This model is a fine-tuned version of ybelkada/falcon-7b-sharded-bf16 on the BramVanroy/alpaca-dolly-dutch dataset. See the original Falcon 7B model for more information, intended use, and biases.

Intended uses & limitations

This model is intended as a (poor) baseline for Dutch generative LLMs. It by no means aims to provide SOTA performance and is specifically intended for research purposes, and an opportunity for me to test hyperparameters and stability.

Importantly, the original Falcon 7B model was only trained on English and French. Therefore, Dutch generations should be taken with a massive grain of salt.

Training and evaluation data

Trained on the synthetic BramVanroy/alpaca-dolly-dutch instruction dataset. Therefore, commercial use of this model is forbidden. The model is intended for research purposes only.

Training procedure

Trained with LoRA and merged before upload. The adapters are in the adapters branch.

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 0.001
train_batch_size: 32
eval_batch_size: 32
seed: 42
distributed_type: multi-GPU
num_devices: 16
total_train_batch_size: 512
total_eval_batch_size: 512
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.03
num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss
1.8677	0.16	20	1.6766
1.5635	0.32	40	1.5643
1.6353	0.48	60	1.4980
1.5166	0.65	80	1.4516
1.4287	0.81	100	1.4096
1.5791	0.97	120	1.3802
1.3911	1.13	140	1.3633
1.356	1.29	160	1.3419
1.2524	1.45	180	1.3263
1.4224	1.61	200	1.3056
1.2266	1.77	220	1.2897
1.3242	1.94	240	1.2785
1.03	2.1	260	1.2957
1.1643	2.26	280	1.2970
1.1492	2.42	300	1.2779
1.0679	2.58	320	1.2770
1.2695	2.74	340	1.2658
1.0439	2.9	360	1.2612
0.9453	3.06	380	1.3157
0.8494	3.23	400	1.3189
1.0745	3.39	420	1.3073
0.8679	3.55	440	1.3019
1.0569	3.71	460	1.2955
1.0186	3.87	480	1.2890
0.8413	4.03	500	1.3445

Framework versions

Transformers 4.30.1
Pytorch 2.0.1+cu117
Datasets 2.13.1
Tokenizers 0.13.3

BramVanroy/falcon-7b-ft-alpaca-dolly-dutch

falcon-7b-ft-alpaca-dolly-dutch

Model description