metadata
library_name: transformers
license: apache-2.0
base_model: alignment-handbook/zephyr-7b-sft-full
tags:
- alignment-handbook
- trl
- sft
- generated_from_trainer
- trl
- sft
- generated_from_trainer
datasets:
- EunsuKim/instruct
- EunsuKim/MATH
model-index:
- name: zephyr-7b-math-case-7
results: []
zephyr-7b-math-case-7
This model is a fine-tuned version of alignment-handbook/zephyr-7b-sft-full on the EunsuKim/instruct and the EunsuKim/MATH datasets. It achieves the following results on the evaluation set:
- Loss: 0.0013
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-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- total_train_batch_size: 32
- total_eval_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.03
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.8719 | 1.0 | 18 | 0.5943 |
| 0.4801 | 2.0 | 36 | 0.2952 |
| 0.2452 | 3.0 | 54 | 0.1275 |
| 0.098 | 4.0 | 72 | 0.0448 |
| 0.0439 | 5.0 | 90 | 0.0259 |
| 0.0306 | 6.0 | 108 | 0.0149 |
| 0.017 | 7.0 | 126 | 0.0068 |
| 0.006 | 8.0 | 144 | 0.0025 |
| 0.0025 | 9.0 | 162 | 0.0014 |
| 0.0012 | 10.0 | 180 | 0.0013 |
Framework versions
- Transformers 4.44.2
- Pytorch 2.4.1+cu121
- Datasets 3.0.0
- Tokenizers 0.19.1