oh_scale_x2_compute_equal

This model is a fine-tuned version of meta-llama/Meta-Llama-3.1-8B on the mlfoundations-dev/oh-dcft-v1.3_no-curation_gpt-4o-mini_scale_2x dataset. It achieves the following results on the evaluation set:

• Loss: 0.9233

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: 8
• total_train_batch_size: 512
• 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: 8.0

Training results

Training Loss	Epoch	Step	Validation Loss
0.7631	0.9995	549	0.7639
0.7148	1.9991	1098	0.7494
0.6707	2.9986	1647	0.7487
0.6324	4.0	2197	0.7583
0.5947	4.9995	2746	0.7757
0.5504	5.9991	3295	0.8122
0.5002	6.9986	3844	0.8600
0.4505	7.9964	4392	0.9233

Framework versions

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