This model is a fine-tuned version of JunxiongWang/llama3_mamba_0_5_sft on the HuggingFaceH4/ultrafeedback_binarized, the HuggingFaceH4/orca_dpo_pairs and the JunxiongWang/llama3-ultrafeedback-armorm datasets. It achieves the following results on the evaluation set:

Loss: 0.4393
Rewards/chosen: -1.7352
Rewards/rejected: -3.4828
Rewards/accuracies: 0.8286
Rewards/margins: 1.7476
Logps/rejected: -623.4827
Logps/chosen: -439.4573
Logits/rejected: -0.6261
Logits/chosen: -0.6084

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: 8
seed: 42
distributed_type: multi-GPU
num_devices: 8
total_train_batch_size: 32
total_eval_batch_size: 64
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.1
num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Rewards/chosen	Rewards/rejected	Rewards/accuracies	Rewards/margins	Logps/rejected	Logps/chosen	Logits/rejected	Logits/chosen
0.4627	0.4798	2000	0.4641	-1.4789	-2.9625	0.8036	1.4836	-571.4525	-413.8262	-0.7196	-0.7049
0.4644	0.9597	4000	0.4393	-1.7352	-3.4828	0.8286	1.7476	-623.4827	-439.4573	-0.6261	-0.6084

Framework versions

Transformers 4.43.1
Pytorch 2.1.1+cu118
Datasets 2.20.0
Tokenizers 0.19.1

MambaInLlama

@article{junxiongdaniele2024mambainllama,
  title   = {The Mamba in the Llama: Distilling and Accelerating Hybrid Models},
  author  = {Junxiong Wang and Daniele Paliotta and Avner May and Alexander M. Rush and Tri Dao},
  journal = {arXiv preprint arXiv:2408.15237},
  year    = {2024}
}