full-4e4-nodefs

This model is a fine-tuned version of deepseek-ai/deepseek-coder-6.7b-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 7.8778

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: 0.0004
• train_batch_size: 4
• eval_batch_size: 2
• seed: 42
• distributed_type: multi-GPU
• num_devices: 2
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• total_eval_batch_size: 4
• 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: 12
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.3747	0.2	3094	0.1750
0.3377	0.4	6188	0.1573
0.3179	0.6	9282	0.1492
0.3015	0.8	12376	0.1446
0.2916	1.0	15470	0.1394
0.2723	1.2	18564	0.1373
0.2765	1.4	21658	0.1329
0.2637	1.6	24752	0.1312
0.2648	1.8	27846	0.1286
0.2551	2.0	30940	0.1283
2.8869	2.2	34034	2.0487
6.2984	2.4	37128	nan
15.7169	2.6	40222	7.8778
15.7148	2.8	43316	7.8778
15.7185	3.0	46410	7.8778
15.7242	3.2	49504	7.8778
15.7111	3.4	52598	7.8778
15.7226	3.6	55692	7.8778
15.7121	3.8	58786	7.8778
15.7189	4.0	61880	7.8778
15.7218	4.2	64974	7.8778
15.7247	4.4	68068	7.8778
15.7286	4.6	71162	7.8778
15.7142	4.8	74256	7.8778
15.7126	5.0	77350	7.8778
15.7211	5.2	80444	7.8778
15.7267	5.4	83538	7.8778
15.7275	5.6	86632	7.8778
15.7127	5.8	89726	7.8778
15.7225	6.0	92820	7.8778
15.7121	6.2	95914	7.8778
15.724	6.4	99008	7.8778
15.7068	6.6	102102	7.8778
15.7184	6.8	105196	7.8778
15.7215	7.0	108290	7.8778
15.7447	7.2	111384	7.8778
15.6922	7.4	114478	7.8778
15.7276	7.6	117572	7.8778
15.7394	7.8	120666	7.8778
15.7261	8.0	123760	7.8778
15.7231	8.2	126854	7.8778
15.7186	8.4	129948	7.8778
15.7168	8.6	133042	7.8778
15.7139	8.8	136136	7.8778
15.7314	9.0	139230	7.8778
15.7223	9.2	142324	7.8778
15.7068	9.4	145418	7.8778
15.6915	9.6	148512	7.8778
15.7315	9.8	151606	7.8778
15.7142	10.0	154700	7.8778
15.7252	10.2	157794	7.8778
15.7295	10.4	160888	7.8778
15.7258	10.6	163982	7.8778
15.7162	10.8	167076	7.8778
15.7111	11.0	170170	7.8778
15.7289	11.2	173264	7.8778
15.7031	11.4	176358	7.8778
15.7102	11.6	179452	7.8778
15.7186	11.8	182546	7.8778
15.7179	12.0	185640	7.8778

Framework versions

• PEFT 0.14.0
• Transformers 4.47.0
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.1