---
library_name: transformers
base_model: timarni/qwen3_dpo
tags:
- generated_from_trainer
datasets:
- timarni/MNLP_STEM_IT
model-index:
- name: outputs/dpo_stem_it
  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: `0.9.2`
```yaml
base_model: timarni/qwen3_dpo
# Automatically upload checkpoint and final model to HF
# hub_model_id: username/custom_model_name

plugins:
  - axolotl.integrations.cut_cross_entropy.CutCrossEntropyPlugin
strict: false

chat_template: qwen3
datasets:
  - path: timarni/MNLP_STEM_IT
    type: alpaca
    split: train

shuffle_merged_datasets: true

val_set_size: 0.1
output_dir: ./outputs/dpo_stem_it
dataset_prepared_path: last_run_prepared

sequence_len: 4096 #2048
sample_packing: true # was true -> need to check if it actually learns on the samples or not (better understand te hyperparam and event. install axolotl to debug)
eval_sample_packing: true
pad_to_sequence_len: true
# train_on_inputs: true # NEW
# group_by_length: false NEW?

# To be sure that no LORA is done
adapter: null
lora: false
merge_lora: false

wandb_project: mnlp_project
wandb_entity: tim-arni
wandb_watch:
wandb_name: dpo_stem_it
wandb_log_model:

gradient_accumulation_steps: 16 # 2
micro_batch_size: 2 # 1
num_epochs: 3
optimizer: adamw_torch
lr_scheduler: cosine
learning_rate: 0.00005 # 0.00005
# cosine_min_lr_ratio: 0.1

warmup_ratio: 0.05
weight_decay: 0.01

bf16: auto
tf32: true

gradient_checkpointing: offload
gradient_checkpointing_kwargs:
  use_reentrant: false
resume_from_checkpoint:
logging_steps: 1
gradient_clipping: 1.0 # or max_grad_norm?
flash_attention: true

evals_per_epoch: 4
saves_per_epoch: 2
save_total_limit: 10
special_tokens:

```

</details><br>

# outputs/dpo_stem_it

This model is a fine-tuned version of [timarni/qwen3_dpo](https://huggingface.co/timarni/qwen3_dpo) on the timarni/MNLP_STEM_IT dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1784

## 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: 5e-05
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 16
- total_train_batch_size: 128
- total_eval_batch_size: 8
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 3
- num_epochs: 3.0

### Training results

| Training Loss | Epoch  | Step | Validation Loss |
|:-------------:|:------:|:----:|:---------------:|
| 1.0322        | 0.0497 | 1    | 1.1077          |
| 0.2802        | 0.2484 | 5    | 0.2157          |
| 0.1753        | 0.4969 | 10   | 0.2002          |
| 0.1614        | 0.7453 | 15   | 0.1912          |
| 0.1582        | 0.9938 | 20   | 0.1867          |
| 0.145         | 1.1988 | 25   | 0.1849          |
| 0.1414        | 1.4472 | 30   | 0.1817          |
| 0.1371        | 1.6957 | 35   | 0.1794          |
| 0.1385        | 1.9441 | 40   | 0.1792          |
| 0.1381        | 2.1491 | 45   | 0.1788          |
| 0.133         | 2.3975 | 50   | 0.1785          |
| 0.1297        | 2.6460 | 55   | 0.1785          |
| 0.1338        | 2.8944 | 60   | 0.1784          |


### Framework versions

- Transformers 4.51.3
- Pytorch 2.5.1+cu121
- Datasets 3.5.1
- Tokenizers 0.21.1