library_name: peft
license: cc-by-nc-4.0
datasets:
- BramVanroy/alpaca-cleaned-dutch
language:
- nl
pipeline_tag: text-generation
tags:
- llama
- alpaca
- Transformers
open_llama_7b_alpaca_clean_dutch_qlora
Model description
This adapter model is a fine-tuned version of openlm-research/open_llama_7b on the BramVanroy/alpaca-cleaned-dutch dataset.
See openlm-research/open_llama_7b for all information about the base model.
Model usage
A basic example of how to use the finetuned model.
import torch
from peft import PeftModel, PeftConfig
from transformers import AutoModelForCausalLM, AutoTokenizer
model_name = "robinsmits/open_llama_7b_alpaca_clean_dutch_qlora"
tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast = False, add_eos_token = True)
config = PeftConfig.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(config.base_model_name_or_path, load_in_8bit = True, device_map = "auto")
model = PeftModel.from_pretrained(model, model_name)
prompt = "### Instructie:\nWat zijn de drie belangrijkste softwareonderdelen die worden gebruikt bij webontwikkeling?\n\n### Antwoord:\n"
inputs = tokenizer(prompt, return_tensors = "pt", truncation = True).input_ids.cuda()
sample = model.generate(input_ids = inputs, max_new_tokens = 512, num_beams = 2, early_stopping = True, eos_token_id = tokenizer.eos_token_id)
output = tokenizer.decode(sample[0], skip_special_tokens = True)
print(output.split(prompt)[1])
For more extensive usage and a lot of generated samples (both good and bad samples) see the following Inference Notebook
Intended uses & limitations
The open_llama_7b model was primarily trained on the English language. Part of the dataset was a Wikipedia dump containing pages in 20 languages. Dutch was one of those languages. Given the size of the total dataset and the wikipedia part the Dutch language was very likely less than 0.5% of the total data.
The primary intention of this model is to explore the use of the Dutch language in combination with an Open LLM.
Training and evaluation data
This model was trained on the BramVanroy/alpaca-cleaned-dutch dataset.
Commercial use is forbidden. This model is intended for research only.
Training procedure
This model was finetuned with a QLoRA setup on a Google Colab A100 GPU in about 6.5 hours.
The notebook used for training can be found here: Training Notebook
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0002
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 64
- training_steps: 1536
The following bitsandbytes quantization config was used during training:
- 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
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.1240 | 1.0 | 768 | 1.1227 |
| 1.0177 | 2.0 | 1536 | 1.0645 |
Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3
- PEFT 0.4.0.dev0