|
--- |
|
license: llama2 |
|
library_name: peft |
|
tags: |
|
- axolotl |
|
- generated_from_trainer |
|
base_model: codellama/CodeLlama-7b-hf |
|
model-index: |
|
- name: EvolCodeLlama-7b |
|
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: codellama/CodeLlama-7b-hf |
|
base_model_config: codellama/CodeLlama-7b-hf |
|
model_type: LlamaForCausalLM |
|
tokenizer_type: LlamaTokenizer |
|
is_llama_derived_model: true |
|
hub_model_id: EvolCodeLlama-7b |
|
|
|
load_in_8bit: false |
|
load_in_4bit: true |
|
strict: false |
|
|
|
datasets: |
|
- path: mlabonne/Evol-Instruct-Python-1k |
|
type: alpaca |
|
dataset_prepared_path: last_run_prepared |
|
val_set_size: 0.02 |
|
output_dir: ./qlora-out |
|
|
|
adapter: qlora |
|
lora_model_dir: |
|
|
|
sequence_len: 2048 |
|
sample_packing: 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: FTCodeLlama-2 |
|
wandb_entity: |
|
wandb_watch: |
|
wandb_run_id: |
|
wandb_log_model: |
|
|
|
gradient_accumulation_steps: 2 |
|
micro_batch_size: 4 |
|
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: 100 |
|
eval_steps: 0.01 |
|
save_strategy: epoch |
|
save_steps: |
|
debug: |
|
deepspeed: |
|
weight_decay: 0.0 |
|
fsdp: |
|
fsdp_config: |
|
special_tokens: |
|
bos_token: "<s>" |
|
eos_token: "</s>" |
|
unk_token: "<unk>" |
|
|
|
``` |
|
|
|
</details><br> |
|
|
|
# EvolCodeLlama-7b |
|
|
|
This model is a fine-tuned version of [codellama/CodeLlama-7b-hf](https://huggingface.co/codellama/CodeLlama-7b-hf) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.3754 |
|
|
|
## 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 |
|
- gradient_accumulation_steps: 2 |
|
- 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: 100 |
|
- num_epochs: 3 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:-----:|:----:|:---------------:| |
|
| 0.3686 | 0.01 | 1 | 0.5015 | |
|
| 0.4397 | 0.03 | 3 | 0.5013 | |
|
| 0.4919 | 0.06 | 6 | 0.5013 | |
|
| 0.3191 | 0.09 | 9 | 0.5011 | |
|
| 0.2514 | 0.12 | 12 | 0.5003 | |
|
| 0.3379 | 0.15 | 15 | 0.4992 | |
|
| 0.4712 | 0.19 | 18 | 0.4969 | |
|
| 0.3801 | 0.22 | 21 | 0.4922 | |
|
| 0.3482 | 0.25 | 24 | 0.4833 | |
|
| 0.4113 | 0.28 | 27 | 0.4702 | |
|
| 0.2524 | 0.31 | 30 | 0.4552 | |
|
| 0.2641 | 0.34 | 33 | 0.4415 | |
|
| 0.3554 | 0.37 | 36 | 0.4302 | |
|
| 0.2384 | 0.4 | 39 | 0.4213 | |
|
| 0.2131 | 0.43 | 42 | 0.4153 | |
|
| 0.2308 | 0.46 | 45 | 0.4105 | |
|
| 0.3478 | 0.49 | 48 | 0.4053 | |
|
| 0.2913 | 0.53 | 51 | 0.4003 | |
|
| 0.2909 | 0.56 | 54 | 0.3956 | |
|
| 0.2032 | 0.59 | 57 | 0.3928 | |
|
| 0.2479 | 0.62 | 60 | 0.3906 | |
|
| 0.2145 | 0.65 | 63 | 0.3890 | |
|
| 0.2447 | 0.68 | 66 | 0.3882 | |
|
| 0.2928 | 0.71 | 69 | 0.3876 | |
|
| 0.384 | 0.74 | 72 | 0.3854 | |
|
| 0.1751 | 0.77 | 75 | 0.3835 | |
|
| 0.352 | 0.8 | 78 | 0.3818 | |
|
| 0.2443 | 0.84 | 81 | 0.3808 | |
|
| 0.3211 | 0.87 | 84 | 0.3798 | |
|
| 0.3026 | 0.9 | 87 | 0.3788 | |
|
| 0.2357 | 0.93 | 90 | 0.3776 | |
|
| 0.2661 | 0.96 | 93 | 0.3755 | |
|
| 0.3314 | 0.99 | 96 | 0.3751 | |
|
| 0.2789 | 1.02 | 99 | 0.3742 | |
|
| 0.1734 | 1.03 | 102 | 0.3744 | |
|
| 0.1928 | 1.06 | 105 | 0.3761 | |
|
| 0.2681 | 1.09 | 108 | 0.3753 | |
|
| 0.4148 | 1.12 | 111 | 0.3750 | |
|
| 0.1977 | 1.15 | 114 | 0.3744 | |
|
| 0.1977 | 1.19 | 117 | 0.3740 | |
|
| 0.2499 | 1.22 | 120 | 0.3742 | |
|
| 0.2192 | 1.25 | 123 | 0.3730 | |
|
| 0.2207 | 1.28 | 126 | 0.3723 | |
|
| 0.2179 | 1.31 | 129 | 0.3718 | |
|
| 0.2843 | 1.34 | 132 | 0.3734 | |
|
| 0.2614 | 1.37 | 135 | 0.3721 | |
|
| 0.2033 | 1.4 | 138 | 0.3705 | |
|
| 0.212 | 1.43 | 141 | 0.3705 | |
|
| 0.2307 | 1.46 | 144 | 0.3712 | |
|
| 0.3182 | 1.49 | 147 | 0.3698 | |
|
| 0.2467 | 1.53 | 150 | 0.3664 | |
|
| 0.1909 | 1.56 | 153 | 0.3665 | |
|
| 0.3286 | 1.59 | 156 | 0.3655 | |
|
| 0.2195 | 1.62 | 159 | 0.3648 | |
|
| 0.3231 | 1.65 | 162 | 0.3650 | |
|
| 0.2922 | 1.68 | 165 | 0.3631 | |
|
| 0.1956 | 1.71 | 168 | 0.3627 | |
|
| 0.2299 | 1.74 | 171 | 0.3639 | |
|
| 0.1585 | 1.77 | 174 | 0.3649 | |
|
| 0.2289 | 1.8 | 177 | 0.3650 | |
|
| 0.189 | 1.84 | 180 | 0.3643 | |
|
| 0.2736 | 1.87 | 183 | 0.3628 | |
|
| 0.3591 | 1.9 | 186 | 0.3614 | |
|
| 0.3181 | 1.93 | 189 | 0.3612 | |
|
| 0.1994 | 1.96 | 192 | 0.3612 | |
|
| 0.2499 | 1.99 | 195 | 0.3618 | |
|
| 0.1659 | 2.01 | 198 | 0.3627 | |
|
| 0.231 | 2.04 | 201 | 0.3665 | |
|
| 0.169 | 2.07 | 204 | 0.3744 | |
|
| 0.2082 | 2.1 | 207 | 0.3800 | |
|
| 0.1755 | 2.13 | 210 | 0.3770 | |
|
| 0.1959 | 2.16 | 213 | 0.3721 | |
|
| 0.1933 | 2.19 | 216 | 0.3705 | |
|
| 0.1213 | 2.22 | 219 | 0.3712 | |
|
| 0.237 | 2.25 | 222 | 0.3738 | |
|
| 0.2277 | 2.28 | 225 | 0.3771 | |
|
| 0.2832 | 2.31 | 228 | 0.3789 | |
|
| 0.2039 | 2.35 | 231 | 0.3783 | |
|
| 0.2302 | 2.38 | 234 | 0.3764 | |
|
| 0.1562 | 2.41 | 237 | 0.3750 | |
|
| 0.1688 | 2.44 | 240 | 0.3742 | |
|
| 0.126 | 2.47 | 243 | 0.3741 | |
|
| 0.1846 | 2.5 | 246 | 0.3746 | |
|
| 0.2195 | 2.53 | 249 | 0.3745 | |
|
| 0.2335 | 2.56 | 252 | 0.3749 | |
|
| 0.1542 | 2.59 | 255 | 0.3750 | |
|
| 0.1783 | 2.62 | 258 | 0.3755 | |
|
| 0.2409 | 2.65 | 261 | 0.3762 | |
|
| 0.1777 | 2.69 | 264 | 0.3762 | |
|
| 0.2591 | 2.72 | 267 | 0.3761 | |
|
| 0.2092 | 2.75 | 270 | 0.3757 | |
|
| 0.2256 | 2.78 | 273 | 0.3757 | |
|
| 0.1923 | 2.81 | 276 | 0.3756 | |
|
| 0.156 | 2.84 | 279 | 0.3755 | |
|
| 0.2036 | 2.87 | 282 | 0.3754 | |
|
| 0.2254 | 2.9 | 285 | 0.3753 | |
|
| 0.1683 | 2.93 | 288 | 0.3753 | |
|
| 0.1528 | 2.96 | 291 | 0.3754 | |
|
|
|
|
|
### Framework versions |
|
|
|
- PEFT 0.8.2 |
|
- Transformers 4.38.0.dev0 |
|
- Pytorch 2.1.2+cu118 |
|
- Datasets 2.17.0 |
|
- Tokenizers 0.15.0 |
