llama-factory/saves/Qwen2.5-3B-Instruct/README.md

metadata

license: other
library_name: peft
tags:
  - llama-factory
  - lora
  - generated_from_trainer
base_model: Qwen/Qwen2.5-3B-Instruct
model-index:
  - name: Qwen2.5-3B-Instruct
    results: []

Qwen2.5-3B-Instruct

This model is a fine-tuned version of Qwen/Qwen2.5-3B-Instruct on the alpaca_mgtv_p2 dataset. It achieves the following results on the evaluation set:

• Loss: 0.2272

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0001
• train_batch_size: 16
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 2.0

Training results

Training Loss	Epoch	Step	Validation Loss
0.3533	0.1990	35	0.3218
0.2926	0.3980	70	0.2747
0.277	0.5970	105	0.2519
0.2691	0.7960	140	0.2480
0.2597	0.9950	175	0.2401
0.2218	1.1940	210	0.2376
0.2594	1.3930	245	0.2271
0.2232	1.5920	280	0.2287
0.2145	1.7910	315	0.2291
0.2284	1.9900	350	0.2272

Framework versions

• PEFT 0.11.1
• Transformers 4.43.3
• Pytorch 2.4.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1