61 lines
1.3 KiB
Markdown
61 lines
1.3 KiB
Markdown
|
---
|
||
|
|
license: apache-2.0
|
||
|
|
base_model: mistralai/Mistral-7B-v0.1
|
||
|
|
language:
|
||
|
|
- en
|
||
|
|
pipeline_tag: text-generation
|
||
|
|
tags:
|
||
|
|
- alignment-handbook
|
||
|
|
- generated_from_trainer
|
||
|
|
model-index:
|
||
|
|
- name: babygrok
|
||
|
|
results: []
|
||
|
|
library_name: transformers
|
||
|
|
---
|
||
|
|
|
||
|
|
# babygrok
|
||
|
|
|
||
|
|
A mini version of Grok, by [ripbaggie](https://huggingface.co/ripbaggie).
|
||
|
|
|
||
|
|
## About
|
||
|
|
|
||
|
|
babygrok is Grok 4.3 condensed into a 7B Mistral model, trained to mirror the behavior of Grok — essentially Grok in a tiny package, allowing you to host it locally.
|
||
|
|
|
||
|
|
## Usage
|
||
|
|
|
||
|
|
A 7B-parameter model designed to run locally.
|
||
|
|
|
||
|
|
## License
|
||
|
|
|
||
|
|
Apache 2.0.
|
||
|
|
|
||
|
|
### 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
|
||
|
|
- gradient_accumulation_steps: 4
|
||
|
|
- total_train_batch_size: 256
|
||
|
|
- 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
|
||
|
|
|
||
|
|
### Training results
|
||
|
|
|
||
|
|
| Training Loss | Epoch | Step | Validation Loss |
|
||
|
|
|:-------------:|:-----:|:----:|:---------------:|
|
||
|
|
| 0.9326 | 1.0 | 545 | 0.9348 |
|
||
|
|
|
||
|
|
|
||
|
|
### Framework versions
|
||
|
|
|
||
|
|
- Transformers 4.36.2
|
||
|
|
- Pytorch 2.1.2+cu121
|
||
|
|
- Datasets 2.16.1
|
||
|
|
- Tokenizers 0.15.0
|