This model is a fine-tuned version of Qwen/Qwen2.5-0.5B on the lunar_lander_270_reward_train dataset. It achieves the following results on the evaluation set:

Loss: 0.0253
Accuracy: 0.9905

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
gradient_accumulation_steps: 32
total_train_batch_size: 32
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: 100.0
num_epochs: 1.0
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.1271	0.0327	200	0.1269	0.9408
0.1134	0.0653	400	0.1141	0.9496
0.0951	0.0980	600	0.0991	0.9562
0.0866	0.1306	800	0.0930	0.9564
0.0909	0.1633	1000	0.0883	0.9608
0.0980	0.1959	1200	0.0856	0.963
0.0836	0.2286	1400	0.0839	0.9611
0.0709	0.2612	1600	0.0811	0.9644
0.0757	0.2939	1800	0.0765	0.9665
0.0701	0.3265	2000	0.0791	0.9655
0.0717	0.3592	2200	0.0681	0.9706
0.0637	0.3918	2400	0.0704	0.9686
0.0619	0.4245	2600	0.0618	0.9747
0.0572	0.4571	2800	0.0562	0.9774
0.0605	0.4898	3000	0.0561	0.9767
0.0648	0.5224	3200	0.0584	0.9745
0.0659	0.5551	3400	0.0540	0.9771
0.0564	0.5878	3600	0.0479	0.9799
0.0545	0.6204	3800	0.0432	0.9819
0.0468	0.6531	4000	0.0449	0.9817
0.0388	0.6857	4200	0.0395	0.9840
0.0328	0.7184	4400	0.0397	0.9831
0.0363	0.7510	4600	0.0357	0.9856
0.0314	0.7837	4800	0.0362	0.985
0.0280	0.8163	5000	0.0336	0.9861
0.0308	0.8490	5200	0.0292	0.9882
0.0252	0.8816	5400	0.0268	0.9894
0.0202	0.9143	5600	0.0258	0.9901
0.0252	0.9469	5800	0.0255	0.9902
0.0201	0.9796	6000	0.0253	0.9904

Framework versions

Transformers 5.2.0
Pytorch 2.11.0+cu130
Datasets 4.0.0
Tokenizers 0.22.2

alirizaercan/qwen25_05b_base_full_ft_lunarlander_a4000

qwen25_05b_base_full_ft_lunarlander_a4000

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

alirizaercan/
qwen25_05b_base_full_ft_lunarlander_a4000