|
--- |
|
license: apache-2.0 |
|
base_model: JackFram/llama-68m |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- accuracy |
|
model-index: |
|
- name: recreate_llama_68M_vanilla |
|
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. --> |
|
|
|
# recreate_llama_68M_vanilla |
|
|
|
This model is a fine-tuned version of [JackFram/llama-68m](https://huggingface.co/JackFram/llama-68m) on the anon8231489123/ShareGPT_Vicuna_unfiltered/ShareGPT_V3_unfiltered_cleaned_split.json dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 9.5494 |
|
- Accuracy: 0.3512 |
|
|
|
## 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: 0.005 |
|
- train_batch_size: 32 |
|
- eval_batch_size: 16 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 3.0 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | |
|
|:-------------:|:------:|:----:|:---------------:|:--------:| |
|
| No log | 0.3125 | 10 | 7.9370 | 0.3676 | |
|
| No log | 0.625 | 20 | 8.6808 | 0.3478 | |
|
| No log | 0.9375 | 30 | 10.9798 | 0.1029 | |
|
| No log | 1.25 | 40 | 10.3023 | 0.2493 | |
|
| No log | 1.5625 | 50 | 9.7688 | 0.3501 | |
|
| No log | 1.875 | 60 | 9.6190 | 0.3510 | |
|
| No log | 2.1875 | 70 | 9.5617 | 0.3510 | |
|
| No log | 2.5 | 80 | 9.5470 | 0.3511 | |
|
| No log | 2.8125 | 90 | 9.5487 | 0.3511 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.41.0.dev0 |
|
- Pytorch 2.1.0+cu121 |
|
- Datasets 2.19.1 |
|
- Tokenizers 0.19.1 |
|
|
