lemonhat/Qwen2.5-Coder-7B-Instruct-swe_30k_v2_tag5mini

swe_30k_v2_tag5mini

This model is a fine-tuned version of Qwen/Qwen2.5-Coder-7B-Instruct on the swe_30k_v2_tag5mini dataset. It achieves the following results on the evaluation set:

• Loss: 0.5374

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: 1e-05
• 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
0.6038	0.0646	100	0.6196
0.6503	0.1293	200	0.5960
0.5604	0.1939	300	0.5876
0.7042	0.2586	400	0.5773
0.6817	0.3232	500	0.5707
0.5783	0.3878	600	0.5644
0.5492	0.4525	700	0.5592
0.5269	0.5171	800	0.5534
0.6711	0.5818	900	0.5493
0.5468	0.6464	1000	0.5443
0.5499	0.7111	1100	0.5413
0.6856	0.7757	1200	0.5392
0.6469	0.8403	1300	0.5385
0.5113	0.9050	1400	0.5379
0.538	0.9696	1500	0.5374

Framework versions

• Transformers 4.46.1
• Pytorch 2.6.0+cu124
• Datasets 3.1.0
• Tokenizers 0.20.3