|
--- |
|
license: llama2 |
|
library_name: peft |
|
tags: |
|
- axolotl |
|
- generated_from_trainer |
|
base_model: codellama/CodeLlama-7b-hf |
|
model-index: |
|
- name: EvolCodeLlama-JS-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: `0.3.0` |
|
```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-JS-7b |
|
|
|
load_in_8bit: false |
|
load_in_4bit: true |
|
strict: false |
|
|
|
datasets: |
|
- path: harryng4869/Evol-Instruct-JS-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: axolotl |
|
wandb_entity: |
|
wandb_watch: |
|
wandb_run_id: |
|
wandb_log_model: |
|
|
|
gradient_accumulation_steps: 4 |
|
micro_batch_size: 2 |
|
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-JS-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: 0.2897 |
|
|
|
## 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: 2 |
|
- eval_batch_size: 2 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 4 |
|
- 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 | |
|
|:-------------:|:-----:|:----:|:---------------:| |
|
| 0.4099 | 0.02 | 1 | 0.4313 | |
|
| 0.5677 | 0.04 | 2 | 0.4313 | |
|
| 0.4255 | 0.08 | 4 | 0.4315 | |
|
| 0.4352 | 0.12 | 6 | 0.4312 | |
|
| 0.4457 | 0.17 | 8 | 0.4312 | |
|
| 0.4705 | 0.21 | 10 | 0.4309 | |
|
| 0.4492 | 0.25 | 12 | 0.4303 | |
|
| 0.5233 | 0.29 | 14 | 0.4294 | |
|
| 0.3795 | 0.33 | 16 | 0.4275 | |
|
| 0.456 | 0.37 | 18 | 0.4248 | |
|
| 0.5132 | 0.41 | 20 | 0.4204 | |
|
| 0.3543 | 0.46 | 22 | 0.4136 | |
|
| 0.4132 | 0.5 | 24 | 0.4046 | |
|
| 0.4219 | 0.54 | 26 | 0.3936 | |
|
| 0.3956 | 0.58 | 28 | 0.3813 | |
|
| 0.3587 | 0.62 | 30 | 0.3697 | |
|
| 0.409 | 0.66 | 32 | 0.3587 | |
|
| 0.3093 | 0.7 | 34 | 0.3483 | |
|
| 0.3717 | 0.75 | 36 | 0.3407 | |
|
| 0.3357 | 0.79 | 38 | 0.3345 | |
|
| 0.2912 | 0.83 | 40 | 0.3289 | |
|
| 0.3171 | 0.87 | 42 | 0.3243 | |
|
| 0.3368 | 0.91 | 44 | 0.3210 | |
|
| 0.3906 | 0.95 | 46 | 0.3180 | |
|
| 0.3491 | 0.99 | 48 | 0.3159 | |
|
| 0.274 | 1.02 | 50 | 0.3133 | |
|
| 0.2474 | 1.06 | 52 | 0.3126 | |
|
| 0.3236 | 1.1 | 54 | 0.3106 | |
|
| 0.3327 | 1.14 | 56 | 0.3092 | |
|
| 0.3153 | 1.18 | 58 | 0.3081 | |
|
| 0.3809 | 1.22 | 60 | 0.3079 | |
|
| 0.2792 | 1.26 | 62 | 0.3072 | |
|
| 0.2465 | 1.31 | 64 | 0.3055 | |
|
| 0.2831 | 1.35 | 66 | 0.3060 | |
|
| 0.408 | 1.39 | 68 | 0.3064 | |
|
| 0.2881 | 1.43 | 70 | 0.3045 | |
|
| 0.2715 | 1.47 | 72 | 0.3018 | |
|
| 0.2686 | 1.51 | 74 | 0.3008 | |
|
| 0.3605 | 1.55 | 76 | 0.3008 | |
|
| 0.2644 | 1.6 | 78 | 0.3002 | |
|
| 0.3479 | 1.64 | 80 | 0.2990 | |
|
| 0.2821 | 1.68 | 82 | 0.2983 | |
|
| 0.3193 | 1.72 | 84 | 0.2980 | |
|
| 0.2857 | 1.76 | 86 | 0.2969 | |
|
| 0.2484 | 1.8 | 88 | 0.2965 | |
|
| 0.236 | 1.84 | 90 | 0.2957 | |
|
| 0.3554 | 1.89 | 92 | 0.2946 | |
|
| 0.2968 | 1.93 | 94 | 0.2931 | |
|
| 0.3792 | 1.97 | 96 | 0.2914 | |
|
| 0.2574 | 2.01 | 98 | 0.2909 | |
|
| 0.3192 | 2.02 | 100 | 0.2915 | |
|
| 0.2519 | 2.06 | 102 | 0.2934 | |
|
| 0.2165 | 2.1 | 104 | 0.2968 | |
|
| 0.2499 | 2.14 | 106 | 0.2960 | |
|
| 0.2243 | 2.18 | 108 | 0.2931 | |
|
| 0.2523 | 2.22 | 110 | 0.2923 | |
|
| 0.2644 | 2.26 | 112 | 0.2943 | |
|
| 0.2048 | 2.31 | 114 | 0.2946 | |
|
| 0.1853 | 2.35 | 116 | 0.2932 | |
|
| 0.2441 | 2.39 | 118 | 0.2927 | |
|
| 0.2494 | 2.43 | 120 | 0.2928 | |
|
| 0.2184 | 2.47 | 122 | 0.2927 | |
|
| 0.2376 | 2.51 | 124 | 0.2932 | |
|
| 0.2496 | 2.55 | 126 | 0.2924 | |
|
| 0.2029 | 2.6 | 128 | 0.2915 | |
|
| 0.2602 | 2.64 | 130 | 0.2908 | |
|
| 0.2137 | 2.68 | 132 | 0.2907 | |
|
| 0.2617 | 2.72 | 134 | 0.2901 | |
|
| 0.2532 | 2.76 | 136 | 0.2901 | |
|
| 0.2743 | 2.8 | 138 | 0.2900 | |
|
| 0.2181 | 2.84 | 140 | 0.2900 | |
|
| 0.254 | 2.89 | 142 | 0.2899 | |
|
| 0.2463 | 2.93 | 144 | 0.2897 | |
|
|
|
|
|
### Framework versions |
|
|
|
- PEFT 0.7.2.dev0 |
|
- Transformers 4.37.0.dev0 |
|
- Pytorch 2.0.1+cu118 |
|
- Datasets 2.16.1 |
|
- Tokenizers 0.15.0 |