metadata

library_name: transformers
license: apache-2.0
base_model: mistralai/Mistral-7B-v0.1
tags:
  - llama-factory
  - full
  - generated_from_trainer
model-index:
  - name: hp_ablations_grid_mistral_bsz1024_lr5e-6_scheduler-cosine-warmup0.15
    results: []

hp_ablations_grid_mistral_bsz1024_lr5e-6_scheduler-cosine-warmup0.15

This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on the mlfoundations-dev/oh-dcft-v3-llama3.1-nemotron-70b_shareGPT_format dataset. It achieves the following results on the evaluation set:

Loss: 0.0575

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-06
train_batch_size: 8
eval_batch_size: 8
seed: 42
distributed_type: multi-GPU
num_devices: 16
gradient_accumulation_steps: 8
total_train_batch_size: 1024
total_eval_batch_size: 128
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: 0.15
num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss
0.4775	1.0	336	0.0599
0.4039	2.0	672	0.0565
0.3325	3.0	1008	0.0575

Framework versions

Transformers 4.46.1
Pytorch 2.3.0
Datasets 3.1.0
Tokenizers 0.20.3