Llama_CDG

<<<<<<< HEAD

license: mit

=======

library_name: transformers license: other base_model: /online1/ycsc_lijt1/lijt1/wpz/hf_models/Meta-Llama-3.1-8B-Instruct tags:

• llama-factory
• full
• generated_from_trainer model-index:
• name: CDG_sl_bsz256 results: []

---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

CDG_sl_bsz256

This model is a fine-tuned version of /online1/ycsc_lijt1/lijt1/wpz/hf_models/Meta-Llama-3.1-8B-Instruct on the EI_round1-loop1, the EI_round1-loop2, the mislead-round2-loop2, the correct-round2-loop2, the mislead-round2-loop1 and the correct-round2-loop1 datasets.

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-06
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• gradient_accumulation_steps: 8
• total_train_batch_size: 256
• total_eval_batch_size: 64
• 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
• num_epochs: 1.0

Training results

Framework versions

• Transformers 4.48.0.dev0
• Pytorch 2.4.0+cu121
• Datasets 3.1.0
• Tokenizers 0.21.0

d623a31 (first_commit)