Quantization made by Richard Erkhov.
llama_3b_step2_batch_v4 - GGUF
Original model description:
library_name: transformers
tags:
- trl
- sft
- generated_from_trainer
model-index:
- name: llama_3b_step2_batch_v4
results: []
llama_3b_step2_batch_v4
This model was trained from scratch on an unknown dataset.
It achieves the following results on the evaluation set:
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: 3e-05
- train_batch_size: 6
- eval_batch_size: 40
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 24
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 2
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|
| 0.897 | 0.1022 | 50 | 0.8837 |
| 0.6522 | 0.2045 | 100 | 0.6790 |
| 0.4811 | 0.3067 | 150 | 0.5709 |
| 0.4821 | 0.4090 | 200 | 0.4997 |
| 0.4577 | 0.5112 | 250 | 0.4469 |
| 0.5176 | 0.6135 | 300 | 0.4128 |
| 0.3274 | 0.7157 | 350 | 0.3879 |
| 0.3363 | 0.8180 | 400 | 0.3672 |
| 0.3491 | 0.9202 | 450 | 0.3495 |
| 0.2589 | 1.0225 | 500 | 0.3429 |
| 0.2783 | 1.1247 | 550 | 0.3360 |
| 0.1976 | 1.2270 | 600 | 0.3314 |
| 0.2332 | 1.3292 | 650 | 0.3274 |
| 0.2571 | 1.4315 | 700 | 0.3237 |
| 0.2596 | 1.5337 | 750 | 0.3210 |
| 0.2918 | 1.6360 | 800 | 0.3169 |
| 0.1956 | 1.7382 | 850 | 0.3167 |
| 0.2186 | 1.8405 | 900 | 0.3159 |
| 0.2477 | 1.9427 | 950 | 0.3157 |
Framework versions
- Transformers 4.46.1
- Pytorch 2.1.0+cu118
- Datasets 3.0.2
- Tokenizers 0.20.1
