|
--- |
|
base_model: ondevicellm/tinyllama_mole_v1 |
|
tags: |
|
- alignment-handbook |
|
- trl |
|
- sft |
|
- generated_from_trainer |
|
- trl |
|
- sft |
|
- generated_from_trainer |
|
datasets: |
|
- HuggingFaceH4/ultrachat_200k |
|
model-index: |
|
- name: tinyllama_mole_sft_router05_ep3 |
|
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. --> |
|
|
|
# tinyllama_mole_sft_router05_ep3 |
|
|
|
This model is a fine-tuned version of [ondevicellm/tinyllama_mole_v1](https://huggingface.co./ondevicellm/tinyllama_mole_v1) on the HuggingFaceH4/ultrachat_200k dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 2.1129 |
|
|
|
## 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: 2e-05 |
|
- train_batch_size: 16 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- distributed_type: multi-GPU |
|
- num_devices: 4 |
|
- gradient_accumulation_steps: 2 |
|
- total_train_batch_size: 128 |
|
- total_eval_batch_size: 32 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: cosine |
|
- lr_scheduler_warmup_steps: 120 |
|
- num_epochs: 3 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:-----:|:----:|:---------------:| |
|
| 2.3008 | 0.09 | 100 | 2.2785 | |
|
| 2.2257 | 0.18 | 200 | 2.2161 | |
|
| 2.1922 | 0.26 | 300 | 2.1924 | |
|
| 2.1698 | 0.35 | 400 | 2.1773 | |
|
| 2.1428 | 0.44 | 500 | 2.1668 | |
|
| 2.1632 | 0.53 | 600 | 2.1586 | |
|
| 2.1503 | 0.61 | 700 | 2.1516 | |
|
| 2.1369 | 0.7 | 800 | 2.1460 | |
|
| 2.1324 | 0.79 | 900 | 2.1409 | |
|
| 2.1158 | 0.88 | 1000 | 2.1362 | |
|
| 2.1396 | 0.96 | 1100 | 2.1321 | |
|
| 2.0565 | 1.05 | 1200 | 2.1317 | |
|
| 2.0596 | 1.14 | 1300 | 2.1297 | |
|
| 2.0712 | 1.23 | 1400 | 2.1276 | |
|
| 2.0626 | 1.31 | 1500 | 2.1259 | |
|
| 2.0654 | 1.4 | 1600 | 2.1235 | |
|
| 2.0628 | 1.49 | 1700 | 2.1216 | |
|
| 2.046 | 1.58 | 1800 | 2.1197 | |
|
| 2.067 | 1.66 | 1900 | 2.1180 | |
|
| 2.0702 | 1.75 | 2000 | 2.1161 | |
|
| 2.057 | 1.84 | 2100 | 2.1144 | |
|
| 2.0307 | 1.93 | 2200 | 2.1129 | |
|
| 2.0134 | 2.01 | 2300 | 2.1172 | |
|
| 2.0205 | 2.1 | 2400 | 2.1172 | |
|
| 2.0091 | 2.19 | 2500 | 2.1170 | |
|
| 2.0021 | 2.28 | 2600 | 2.1164 | |
|
| 2.0006 | 2.37 | 2700 | 2.1159 | |
|
| 2.006 | 2.45 | 2800 | 2.1158 | |
|
| 2.0121 | 2.54 | 2900 | 2.1152 | |
|
| 1.9942 | 2.63 | 3000 | 2.1150 | |
|
| 2.0129 | 2.72 | 3100 | 2.1149 | |
|
| 2.0041 | 2.8 | 3200 | 2.1146 | |
|
| 2.0002 | 2.89 | 3300 | 2.1146 | |
|
| 2.019 | 2.98 | 3400 | 2.1146 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.37.0 |
|
- Pytorch 2.1.2+cu118 |
|
- Datasets 2.16.1 |
|
- Tokenizers 0.15.0 |
|
|
