CCsky123/Qwen2.5-7B-Instruct-Eng-ContinueWriting-Optimizer

train_2025-06-07-18-50-03

This model is a fine-tuned version of Qwen/Qwen2.5-7B-Instruct on the 你的数据集 dataset. It achieves the following results on the evaluation set:

• Loss: 2.7113
• Num Input Tokens Seen: 2649088

Model description

This model is a fine-tuned LoRA adapter based on Qwen2.5-7B-Instruct , designed to optimize continuation writing tasks for high school students .

Intended uses & limitations

Inputs:

1. Writing prompt (the given story segment to continue)
2. • Prompt analysis (generated by CCsky123/Qwen2.5-7B-Instruct-Eng-ContinueWriting-Analyzer)
3. Student's essay
4. Essay analysis (also generated by the Analyzer model)

Outputs: An improved version of the student's essay, with enhanced coherence, language use, and logical flow while preserving the original intent.

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• num_epochs: 6.0

Training results

Training Loss	Epoch	Step	Validation Loss	Input Tokens Seen
0.9381	3.6036	100	2.3569	1635328

Framework versions

• PEFT 0.12.0
• Transformers 4.45.2
• Pytorch 2.4.1+cu121
• Datasets 2.21.0
• Tokenizers 0.20.1