---
base_model: Aculi/Tinyllama-2B
library_name: peft
tags:
- generated_from_trainer
model-index:
- name: outputs/thinking-tiny-llama
  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. -->

[<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl)
<details><summary>See axolotl config</summary>

axolotl version: `0.4.1`
```yaml
base_model: Aculi/Tinyllama-2B
model_type: LlamaForCausalLM
tokenizer_type: LlamaTokenizer

load_in_8bit: false
load_in_4bit: true
strict: false

datasets:
  - path: ./datas/1.json
    type: alpaca
  - path: ./datas/2.json
    type: alpaca
    
dataset_prepared_path:
val_set_size: 0.05
output_dir: ./outputs/thinking-tiny-llama

adapter: qlora
lora_model_dir:

sequence_len: 4096
sample_packing: true
eval_sample_packing: false
pad_to_sequence_len: true

lora_r: 32
lora_alpha: 16
lora_dropout: 0.05
lora_target_modules:
lora_target_linear: true
lora_fan_in_fan_out:

wandb_project:
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 4
micro_batch_size: 2
num_epochs: 4
optimizer: paged_adamw_32bit
lr_scheduler: cosine
learning_rate: 0.0002

train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: false

gradient_checkpointing: true
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention: false
flash_attention: true

warmup_steps: 10
evals_per_epoch: 2
saves_per_epoch: 2
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:
```

</details><br>

# outputs/thinking-tiny-llama

This model is a fine-tuned version of [Aculi/Tinyllama-2B](https://huggingface.co/Aculi/Tinyllama-2B) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.0222

## 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: 0.0002
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 10
- num_epochs: 4

### Training results

| Training Loss | Epoch  | Step | Validation Loss |
|:-------------:|:------:|:----:|:---------------:|
| 1.5625        | 0.0013 | 1    | 1.5692          |
| 1.1161        | 0.5002 | 400  | 1.0995          |
| 1.0509        | 1.0003 | 800  | 1.0633          |
| 1.0665        | 1.4867 | 1200 | 1.0422          |
| 1.012         | 1.9869 | 1600 | 1.0287          |
| 1.0124        | 2.4733 | 2000 | 1.0250          |
| 0.8544        | 2.9734 | 2400 | 1.0212          |
| 0.9435        | 3.4605 | 2800 | 1.0222          |


### Framework versions

- PEFT 0.11.1
- Transformers 4.43.1
- Pytorch 2.3.1+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1