|
--- |
|
base_model: concedo/KobbleTinyV2-1.1B |
|
library_name: peft |
|
license: apache-2.0 |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
- name: outputs/32r |
|
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: `0.4.1` |
|
```yaml |
|
base_model: concedo/KobbleTinyV2-1.1B |
|
model_type: LlamaForCausalLM |
|
tokenizer_type: LlamaTokenizer |
|
|
|
load_in_8bit: false |
|
load_in_4bit: true |
|
strict: false |
|
|
|
datasets: |
|
- path: NobodyExistsOnTheInternet/AlpacaToxicQA |
|
type: alpaca |
|
- path: Fischerboot/freedom-rp-alpaca-shortend |
|
type: alpaca |
|
dataset_prepared_path: |
|
val_set_size: 0.05 |
|
output_dir: ./outputs/32r |
|
|
|
adapter: qlora |
|
lora_model_dir: |
|
|
|
sequence_len: 4096 |
|
sample_packing: true |
|
eval_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: |
|
wandb_entity: |
|
wandb_watch: |
|
wandb_name: |
|
wandb_log_model: |
|
|
|
gradient_accumulation_steps: 4 |
|
micro_batch_size: 2 |
|
num_epochs: 4 |
|
optimizer: paged_adamw_32bit |
|
lr_scheduler: cosine |
|
learning_rate: 0.0002 |
|
|
|
train_on_inputs: false |
|
group_by_length: false |
|
bf16: auto |
|
fp16: |
|
tf32: false |
|
|
|
gradient_checkpointing: true |
|
early_stopping_patience: |
|
resume_from_checkpoint: |
|
local_rank: |
|
logging_steps: 1 |
|
xformers_attention: |
|
flash_attention: true |
|
|
|
warmup_steps: 10 |
|
evals_per_epoch: 4 |
|
saves_per_epoch: 1 |
|
debug: |
|
deepspeed: |
|
weight_decay: 0.0 |
|
fsdp: |
|
fsdp_config: |
|
special_tokens: |
|
|
|
``` |
|
|
|
</details><br> |
|
|
|
# outputs/32r |
|
|
|
This model is a fine-tuned version of [concedo/KobbleTinyV2-1.1B](https://huggingface.co/concedo/KobbleTinyV2-1.1B) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 1.3368 |
|
|
|
## 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: 10 |
|
- num_epochs: 4 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:------:|:----:|:---------------:| |
|
| 1.9821 | 0.0034 | 1 | 1.8932 | |
|
| 1.6851 | 0.2517 | 73 | 1.5089 | |
|
| 1.4335 | 0.5034 | 146 | 1.4387 | |
|
| 1.3165 | 0.7552 | 219 | 1.4085 | |
|
| 2.0848 | 1.0069 | 292 | 1.3896 | |
|
| 1.3564 | 1.2379 | 365 | 1.3757 | |
|
| 1.2587 | 1.4897 | 438 | 1.3640 | |
|
| 1.2955 | 1.7414 | 511 | 1.3552 | |
|
| 1.4962 | 1.9931 | 584 | 1.3487 | |
|
| 1.3458 | 2.2284 | 657 | 1.3455 | |
|
| 1.301 | 2.4802 | 730 | 1.3413 | |
|
| 1.2458 | 2.7319 | 803 | 1.3389 | |
|
| 1.1965 | 2.9836 | 876 | 1.3367 | |
|
| 1.4968 | 3.2172 | 949 | 1.3369 | |
|
| 1.2504 | 3.4690 | 1022 | 1.3368 | |
|
| 1.5103 | 3.7207 | 1095 | 1.3368 | |
|
|
|
|
|
### Framework versions |
|
|
|
- PEFT 0.11.1 |
|
- Transformers 4.42.3 |
|
- Pytorch 2.1.2+cu118 |
|
- Datasets 2.19.1 |
|
- Tokenizers 0.19.1 |
