lemonhat/Qwen2.5-7B-gaia_5k_v3

gaia_5k_v3

This model is a fine-tuned version of Qwen/Qwen2.5-7B on the gaia_5k_v3 dataset. It achieves the following results on the evaluation set:

• Loss: 0.3064

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-06
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 4
• total_eval_batch_size: 4
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss
0.4184	0.1099	100	0.4009
0.525	0.2198	200	0.3417
0.3922	0.3297	300	0.3336
0.3402	0.4396	400	0.3218
0.2915	0.5495	500	0.3122
0.3677	0.6593	600	0.3102
0.2704	0.7692	700	0.3136
0.3887	0.8791	800	0.3077
0.4175	0.9890	900	0.3071

Framework versions

• Transformers 4.46.1
• Pytorch 2.7.1+cu126
• Datasets 3.1.0
• Tokenizers 0.20.3