---
library_name: peft
tags:
- generated_from_trainer
base_model: bofenghuang/vigogne-2-7b-instruct
model-index:
- name: qlora-out
  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: bofenghuang/vigogne-2-7b-instruct
base_model_config: bofenghuang/vigogne-2-7b-instruct
model_type: LlamaForCausalLM
tokenizer_type: LlamaTokenizer
is_llama_derived_model: true

load_in_8bit: false
load_in_4bit: true
strict: false

datasets:
  - path: bobyres/LabelV01
    type: alpaca 

dataset_prepared_path: last_run_prepared
val_set_size: 0.05
output_dir: ./qlora-out
adapter: qlora
lora_model_dir:

sequence_len: 4096
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: "finetune_labelisation_v011"
wandb_entity:
wandb_watch:
wandb_run_id:
wandb_log_model: "checkpoint"

gradient_accumulation_steps: 1
micro_batch_size: 1
num_epochs: 3
optimizer: paged_adamw_32bit
lr_scheduler: cosine
learning_rate: 0.0002

train_on_inputs: false
group_by_length: false
bf16: true
fp16: false
tf32: false

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

warmup_steps: 10
eval_steps: 0.05
eval_table_size:
save_steps:
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:
  bos_token: "<s>"
  eos_token: "</s>"
  unk_token: "<unk>"

```

</details><br>

# qlora-out

This model is a fine-tuned version of [bofenghuang/vigogne-2-7b-instruct](https://huggingface.co/bofenghuang/vigogne-2-7b-instruct) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6592

## 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: 1
- eval_batch_size: 1
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 10
- num_epochs: 3

### Training results

| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 1.0134        | 0.03  | 1    | 1.0981          |
| 0.972         | 0.15  | 6    | 1.0736          |
| 0.7982        | 0.31  | 12   | 0.8548          |
| 0.6944        | 0.46  | 18   | 0.7151          |
| 0.6808        | 0.62  | 24   | 0.6943          |
| 0.6763        | 0.77  | 30   | 0.6821          |
| 0.67          | 0.92  | 36   | 0.6764          |
| 0.6424        | 1.08  | 42   | 0.6730          |
| 0.6552        | 1.23  | 48   | 0.6780          |
| 0.6527        | 1.38  | 54   | 0.6690          |
| 0.6624        | 1.54  | 60   | 0.6632          |
| 0.6228        | 1.69  | 66   | 0.6625          |
| 0.6447        | 1.85  | 72   | 0.6617          |
| 0.6409        | 2.0   | 78   | 0.6599          |
| 0.6356        | 2.15  | 84   | 0.6589          |
| 0.648         | 2.31  | 90   | 0.6584          |
| 0.6254        | 2.46  | 96   | 0.6593          |
| 0.6167        | 2.62  | 102  | 0.6596          |
| 0.6451        | 2.77  | 108  | 0.6590          |
| 0.6144        | 2.92  | 114  | 0.6592          |


### Framework versions

- PEFT 0.11.0
- Transformers 4.39.0.dev0
- Pytorch 2.0.1+cu118
- Datasets 2.17.1
- Tokenizers 0.15.0