lemexp-processed-task1_min_symbols_lemma_command_small-deepseek-coder-1.3b-base

This model is a fine-tuned version of deepseek-ai/deepseek-coder-1.3b-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4329

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: 0.0002
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 6
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.6364	0.2000	3683	0.6357
0.5857	0.4001	7366	0.5827
0.5682	0.6001	11049	0.5516
0.5421	0.8001	14732	0.5293
0.5142	1.0002	18415	0.5177
0.4674	1.2002	22098	0.5015
0.4615	1.4002	25781	0.5000
0.453	1.6003	29464	0.4770
0.4506	1.8003	33147	0.4701
0.4309	2.0003	36830	0.4646
0.3829	2.2004	40513	0.4667
0.3925	2.4004	44196	0.4595
0.3858	2.6004	47879	0.4566
0.3879	2.8005	51562	0.4439
0.3764	3.0005	55245	0.4379
0.3267	3.2005	58928	0.4502
0.3346	3.4006	62611	0.4443
0.3363	3.6006	66294	0.4339
0.3321	3.8006	69977	0.4350
0.3423	4.0007	73660	0.4288
0.2789	4.2007	77343	0.4458
0.2928	4.4007	81026	0.4379
0.2963	4.6007	84709	0.4325
0.2887	4.8008	88392	0.4275
0.2949	5.0008	92075	0.4292
0.2437	5.2008	95758	0.4366
0.2424	5.4009	99441	0.4358
0.2528	5.6009	103124	0.4331
0.2477	5.8009	106807	0.4329

Framework versions

• PEFT 0.14.0
• Transformers 4.47.0
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0