gptoss_20b_all_zhtw_lr5e-7_ep3_16_64_128_turn_dup

This model is a fine-tuned version of openai/gpt-oss-20b on the all_dupcleaned_data_turn_gpt_oss_20b_pretokenized, the new_irrelevance_zh_tw_oss3000_oss_gpt_oss_20b_pretokenized and the apigen_zhtwV3_remove_sys_gpt_oss_20b_pretokenized datasets. It achieves the following results on the evaluation set:

• Loss: 0.9870

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: 5e-07
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 128
• total_train_batch_size: 128
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 50
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss
1.3199	0.3888	100	1.9226
1.2301	0.7777	200	1.6118
0.9273	1.1633	300	1.3543
1.0135	1.5522	400	1.1740
1.0803	1.9410	500	1.0577
1.0983	2.3266	600	1.0026
0.9965	2.7155	700	0.9903

Framework versions

• PEFT 0.18.1
• Transformers 4.57.6
• Pytorch 2.8.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2