mini_llama_crafting_sft_success_new_mem

This model is a fine-tuned version of meta-llama/Llama-3.2-1B on the identity and the crafting_sft_success_new_mem datasets. It achieves the following results on the evaluation set:

• Loss: 0.4032

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: 8
• gradient_accumulation_steps: 16
• total_train_batch_size: 128
• total_eval_batch_size: 8
• 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
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss
0.8427	0.3380	50	1.1575
0.5411	0.6760	100	0.5065
0.519	1.0203	150	0.4361
0.3662	1.3583	200	0.4007
0.3679	1.6962	250	0.3948
0.3176	2.0406	300	0.3846
0.2141	2.3785	350	0.4076
0.2089	2.7165	400	0.3996

Framework versions

• Transformers 4.49.0
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0