|
|
---
|
|
|
library_name: peft
|
|
|
license: llama2
|
|
|
base_model: Yellow-AI-NLP/komodo-7b-base
|
|
|
tags:
|
|
|
- trl
|
|
|
- sft
|
|
|
- generated_from_trainer
|
|
|
datasets:
|
|
|
- id_nergrit_corpus
|
|
|
model-index:
|
|
|
- name: result
|
|
|
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. -->
|
|
|
|
|
|
# result
|
|
|
|
|
|
This model is a fine-tuned version of [Yellow-AI-NLP/komodo-7b-base](https://huggingface.co/Yellow-AI-NLP/komodo-7b-base) on the id_nergrit_corpus dataset.
|
|
|
It achieves the following results on the evaluation set:
|
|
|
- eval_loss: 0.0598
|
|
|
- eval_runtime: 429.4851
|
|
|
- eval_samples_per_second: 1.469
|
|
|
- eval_steps_per_second: 0.368
|
|
|
- epoch: 0.8265
|
|
|
- step: 162
|
|
|
|
|
|
## 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: bitsandbytes
|
|
|
- _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: float16
|
|
|
- bnb_4bit_quant_storage: uint8
|
|
|
- load_in_4bit: True
|
|
|
- load_in_8bit: False
|
|
|
### Training hyperparameters
|
|
|
|
|
|
The following hyperparameters were used during training:
|
|
|
- learning_rate: 0.0001
|
|
|
- train_batch_size: 4
|
|
|
- eval_batch_size: 4
|
|
|
- seed: 42
|
|
|
- gradient_accumulation_steps: 4
|
|
|
- total_train_batch_size: 16
|
|
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
|
|
- lr_scheduler_type: cosine
|
|
|
- lr_scheduler_warmup_ratio: 0.05
|
|
|
- num_epochs: 2
|
|
|
- mixed_precision_training: Native AMP
|
|
|
|
|
|
### Framework versions
|
|
|
|
|
|
- PEFT 0.4.0
|
|
|
- Transformers 4.44.1
|
|
|
- Pytorch 2.4.0+cu118
|
|
|
- Datasets 2.21.0
|
|
|
- Tokenizers 0.19.1
|
|
|
|
