|
|
--- |
|
|
license: other |
|
|
base_model: yandex/YandexGPT-5-Lite-8B-instruct |
|
|
tags: |
|
|
- generated_from_trainer |
|
|
model-index: |
|
|
- name: outputs |
|
|
results: [] |
|
|
library_name: peft |
|
|
datasets: |
|
|
- samedad/mem-and-russian-jokes-dataset |
|
|
--- |
|
|
|
|
|
<!-- 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. --> |
|
|
|
|
|
# outputs |
|
|
|
|
|
This model is a fine-tuned version of [yandex/YandexGPT-5-Lite-8B-instruct](https://huggingface.co/yandex/YandexGPT-5-Lite-8B-instruct) on an |
|
|
[samedad/mem-and-russian-jokes-dataset](https://huggingface.co/datasets/samedad/mem-and-russian-jokes-dataset). |
|
|
It achieves the following results on the evaluation set: |
|
|
- Loss: 2.4964 |
|
|
|
|
|
## Model description |
|
|
|
|
|
More information needed |
|
|
|
|
|
## Intended uses & limitations |
|
|
|
|
|
More information needed |
|
|
|
|
|
## Training and evaluation data |
|
|
|
|
|
More information needed |
|
|
|
|
|
## Training procedure |
|
|
|
|
|
|
|
|
The following `bitsandbytes` quantization config was used during training: |
|
|
- quant_method: QuantizationMethod.BITS_AND_BYTES |
|
|
- _load_in_8bit: False |
|
|
- _load_in_4bit: True |
|
|
- llm_int8_threshold: 6.0 |
|
|
- llm_int8_skip_modules: None |
|
|
- llm_int8_enable_fp32_cpu_offload: False |
|
|
- llm_int8_has_fp16_weight: False |
|
|
- bnb_4bit_quant_type: nf4 |
|
|
- bnb_4bit_use_double_quant: True |
|
|
- bnb_4bit_compute_dtype: bfloat16 |
|
|
- load_in_4bit: True |
|
|
- load_in_8bit: False |
|
|
### Training hyperparameters |
|
|
|
|
|
The following hyperparameters were used during training: |
|
|
- learning_rate: 0.0001 |
|
|
- train_batch_size: 1 |
|
|
- eval_batch_size: 1 |
|
|
- seed: 42 |
|
|
- gradient_accumulation_steps: 8 |
|
|
- total_train_batch_size: 8 |
|
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
|
- lr_scheduler_type: linear |
|
|
- lr_scheduler_warmup_steps: 250 |
|
|
- training_steps: 1000 |
|
|
- mixed_precision_training: Native AMP |
|
|
|
|
|
### Training results |
|
|
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|
|:-------------:|:-----:|:----:|:---------------:| |
|
|
| 6.3533 | 0.04 | 100 | 4.2878 | |
|
|
| 3.3336 | 0.08 | 200 | 2.8855 | |
|
|
| 2.8695 | 0.12 | 300 | 2.8085 | |
|
|
| 2.7996 | 0.16 | 400 | 2.7477 | |
|
|
| 2.7557 | 0.2 | 500 | 2.6723 | |
|
|
| 2.6529 | 0.24 | 600 | 2.5928 | |
|
|
| 2.6168 | 0.29 | 700 | 2.5523 | |
|
|
| 2.585 | 0.33 | 800 | 2.5235 | |
|
|
| 2.5576 | 0.37 | 900 | 2.5039 | |
|
|
| 2.5305 | 0.41 | 1000 | 2.4964 | |
|
|
|
|
|
|
|
|
### Framework versions |
|
|
|
|
|
- PEFT 0.5.0 |
|
|
- Transformers 4.38.2 |
|
|
- Pytorch 2.6.0+cu124 |
|
|
- Datasets 4.1.1 |
|
|
- Tokenizers 0.15.2 |
