qwen-3b-sft-full

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

• Loss: 1.1583

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: 2e-05
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• total_eval_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss
1.18	0.1046	100	1.1877
1.1739	0.2092	200	1.1825
1.1693	0.3138	300	1.1775
1.1555	0.4184	400	1.1726
1.165	0.5230	500	1.1683
1.1517	0.6276	600	1.1640
1.1712	0.7322	700	1.1609
1.1535	0.8368	800	1.1590
1.1528	0.9414	900	1.1583

Framework versions

• Transformers 4.45.2
• Pytorch 2.5.1+rocm6.2
• Datasets 3.2.0
• Tokenizers 0.20.3