gekun/my-awesome-model

MyAwesomeModel (step_1000)

This repository contains the best checkpoint (step_1000) of MyAwesomeModel, selected based on the highest overall eval_accuracy (0.712) across all training checkpoints.

Evaluation Results

Detailed evaluation results for all 15 benchmarks on the step_1000 checkpoint:

Benchmark	Score
text_classification	0.828
logical_reasoning	0.819
translation	0.804
sentiment_analysis	0.792
summarization	0.767
instruction_following	0.758
safety_evaluation	0.739
common_sense	0.736
reading_comprehension	0.700
knowledge_retrieval	0.676
code_generation	0.650
dialogue_generation	0.644
creative_writing	0.610
question_answering	0.607
math_reasoning	0.550

Overall weighted eval_accuracy: 0.710

Usage

from transformers import AutoModel
model = AutoModel.from_pretrained("gekun/my-awesome-model")