wmt_llama2-7b_sft_reward_mtst3_fixemb
This model is a fine-tuned version of meta-llama/Llama-2-7b-hf using TIM method.
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: 2e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 4
- total_train_batch_size: 128
- total_eval_batch_size: 32
- optimizer: Adam with betas=(0.9,0.95) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.03
- training_steps: 5000
Training results
Framework versions
- Transformers 4.34.1
- Pytorch 2.1.0a0+gitf8b6084
- Datasets 2.14.7
- Tokenizers 0.14.1
Model tree for model-trial/llama2-7b-TIM-fixemb
Base model
meta-llama/Llama-2-7b-hf
Finetuned
this model