULLM์—์„œ ์ œ๊ณตํ•œ ํ›ˆ๋ จ ๋ฐ์ดํ„ฐ, ์„์‚ฌ ๋ฐ ๋ฐ•์‚ฌํ•™์œ„ ๋…ผ๋ฌธ์˜ ์ดˆ๋ก, AI Hub์—์„œ ์ œ๊ณตํ•œ ํ•œ๊ตญ์–ด ๋ฐ์ดํ„ฐ ์ƒ˜ํ”Œ, alpaca-gpt4-data, ๊ทธ๋ฆฌ๊ณ  OpenOrca Dataset์—์„œ ์ œ๊ณตํ•œ ์ƒ˜ํ”Œ๋“ค์ด ํฌํ•จ๋ฉ๋‹ˆ๋‹ค.

Training Method

  • This model was fine-tuned on the "upstage/SOLAR-10.7B-v1.0" base model using a full parameter tuning method with SFT (Supervised Fine-Tuning).
  • ์ด ๋ชจ๋ธ์€ "upstage/SOLAR-10.7B-v1.0" ๊ธฐ๋ฐ˜ ๋ชจ๋ธ์„ SFT๋ฅผ ์‚ฌ์šฉํ•˜์—ฌ ์ „์ฒด ํŒŒ๋ผ๋ฏธํ„ฐ ์กฐ์ • ๋ฐฉ๋ฒ•์œผ๋กœ ๋ฏธ์„ธ์กฐ์ •๋˜์—ˆ์Šต๋‹ˆ๋‹ค.

Hardware

  • Hardware: Utilized two A100 (80G*2EA) GPUs for training.
  • Training Factors: This model was fine-tuned with SFT, using the HuggingFace SFTtrainer and applied fsdp.
  • ์ด ๋ชจ๋ธ์€ SFT๋ฅผ ์‚ฌ์šฉํ•˜์—ฌ HuggingFace SFTtrainer์™€ fsdp๋ฅผ ์ ์šฉํ•˜์—ฌ ๋ฏธ์„ธ์กฐ์ •๋˜์—ˆ์Šต๋‹ˆ๋‹ค.
Downloads last month
2,406
Safetensors
Model size
10.7B params
Tensor type
FP16
ยท
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for Changgil/K2S3-SOLAR-11b-v3.0

Quantizations
4 models

Spaces using Changgil/K2S3-SOLAR-11b-v3.0 5