---
license: other
library_name: peft
tags:
- generated_from_trainer
base_model: deepseek-ai/deepseek-coder-33b-instruct
model-index:
- name: lora-logo_fix_full_deepseek33b_gpt35i_lr_0.0002_alpha_1024_r_1024
  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
adapter: lora
base_model: deepseek-ai/deepseek-coder-33b-instruct
bf16: auto
dataset_prepared_path: ./logo_ds_preprocess_list_gpt35
datasets:
- path: ../logo/fix_logo_synthetic_training_data_full.json
  type:
    field_instruction: input
    field_output: output
    format: '### Instruction:

      {input}

      ### Response:

      '
    no_input_format: '{instruction}'
debug: null
deepspeed: ./deepspeed_configs/zero2.json
early_stopping_patience: null
eval_sample_packing: true
evals_per_epoch: 4
flash_attention: true
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 2
gradient_checkpointing: true
group_by_length: false
is_llama_derived_model: true
learning_rate: 0.0002
load_in_4bit: false
load_in_8bit: true
local_rank: null
logging_steps: 1
lora_alpha: 1024
lora_dropout: 0.05
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 1024
lora_target_linear: true
lr_scheduler: cosine
micro_batch_size: 4
model_type: AutoModelForCausalLM
num_epochs: 2
optimizer: adamw_bnb_8bit
output_dir: ./lora-logo_fix_full_deepseek33b_gpt35i_lr_0.0002_alpha_1024_r_1024
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: true
saves_per_epoch: 1
sequence_len: 1800
special_tokens:
  bos_token: "<\uFF5Cbegin\u2581of\u2581sentence\uFF5C>"
  eos_token: <|EOT|>
strict: true
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
val_set_size: 0.05
wandb_entity: null
wandb_log_model: null
wandb_name: logo_fix_full_deepseek33b_gpt35i_lr_0.0002_alpha_1024_r_1024
wandb_project: pbe-axo
wandb_watch: null
warmup_steps: 50
weight_decay: 0.0
xformers_attention: null

```

</details><br>

# lora-logo_fix_full_deepseek33b_gpt35i_lr_0.0002_alpha_1024_r_1024

This model is a fine-tuned version of [deepseek-ai/deepseek-coder-33b-instruct](https://huggingface.co/deepseek-ai/deepseek-coder-33b-instruct) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2380

## 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: 4
- eval_batch_size: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- total_eval_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 50
- num_epochs: 2

### Training results

| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 2.1469        | 0.0   | 1    | 2.1795          |
| 0.3319        | 0.25  | 113  | 0.3324          |
| 0.2883        | 0.5   | 226  | 0.2976          |
| 0.2748        | 0.75  | 339  | 0.2785          |
| 0.2812        | 1.0   | 452  | 0.2612          |
| 0.2276        | 1.23  | 565  | 0.2523          |
| 0.2483        | 1.48  | 678  | 0.2440          |
| 0.1982        | 1.73  | 791  | 0.2380          |


### Framework versions

- PEFT 0.10.0
- Transformers 4.40.0.dev0
- Pytorch 2.1.2+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0