---
license: other
base_model: lightblue/suzume-llama-3-8B-multilingual
tags:
- generated_from_trainer
model-index:
- name: workspace/llm_training/axolotl/llama3-multilingual-orpo/output_mitsu_top75_borda
  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/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)
<details><summary>See axolotl config</summary>

axolotl version: `0.4.0`
```yaml
base_model: lightblue/suzume-llama-3-8B-multilingual
model_type: LlamaForCausalLM
tokenizer_type: AutoTokenizer  # PreTrainedTokenizerFast

load_in_8bit: false
load_in_4bit: false
strict: false

rl: orpo
orpo_alpha: 0.1
remove_unused_columns: false

chat_template: chatml
datasets:
  - path: lightblue/mitsu_top75_borda
    type: orpo.chat_template
    conversation: llama-3
dataset_prepared_path: /workspace/llm_training/axolotl/llama3-multilingual-orpo/prepared_mitsu_top75_borda
val_set_size: 0.02
output_dir: /workspace/llm_training/axolotl/llama3-multilingual-orpo/output_mitsu_top75_borda

sequence_len: 8192
sample_packing: false
pad_to_sequence_len: true

use_wandb: true
wandb_project: axolotl
wandb_entity: peterd
wandb_name: mitsu_top75_borda

gradient_accumulation_steps: 8
micro_batch_size: 1
num_epochs: 1
optimizer: paged_adamw_8bit
lr_scheduler: cosine
learning_rate: 8e-6

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

gradient_checkpointing: true
gradient_checkpointing_kwargs:
  use_reentrant: false
early_stopping_patience:
resume_from_checkpoint:
logging_steps: 1
xformers_attention:
flash_attention: true

warmup_steps: 10
evals_per_epoch: 20
eval_table_size:
saves_per_epoch: 1
debug:
deepspeed: /workspace/axolotl/deepspeed_configs/zero3_bf16.json
weight_decay: 0.0
special_tokens:
  pad_token: <|end_of_text|>
```

</details><br>

# workspace/llm_training/axolotl/llama3-multilingual-orpo/output_mitsu_top75_borda

This model is a fine-tuned version of [lightblue/suzume-llama-3-8B-multilingual](https://huggingface.co/lightblue/suzume-llama-3-8B-multilingual) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0863

## 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: 8e-06
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 8
- total_train_batch_size: 32
- total_eval_batch_size: 4
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 10
- num_epochs: 1

### Training results

| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 7.6309        | 0.02  | 1    | 7.7104          |
| 3.9307        | 0.06  | 4    | 2.3582          |
| 0.1361        | 0.13  | 8    | 0.1163          |
| 0.1072        | 0.19  | 12   | 0.1045          |
| 0.1087        | 0.26  | 16   | 0.1007          |
| 0.1109        | 0.32  | 20   | 0.0971          |
| 0.1015        | 0.39  | 24   | 0.0908          |
| 0.1032        | 0.45  | 28   | 0.0872          |
| 0.0996        | 0.52  | 32   | 0.0968          |
| 0.1107        | 0.58  | 36   | 0.0982          |
| 0.1079        | 0.65  | 40   | 0.0911          |
| 0.1011        | 0.71  | 44   | 0.0893          |
| 0.1251        | 0.78  | 48   | 0.0866          |
| 0.1008        | 0.84  | 52   | 0.0863          |
| 0.0948        | 0.91  | 56   | 0.0863          |
| 0.0936        | 0.97  | 60   | 0.0863          |


### Framework versions

- Transformers 4.38.2
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.0