Edit model card

image/png

adpater_config.json์˜ ์—ญํ• : adapter_config.json์€ ์–ด๋Œ‘ํ„ฐ ๊ธฐ๋ฐ˜ ๋ชจ๋ธ ํŠธ๋ ˆ์ด๋‹์— ์‚ฌ์šฉ๋˜๋Š” ์„ค์ • ํŒŒ์ผ์ž…๋‹ˆ๋‹ค. ์–ด๋Œ‘ํ„ฐ๋Š” ์‹ ๊ฒฝ๋ง ๋ชจ๋ธ์—์„œ ๊ธฐ์กด์˜ ๋ ˆ์ด์–ด ๊ตฌ์กฐ๋ฅผ ๋ณ€๊ฒฝํ•˜์ง€ ์•Š๊ณ  ์ƒˆ๋กœ์šด ์ž‘์—…์ด๋‚˜ ๋„๋ฉ”์ธ์— ๋Œ€ํ•ด ๋ชจ๋ธ์„ ๋ฏธ์„ธ ์กฐ์ •ํ•˜๋Š” ๊ธฐ์ˆ ์ž…๋‹ˆ๋‹ค. ์ด ํŒŒ์ผ์€ ๋ ˆ์ด์–ด์— ์ถ”๊ฐ€๋˜๋Š” ์–ด๋Œ‘ํ„ฐ์˜ ์ˆ˜, ํฌ๊ธฐ, ํ™œ์„ฑํ™” ํ•จ์ˆ˜, ํ•™์Šต๋ฅ  ๋“ฑ๊ณผ ๊ฐ™์€ ์ค‘์š”ํ•œ ํ•˜์ดํผํŒŒ๋ผ๋ฏธํ„ฐ๋ฅผ ์ •์˜ํ•ฉ๋‹ˆ๋‹ค. ์–ด๋Œ‘ํ„ฐ๋ฅผ ์‚ฌ์šฉํ•˜๋ฉด ์ „์ฒด ๋ชจ๋ธ์„ ์žฌํ•™์Šตํ•  ํ•„์š” ์—†์ด ํšจ์œจ์ ์œผ๋กœ ๋ฏธ์„ธ ์กฐ์ •ํ•  ์ˆ˜ ์žˆ์–ด, ๊ณ„์‚ฐ ๋น„์šฉ๊ณผ ์‹œ๊ฐ„์„ ์ ˆ์•ฝํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์ด๋Ÿฐ ๋ฐฉ์‹์€ ํŠนํžˆ ํฐ ๋ชจ๋ธ์— ์œ ์šฉํ•˜๋ฉฐ, ๋‹ค์–‘ํ•œ ์ž‘์—…์— ํ•˜๋‚˜์˜ ๋ชจ๋ธ์„ ์ ์šฉํ•  ์ˆ˜ ์žˆ๋Š” ์œ ์—ฐ์„ฑ์„ ์ œ๊ณตํ•ฉ๋‹ˆ๋‹ค.

ํ…Œ์ŠคํŠธ ๋ฐ์ดํ„ฐ์— ๋Œ€ํ•œ ์„ค๋ช…: ํ…Œ์ŠคํŠธ ๋ฐ์ดํ„ฐ์…‹ 'nsmc'๋Š” ํ•œ๊ตญ์–ด ์˜ํ™” ๋ฆฌ๋ทฐ ๋ฐ์ดํ„ฐ์…‹์œผ๋กœ, ๊ฐ์„ฑ ๋ถ„์„ ์ž‘์—…์— ์ž์ฃผ ์‚ฌ์šฉ๋ฉ๋‹ˆ๋‹ค. ์ด ๋ฐ์ดํ„ฐ์…‹์€ ๋ฆฌ๋ทฐ ํ…์ŠคํŠธ์™€ ํ•ด๋‹น ๋ฆฌ๋ทฐ๊ฐ€ ๊ธ์ •์ ์ธ์ง€ ๋ถ€์ •์ ์ธ์ง€๋ฅผ ๋‚˜ํƒ€๋‚ด๋Š” ๋ผ๋ฒจ๋กœ ๊ตฌ์„ฑ๋˜์–ด ์žˆ์Šต๋‹ˆ๋‹ค. ์ด ์ฝ”๋“œ์—์„œ๋Š” ์ด ๋ฐ์ดํ„ฐ์…‹์˜ ์ผ๋ถ€๋ฅผ ์‚ฌ์šฉํ•˜์—ฌ ๋ชจ๋ธ์„ ํ…Œ์ŠคํŠธํ•ฉ๋‹ˆ๋‹ค. ํ…Œ์ŠคํŠธ ๋ฐ์ดํ„ฐ์…‹์„ ์‚ฌ์šฉํ•˜๋Š” ๋ชฉ์ ์€ ํ›ˆ๋ จ๋œ ๋ชจ๋ธ์ด ์‹ค์ œ ์„ธ๊ณ„ ๋ฐ์ดํ„ฐ์— ๋Œ€ํ•ด ์–ผ๋งˆ๋‚˜ ์ž˜ ์ž‘๋™ํ•˜๋Š”์ง€๋ฅผ ํ‰๊ฐ€ํ•˜๋Š” ๊ฒƒ์ž…๋‹ˆ๋‹ค. ์ด ๊ณผ์ •์—์„œ ๋ชจ๋ธ์˜ ์ผ๋ฐ˜ํ™” ๋Šฅ๋ ฅ๊ณผ ๊ฐ•์ธ์„ฑ์„ ๊ฒ€์ฆํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

์ด ์ฝ”๋“œ์—์„œ ์‚ฌ์šฉ๋œ ํ…Œ์ŠคํŠธ ์กฐ๊ฑด์€ ๋‹ค์Œ๊ณผ ๊ฐ™์Šต๋‹ˆ๋‹ค:

๋ฐ์ดํ„ฐ์…‹๊ณผ ์ƒ˜ํ”Œ ํฌ๊ธฐ:

์‚ฌ์šฉ๋œ ๋ฐ์ดํ„ฐ์…‹์€ 'nsmc'๋กœ, ํ•œ๊ตญ์–ด ์˜ํ™” ๋ฆฌ๋ทฐ๋ฅผ ๋‹ด๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. ์ด ๋ฐ์ดํ„ฐ์…‹์—์„œ 1000๊ฐœ์˜ ํ…Œ์ŠคํŠธ ์ƒ˜ํ”Œ์„ ์‚ฌ์šฉํ–ˆ์Šต๋‹ˆ๋‹ค. ๊ฐ ์ƒ˜ํ”Œ์€ ํ…์ŠคํŠธ ๋ฆฌ๋ทฐ์™€ ์ด ๋ฆฌ๋ทฐ๊ฐ€ ๊ธ์ •์ ์ธ์ง€ ๋ถ€์ •์ ์ธ์ง€๋ฅผ ๋‚˜ํƒ€๋‚ด๋Š” ๋ผ๋ฒจ๋กœ ๊ตฌ์„ฑ๋ฉ๋‹ˆ๋‹ค. ์ž…๋ ฅ ๋ฐ์ดํ„ฐ์˜ ์ฒ˜๋ฆฌ:

ํ…Œ์ŠคํŠธ ๊ณผ์ •์—์„œ ๊ฐ ๋ฆฌ๋ทฐ ํ…์ŠคํŠธ๋Š” ๋ชจ๋ธ์— ์ž…๋ ฅ๋˜๊ธฐ ์ „์— ํŠน์ • ํฌ๋งท์œผ๋กœ ๋ณ€ํ™˜๋ฉ๋‹ˆ๋‹ค. ์ด ํฌ๋งท์€ ๋ชจ๋ธ์ด ์ดํ•ดํ•  ์ˆ˜ ์žˆ๋„๋ก ํ…์ŠคํŠธ๋ฅผ ๊ตฌ์กฐํ™”ํ•˜๋Š” ๋ฐฉ๋ฒ•์— ๋”ฐ๋ผ ๋‹ฌ๋ผ์งˆ ์ˆ˜ ์žˆ์œผ๋ฉฐ, ์ผ๋ฐ˜์ ์œผ๋กœ ํ† ํฐํ™” ๊ณผ์ •์„ ํฌํ•จํ•ฉ๋‹ˆ๋‹ค. ๋ชจ๋ธ์˜ ์˜ˆ์ธก๊ณผ ํ‰๊ฐ€ ๋ฐฉ๋ฒ•:

๋ชจ๋ธ์€ ๊ฐ ํ…Œ์ŠคํŠธ ์ƒ˜ํ”Œ์— ๋Œ€ํ•ด ๊ธ์ •์  ๋˜๋Š” ๋ถ€์ •์ ์ธ ๊ฐ์„ฑ์„ ์˜ˆ์ธกํ•ฉ๋‹ˆ๋‹ค. ์ด ์˜ˆ์ธก์€ ์‹ค์ œ ๋ผ๋ฒจ(๊ธ์ •/๋ถ€์ •)๊ณผ ๋น„๊ต๋˜์–ด ๋ชจ๋ธ์˜ ์„ฑ๋Šฅ์„ ํ‰๊ฐ€ํ•ฉ๋‹ˆ๋‹ค. ์„ฑ๋Šฅ ํ‰๊ฐ€ ์ง€ํ‘œ:

ํ…Œ์ŠคํŠธ ๊ณผ์ •์—์„œ ์ฃผ๋กœ ์‚ฌ์šฉ๋œ ์„ฑ๋Šฅ ํ‰๊ฐ€ ์ง€ํ‘œ๋Š” ์ •ํ™•๋„(accuracy)์ž…๋‹ˆ๋‹ค. ์ •ํ™•๋„๋Š” ๋ชจ๋ธ์ด ์˜ฌ๋ฐ”๋ฅด๊ฒŒ ์˜ˆ์ธกํ•œ ์ƒ˜ํ”Œ์˜ ๋น„์œจ๋กœ ๊ณ„์‚ฐ๋ฉ๋‹ˆ๋‹ค. ์„ฑ๋Šฅ ํ‰๊ฐ€๋ฅผ ์œ„ํ•œ ๊ณ„์‚ฐ ๋ฐฉ๋ฒ•:

์ฝ”๋“œ๋Š” True Positive (TP), True Negative (TN), False Positive (FP), False Negative (FN)๋ฅผ ๊ณ„์‚ฐํ•˜์—ฌ ๋ชจ๋ธ์˜ ์ •ํ™•๋„๋ฅผ ๋„์ถœํ•ฉ๋‹ˆ๋‹ค. ์ด ๊ฐ’๋“ค์€ ๋ชจ๋ธ์ด ์‹ค์ œ ๊ธ์ •(positive) ๋ผ๋ฒจ์„ ๊ธ์ •์œผ๋กœ, ์‹ค์ œ ๋ถ€์ •(negative) ๋ผ๋ฒจ์„ ๋ถ€์ •์œผ๋กœ ์ •ํ™•ํžˆ ์˜ˆ์ธกํ•œ ๊ฒฝ์šฐ์™€ ๊ทธ๋ ‡์ง€ ๋ชปํ•œ ๊ฒฝ์šฐ๋ฅผ ๋‚˜ํƒ€๋ƒ…๋‹ˆ๋‹ค.

ํ…Œ์ŠคํŠธ ์กฐ๊ฑด ๋ฐ ๋ฐฉ๋ฒ•๋ก : ๋ชจ๋ธ ํ…Œ์ŠคํŠธ๋Š” ์ฃผ์–ด์ง„ ์ž…๋ ฅ์— ๋Œ€ํ•œ ๋ชจ๋ธ์˜ ์˜ˆ์ธก๊ณผ ์‹ค์ œ ๋ผ๋ฒจ์„ ๋น„๊ตํ•˜๋Š” ๊ณผ์ •์ž…๋‹ˆ๋‹ค. ์ด ์ฝ”๋“œ์—์„œ๋Š” True Positive (TP), True Negative (TN), False Positive (FP), False Negative (FN)์˜ ๋„ค ๊ฐ€์ง€ ๊ธฐ๋ณธ ์ง€ํ‘œ๋ฅผ ์‚ฌ์šฉํ•˜์—ฌ ๋ชจ๋ธ์˜ ์˜ˆ์ธก ์„ฑ๋Šฅ์„ ํ‰๊ฐ€ํ•ฉ๋‹ˆ๋‹ค. TP์™€ TN์€ ๋ชจ๋ธ์ด ์˜ฌ๋ฐ”๋ฅด๊ฒŒ ์˜ˆ์ธกํ•œ ๊ฒฝ์šฐ, FP์™€ FN์€ ์ž˜๋ชป ์˜ˆ์ธกํ•œ ๊ฒฝ์šฐ๋ฅผ ๋‚˜ํƒ€๋ƒ…๋‹ˆ๋‹ค. ์ด๋Ÿฌํ•œ ์ง€ํ‘œ๋“ค์€ ๋ชจ๋ธ์˜ ์ •ํ™•๋„๋ฟ๋งŒ ์•„๋‹ˆ๋ผ ๋‹ค๋ฅธ ์ค‘์š”ํ•œ ์„ฑ๋Šฅ ์ง€ํ‘œ๋“ค(์˜ˆ: ์ •๋ฐ€๋„, ์žฌํ˜„์œจ)์„ ๊ณ„์‚ฐํ•˜๋Š” ๋ฐ์—๋„ ์‚ฌ์šฉ๋ฉ๋‹ˆ๋‹ค.

๊ฒฐ๊ณผ ํ•ด์„ ์ •ํ™•๋„ 87.8%: ์ด๋Š” ์ „์ฒด ํ…Œ์ŠคํŠธ ๋ฐ์ดํ„ฐ ์ค‘ ์•ฝ 87.8%๋ฅผ ๋ชจ๋ธ์ด ์˜ฌ๋ฐ”๋ฅด๊ฒŒ ๋ถ„๋ฅ˜ํ–ˆ๋‹ค๋Š” ๊ฒƒ์„ ์˜๋ฏธํ•ฉ๋‹ˆ๋‹ค. ์ด๋Š” ์ƒ๋‹นํžˆ ๋†’์€ ์ •ํ™•๋„์ด๋ฉฐ, ๋ชจ๋ธ์ด ๋Œ€๋ถ€๋ถ„์˜ ๋ฆฌ๋ทฐ๋ฅผ ์˜ฌ๋ฐ”๋ฅด๊ฒŒ ๊ฐ์„ฑ ๋ถ„์„ํ–ˆ์Œ์„ ๋‚˜ํƒ€๋ƒ…๋‹ˆ๋‹ค.

True Positives์™€ True Negatives: ๋†’์€ TP์™€ TN์€ ๋ชจ๋ธ์ด ๊ธ์ •์  ๋ฐ ๋ถ€์ •์  ๋ฆฌ๋ทฐ๋ฅผ ๋Œ€์ฒด๋กœ ์ž˜ ๋ถ„๋ฅ˜ํ•˜๊ณ  ์žˆ์Œ์„ ๋‚˜ํƒ€๋ƒ…๋‹ˆ๋‹ค.

False Positives์™€ False Negatives: ๋น„๊ต์  ๋‚ฎ์€ FP์™€ ๋†’์€ FN์€ ๋ชจ๋ธ์ด ๋ถ€์ •์  ๋ฆฌ๋ทฐ๋ฅผ ๊ธ์ •์œผ๋กœ ์ž˜๋ชป ๋ถ„๋ฅ˜ํ•  ๊ฐ€๋Šฅ์„ฑ์ด ๋‚ฎ์ง€๋งŒ, ๊ธ์ •์  ๋ฆฌ๋ทฐ๋ฅผ ๋ถ€์ •์œผ๋กœ ์ž˜๋ชป ๋ถ„๋ฅ˜ํ•˜๋Š” ๊ฒฝํ–ฅ์ด ์žˆ์Œ์„ ๋‚˜ํƒ€๋ƒ…๋‹ˆ๋‹ค. ์ด๋Š” ๋ชจ๋ธ์ด ๊ธ์ •์  ํ‘œํ˜„์„ ๋ถ€์ •์ ์œผ๋กœ ์˜คํ•ดํ•˜๋Š” ๊ฒฝํ–ฅ์ด ์žˆ์„ ์ˆ˜ ์žˆ์Œ์„ ์˜๋ฏธํ•ฉ๋‹ˆ๋‹ค.

์„ฑ๋Šฅ ํ–ฅ์ƒ์„ ์œ„ํ•œ ๊ณ ๋ ค ์‚ฌํ•ญ ๋ฐ์ดํ„ฐ ๋ถˆ๊ท ํ˜•: ๋ฐ์ดํ„ฐ์…‹์— ๊ธ์ • ๋˜๋Š” ๋ถ€์ • ๋ฆฌ๋ทฐ์˜ ๋ถˆ๊ท ํ˜•์ด ์žˆ์„ ๊ฒฝ์šฐ, ์ด๋Š” ๋ชจ๋ธ์˜ ํ•™์Šต์— ์˜ํ–ฅ์„ ๋ฏธ์น  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ๊ท ํ˜• ์žกํžŒ ๋ฐ์ดํ„ฐ์…‹์„ ์‚ฌ์šฉํ•˜๊ฑฐ๋‚˜, ๊ฐ€์ค‘์น˜ ์กฐ์ •๊ณผ ๊ฐ™์€ ๊ธฐ๋ฒ•์„ ํ†ตํ•ด ์ด๋ฅผ ๋ณด์ •ํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

Model Card for Model ID

Model Details

Model Description

  • **Model type: Causal Language Model
  • **Language(s) (NLP): Korean
  • **Finetuned from model [optional]: Finetuned from 'KT-AI/midm-bitext-S-7B-inst-v1'

Uses

Direct Use

This model is primarily used for sentiment analysis on Korean text, particularly in classifying movie reviews as positive or negative.

Downstream Use [optional]

The model can be adapted for other types of Korean text classification tasks such as customer feedback analysis, social media sentiment analysis, etc.

Out-of-Scope Use

Bias, Risks, and Limitations

This model, while performing with high accuracy, may exhibit biases present in the training data, potentially leading to skewed results in certain scenarios. Further evaluation and monitoring are recommended to identify and mitigate these biases.

Recommendations

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.

How to Get Started with the Model

Use the code below to get started with the model.

[More Information Needed]

Training Details

Training Data

The model was fine-tuned on the NSMC (Naver Sentiment Movie Corpus) dataset, consisting of Korean movie reviews with binary sentiment labels.

Training Procedure

Text data were tokenized using a Korean-specific tokenizer. Standard preprocessing steps such as lowercasing and removal of special characters were applied.

Preprocessing [optional]

[More Information Needed]

Training Hyperparameters

  • **Training regime:**Learning Rate: 1e-4 Batch Size: 8 Epochs: 3

Speeds, Sizes, Times [optional]

[More Information Needed]

Evaluation

Testing Data, Factors & Metrics

Testing Data

The model was evaluated on a separate test set extracted from the NSMC dataset, ensuring no overlap with the training data.

Factors

The evaluation focused on the model's ability to accurately classify sentiment in Korean movie reviews.

Metrics

Metrics used include Accuracy, Precision

Results

The model achieved an accuracy of 87.8%

Summary

๊ฒฐ๊ณผ ํ•ด์„ ์ •ํ™•๋„ 87.8%: ์ด๋Š” ์ „์ฒด ํ…Œ์ŠคํŠธ ๋ฐ์ดํ„ฐ ์ค‘ ์•ฝ 87.8%๋ฅผ ๋ชจ๋ธ์ด ์˜ฌ๋ฐ”๋ฅด๊ฒŒ ๋ถ„๋ฅ˜ํ–ˆ๋‹ค๋Š” ๊ฒƒ์„ ์˜๋ฏธํ•ฉ๋‹ˆ๋‹ค. ์ด๋Š” ์ƒ๋‹นํžˆ ๋†’์€ ์ •ํ™•๋„์ด๋ฉฐ, ๋ชจ๋ธ์ด ๋Œ€๋ถ€๋ถ„์˜ ๋ฆฌ๋ทฐ๋ฅผ ์˜ฌ๋ฐ”๋ฅด๊ฒŒ ๊ฐ์„ฑ ๋ถ„์„ํ–ˆ์Œ์„ ๋‚˜ํƒ€๋ƒ…๋‹ˆ๋‹ค.

True Positives์™€ True Negatives: ๋†’์€ TP์™€ TN์€ ๋ชจ๋ธ์ด ๊ธ์ •์  ๋ฐ ๋ถ€์ •์  ๋ฆฌ๋ทฐ๋ฅผ ๋Œ€์ฒด๋กœ ์ž˜ ๋ถ„๋ฅ˜ํ•˜๊ณ  ์žˆ์Œ์„ ๋‚˜ํƒ€๋ƒ…๋‹ˆ๋‹ค.

False Positives์™€ False Negatives: ๋น„๊ต์  ๋‚ฎ์€ FP์™€ ๋†’์€ FN์€ ๋ชจ๋ธ์ด ๋ถ€์ •์  ๋ฆฌ๋ทฐ๋ฅผ ๊ธ์ •์œผ๋กœ ์ž˜๋ชป ๋ถ„๋ฅ˜ํ•  ๊ฐ€๋Šฅ์„ฑ์ด ๋‚ฎ์ง€๋งŒ, ๊ธ์ •์  ๋ฆฌ๋ทฐ๋ฅผ ๋ถ€์ •์œผ๋กœ ์ž˜๋ชป ๋ถ„๋ฅ˜ํ•˜๋Š” ๊ฒฝํ–ฅ์ด ์žˆ์Œ์„ ๋‚˜ํƒ€๋ƒ…๋‹ˆ๋‹ค. ์ด๋Š” ๋ชจ๋ธ์ด ๊ธ์ •์  ํ‘œํ˜„์„ ๋ถ€์ •์ ์œผ๋กœ ์˜คํ•ดํ•˜๋Š” ๊ฒฝํ–ฅ์ด ์žˆ์„ ์ˆ˜ ์žˆ์Œ์„ ์˜๋ฏธํ•ฉ๋‹ˆ๋‹ค.

์„ฑ๋Šฅ ํ–ฅ์ƒ์„ ์œ„ํ•œ ๊ณ ๋ ค ์‚ฌํ•ญ ๋ฐ์ดํ„ฐ ๋ถˆ๊ท ํ˜•: ๋ฐ์ดํ„ฐ์…‹์— ๊ธ์ • ๋˜๋Š” ๋ถ€์ • ๋ฆฌ๋ทฐ์˜ ๋ถˆ๊ท ํ˜•์ด ์žˆ์„ ๊ฒฝ์šฐ, ์ด๋Š” ๋ชจ๋ธ์˜ ํ•™์Šต์— ์˜ํ–ฅ์„ ๋ฏธ์น  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ๊ท ํ˜• ์žกํžŒ ๋ฐ์ดํ„ฐ์…‹์„ ์‚ฌ์šฉํ•˜๊ฑฐ๋‚˜, ๊ฐ€์ค‘์น˜ ์กฐ์ •๊ณผ ๊ฐ™์€ ๊ธฐ๋ฒ•์„ ํ†ตํ•ด ์ด๋ฅผ ๋ณด์ •ํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

Model Examination [optional]

[More Information Needed]

Environmental Impact

Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).

  • Hardware Type: [More Information Needed]
  • Hours used: [More Information Needed]
  • Cloud Provider: [More Information Needed]
  • Compute Region: [More Information Needed]
  • Carbon Emitted: [More Information Needed]

Technical Specifications [optional]

Model Architecture and Objective

[More Information Needed]

Compute Infrastructure

[More Information Needed]

Hardware

[More Information Needed]

Software

[More Information Needed]

Citation [optional]

BibTeX:

[More Information Needed]

APA:

[More Information Needed]

Glossary [optional]

[More Information Needed]

More Information [optional]

[More Information Needed]

Model Card Authors [optional]

[More Information Needed]

Model Card Contact

[More Information Needed]

Training procedure

The following bitsandbytes quantization config was used during training:

  • quant_method: bitsandbytes
  • load_in_8bit: False
  • load_in_4bit: True
  • llm_int8_threshold: 6.0
  • llm_int8_skip_modules: None
  • llm_int8_enable_fp32_cpu_offload: False
  • llm_int8_has_fp16_weight: False
  • bnb_4bit_quant_type: nf4
  • bnb_4bit_use_double_quant: False
  • bnb_4bit_compute_dtype: bfloat16

Framework versions

  • PEFT 0.7.0
Downloads last month
2
Inference API
Unable to determine this modelโ€™s pipeline type. Check the docs .

Model tree for cheonyumin/lora-midm-7b-food-order-understanding

Adapter
(16)
this model