Hugging Face's logo

Corianas
/
mistral-7b-expert-iteration-iter3-GPTQ

Safetensors
mistral
alignment-handbook
Generated from Trainer
autoquant
gptq
8-bit precision
Model card Files Community
mistral-7b-expert-iteration-iter3-GPTQ / README.md
Corianas's picture
Corianas
Update README.md
a5a3cbf verified
preview code
|
raw
history blame
1.62 kB
metadata
base_model: UCLA-AGI/Mistral7B-PairRM-SPPO-Iter3
datasets:
  - synthetic_data_mistral-7b-instruct-expert-iteration-iter3_score
tags:
  - alignment-handbook
  - generated_from_trainer
  - autoquant
  - gptq
model-index:
  - name: Mistral7B-PairRM-SPPO-Iter3
    results: []

Mistral-7B-Instruct-EI-Iter3

This model is a GPTQ version of UCLA-AGI/Mistral7B-PairRM-SPPO-Iter3

Created with AutoQuant

Model description

I like the GPTQ format, this is 8bit, GROUP_SIZE 32.

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: 8
  • eval_batch_size: 8
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 8
  • total_train_batch_size: 64
  • 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: 1.0

Training results

Training Loss Epoch Step Validation Loss
0.6652 1.0 106 0.4722

Framework versions

  • Transformers 4.42.4
  • Pytorch 2.3.1+cu121
  • Datasets 2.14.6
  • Tokenizers 0.19.1