oh-dcft-v3.1-SN-405B-hacky-qwen

This model is a fine-tuned version of Qwen/Qwen2.5-7B on the mlfoundations-dev/oh-dcft-v3.1-SN-405B-hacky dataset. It achieves the following results on the evaluation set:

• Loss: 0.5737

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: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• gradient_accumulation_steps: 2
• total_train_batch_size: 128
• total_eval_batch_size: 64
• 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: constant
• num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss
0.5404	1.0	1341	0.5464
0.4804	2.0	2682	0.5512
0.4292	3.0	4023	0.5737

Framework versions

• Transformers 4.46.1
• Pytorch 2.3.0
• Datasets 3.1.0
• Tokenizers 0.20.3