This model is a fine-tuned version of W-61/llama-3-8b-base-sft-ultrachat-8xh200 on the HuggingFaceH4/ultrafeedback_binarized dataset. It achieves the following results on the evaluation set:

Loss: 0.5352
Fcm Dpo/beta: 0.0110
Margin Dpo/margin Mean: 54.2375
Margin Dpo/margin Std: 83.0790
Logps/chosen: -383.9891
Logps/rejected: -417.3312
Logps/ref Chosen: -287.8268
Logps/ref Rejected: -266.9314
Logits/chosen: -0.8407
Logits/rejected: -0.8346

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-07
train_batch_size: 4
eval_batch_size: 4
seed: 42
distributed_type: multi-GPU
num_devices: 4
gradient_accumulation_steps: 8
total_train_batch_size: 128
total_eval_batch_size: 16
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: cosine
lr_scheduler_warmup_ratio: 0.1
num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Fcm Dpo/beta	Margin Dpo/margin Mean	Margin Dpo/margin Std	Logps/chosen	Logps/rejected	Logps/ref Chosen	Logps/ref Rejected	Logits/chosen	Logits/rejected
4.6087	0.4188	200	0.5497	0.0189	29.5673	48.3802	-320.8928	-329.5647	-287.8268	-266.9314	-0.8679	-0.8610
4.3167	0.8377	400	0.5352	0.0110	54.2375	83.0790	-383.9891	-417.3312	-287.8268	-266.9314	-0.8407	-0.8346

Framework versions

Transformers 4.51.0
Pytorch 2.3.1+cu121
Datasets 2.21.0
Tokenizers 0.21.4

W-61/llama3-8b-base-new-method-s_star0.6-20260425-180936

llama3-8b-base-new-method-s_star0.6-20260425-180936

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

W-61/
llama3-8b-base-new-method-s_star0.6-20260425-180936