lemexp-task1-min_symbols_template_full-deepseek-coder-1.3b-base-ddp

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: 0.1978

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: 0.001
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• total_train_batch_size: 16
• total_eval_batch_size: 16
• 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
• num_epochs: 6
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.3666	0.2000	2907	0.3499
0.3395	0.4001	5814	0.3318
0.3267	0.6001	8721	0.3205
0.3174	0.8001	11628	0.3100
0.305	1.0001	14535	0.3098
0.3004	1.2002	17442	0.2988
0.3017	1.4002	20349	0.2947
0.2912	1.6002	23256	0.2879
0.2843	1.8002	26163	0.2795
0.2793	2.0003	29070	0.2770
0.27	2.2003	31977	0.2737
0.2697	2.4003	34884	0.2725
0.2641	2.6004	37791	0.2652
0.2586	2.8004	40698	0.2608
0.2564	3.0004	43605	0.2545
0.2464	3.2004	46512	0.2516
0.2428	3.4005	49419	0.2442
0.2395	3.6005	52326	0.2403
0.234	3.8005	55233	0.2367
0.2299	4.0006	58140	0.2311
0.2178	4.2006	61047	0.2271
0.2163	4.4006	63954	0.2227
0.2114	4.6006	66861	0.2175
0.2083	4.8007	69768	0.2150
0.2026	5.0007	72675	0.2100
0.1907	5.2007	75582	0.2065
0.1893	5.4007	78489	0.2020
0.1874	5.6008	81396	0.1982
0.1812	5.8008	84303	0.1978

Framework versions

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