You can deploy this model and start earning money today!

RichardErkhov/
BramVanroy_-_llama2-13b-ft-mc4_nl_cleaned_tiny-gguf

Name: BramVanroy_-_llama2-13b-ft-mc4_nl_cleaned_tiny-gguf
Availability: InStock

Views

No views yet

Quantization made by Richard Erkhov.

Github

Discord

Request more models

llama2-13b-ft-mc4_nl_cleaned_tiny - GGUF

Model creator: https://huggingface.co/BramVanroy/
Original model: https://huggingface.co/BramVanroy/llama2-13b-ft-mc4_nl_cleaned_tiny/

Name	Quant method	Size
llama2-13b-ft-mc4_nl_cleaned_tiny.Q2_K.gguf	Q2_K	4.52GB
llama2-13b-ft-mc4_nl_cleaned_tiny.IQ3_XS.gguf	IQ3_XS	4.99GB
llama2-13b-ft-mc4_nl_cleaned_tiny.IQ3_S.gguf	IQ3_S	5.27GB
llama2-13b-ft-mc4_nl_cleaned_tiny.Q3_K_S.gguf	Q3_K_S	5.27GB
llama2-13b-ft-mc4_nl_cleaned_tiny.IQ3_M.gguf	IQ3_M	5.57GB
llama2-13b-ft-mc4_nl_cleaned_tiny.Q3_K.gguf	Q3_K	5.9GB
llama2-13b-ft-mc4_nl_cleaned_tiny.Q3_K_M.gguf	Q3_K_M	5.9GB
llama2-13b-ft-mc4_nl_cleaned_tiny.Q3_K_L.gguf	Q3_K_L	6.45GB
llama2-13b-ft-mc4_nl_cleaned_tiny.IQ4_XS.gguf	IQ4_XS	6.54GB
llama2-13b-ft-mc4_nl_cleaned_tiny.Q4_0.gguf	Q4_0	6.86GB
llama2-13b-ft-mc4_nl_cleaned_tiny.IQ4_NL.gguf	IQ4_NL	6.9GB
llama2-13b-ft-mc4_nl_cleaned_tiny.Q4_K_S.gguf	Q4_K_S	6.91GB
llama2-13b-ft-mc4_nl_cleaned_tiny.Q4_K.gguf	Q4_K	7.33GB
llama2-13b-ft-mc4_nl_cleaned_tiny.Q4_K_M.gguf	Q4_K_M	7.33GB
llama2-13b-ft-mc4_nl_cleaned_tiny.Q4_1.gguf	Q4_1	7.61GB
llama2-13b-ft-mc4_nl_cleaned_tiny.Q5_0.gguf	Q5_0	8.36GB
llama2-13b-ft-mc4_nl_cleaned_tiny.Q5_K_S.gguf	Q5_K_S	8.36GB
llama2-13b-ft-mc4_nl_cleaned_tiny.Q5_K.gguf	Q5_K	8.6GB
llama2-13b-ft-mc4_nl_cleaned_tiny.Q5_K_M.gguf	Q5_K_M	8.6GB
llama2-13b-ft-mc4_nl_cleaned_tiny.Q5_1.gguf	Q5_1	9.1GB
llama2-13b-ft-mc4_nl_cleaned_tiny.Q6_K.gguf	Q6_K	9.95GB
llama2-13b-ft-mc4_nl_cleaned_tiny.Q8_0.gguf	Q8_0	12.88GB

Original model description:

license: apache-2.0 base_model: meta-llama/Llama-2-13b-hf tags:

generated_from_trainer
llama
lora
adapters datasets:
yhavinga/mc4_nl_cleaned language:
nl model-index:
name: llama2-13b-ft-mc4_nl_cleaned_tiny results: []

llama2-13b-ft-mc4_nl_cleaned_tiny

This model is a fine-tuned version of meta-llama/Llama-2-13b-hf on the yhavinga/mc4_nl_cleaned dataset (tiny partition) on a context of 4096 tokens. See the original meta-llama/Llama-2-13b-hf for more information, intended use, and biases.

If you use this model or refer to it, please use the following citation:

Vanroy, B. (2023). Language Resources for Dutch Large Language Modelling. https://arxiv.org/abs/2312.12852

bibtext

@article{vanroy2023language,
  title={Language Resources for {Dutch} Large Language Modelling},
  author={Vanroy, Bram},
  journal={arXiv preprint arXiv:2312.12852},
  year={2023}
}

Intended uses & limitations

While Llama 2 already contains some proficiency in Dutch, this finetune is intended to improve the fluency of Dutch (not increase its knowledge). It is therefore intended as a generative model for Dutch language. The biases, shortcomings and intended uses are otherwise the same as those of the original model. The model can be used for generative tasks or finetuned further on other tasks such as summarization, adaptation, instruction or chat finetuning.

Training and evaluation data

Trained on the yhavinga/mc4_nl_cleaned dataset (tiny partition) for one epoch. The canonical validation split was not used but instead 5% of train was used as validation.

Training procedure

Trained with LoRA targetting ["q_proj", "v_proj"] in 4 bit and merged before upload. Trained with Flash Attention as borrowed from here.

The adapters are in the adapters branch.

Initial training investigation on the Tier-1 HPC of Vlaams Supercomputer Centrum (VSC) and training on our own server of 4x 3090s.

Training hyperparameters

The following hyperparameters were used during training in the HPC investigation:

learning_rate: 0.0003
train_batch_size: 12
eval_batch_size: 12
seed: 42
distributed_type: multi-GPU
num_devices: 16
gradient_accumulation_steps: 6
total_train_batch_size: 1152
total_eval_batch_size: 192
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.03
num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss
1.8784	0.09	90	1.8820
1.8344	0.19	180	1.8542
1.8351	0.28	270	1.8355
1.8206	0.37	360	1.8212
1.8021	0.47	450	1.8088
1.8102	0.56	540	1.7982
1.7991	0.65	630	1.7890
1.7788	0.74	720	1.7811
1.7915	0.84	810	1.7742
1.7715	0.93	900	1.7676

Framework versions

Transformers 4.31.0.dev0
Pytorch 2.0.1+cu117
Datasets 2.13.1
Tokenizers 0.13.3

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	46.81
ARC (25-shot)	59.3
HellaSwag (10-shot)	82.04
MMLU (5-shot)	54.67
TruthfulQA (0-shot)	38.03
Winogrande (5-shot)	77.27
GSM8K (5-shot)	10.31
DROP (3-shot)	6.08

RichardErkhov/BramVanroy_-_llama2-13b-ft-mc4_nl_cleaned_tiny-gguf