sft_gpt7b_domar_pretuned

thorirhrafn/sft_gpt7b_domar_pretuned

sft_gpt7b_domar_pretuned is a machine learning model from thorirhrafn. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for peft. The card lists the license as other.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

17% of all-time downloads

All-time downloads

Public

Repo size

809 MB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors268 MB · 99%

At a glance

Library: peft
License: other
Access: Public
Created: Apr 9, 2024
Updated: Apr 10, 2024
SHA: 27bad3e0

Base models

AI-Sweden-Models/gpt-sw3-6.7b

Model card

From the Hugging Face model README

sft_gpt7b_domar_pretuned

This model is a fine-tuned version of AI-Sweden-Models/gpt-sw3-6.7b on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 1.5598

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

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: 4
total_train_batch_size: 4
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss
1.6679	0.02	500	1.6397
1.7258	0.04	1000	1.6302
1.6602	0.06	1500	1.6228
1.6731	0.08	2000	1.6173
1.7399	0.09	2500	1.6155
1.8291	0.11	3000	1.6106
1.7113	0.13	3500	1.6078
1.6768	0.15	4000	1.6037
1.7504	0.17	4500	1.6028
1.598	0.19	5000	1.6003
1.5689	0.21	5500	1.5974
1.6727	0.23	6000	1.5961
1.5689	0.25	6500	1.5952
1.6331	0.26	7000	1.5931
1.6459	0.28	7500	1.5921
1.6334	0.3	8000	1.5918
1.6803	0.32	8500	1.5894
1.6182	0.34	9000	1.5879
1.693	0.36	9500	1.5866
1.6276	0.38	10000	1.5857
1.612	0.4	10500	1.5859
1.6412	0.42	11000	1.5843
1.6827	0.43	11500	1.5824
1.584	0.45	12000	1.5826
1.591	0.47	12500	1.5817
1.6641	0.49	13000	1.5805
1.6555	0.51	13500	1.5799
1.689	0.53	14000	1.5798
1.6216	0.55	14500	1.5785
1.6271	0.57	15000	1.5780
1.6898	0.59	15500	1.5771
1.6752	0.6	16000	1.5762
1.5884	0.62	16500	1.5761
1.6094	0.64	17000	1.5755
1.5202	0.66	17500	1.5749
1.6506	0.68	18000	1.5744
1.6805	0.7	18500	1.5736
1.6421	0.72	19000	1.5732
1.652	0.74	19500	1.5731
1.5729	0.76	20000	1.5722
1.6231	0.77	20500	1.5715
1.6527	0.79	21000	1.5710
1.656	0.81	21500	1.5705
1.5076	0.83	22000	1.5708
1.6925	0.85	22500	1.5700
1.6761	0.87	23000	1.5701
1.6376	0.89	23500	1.5697
1.696	0.91	24000	1.5686
1.6921	0.93	24500	1.5688
1.6896	0.94	25000	1.5681
1.7896	0.96	25500	1.5678
1.6342	0.98	26000	1.5679
1.6001	1.0	26500	1.5679
1.7183	1.02	27000	1.5678
1.5685	1.04	27500	1.5675
1.5349	1.06	28000	1.5672
1.6439	1.08	28500	1.5677
1.6201	1.1	29000	1.5670
1.6209	1.11	29500	1.5664
1.5495	1.13	30000	1.5665
1.5573	1.15	30500	1.5661
1.6094	1.17	31000	1.5660
1.625	1.19	31500	1.5662
1.5404	1.21	32000	1.5656
1.547	1.23	32500	1.5655
1.5997	1.25	33000	1.5648
1.6287	1.27	33500	1.5651
1.4998	1.28	34000	1.5650
1.7069	1.3	34500	1.5642
1.5453	1.32	35000	1.5643
1.5378	1.34	35500	1.5640
1.5702	1.36	36000	1.5643
1.6593	1.38	36500	1.5641
1.4526	1.4	37000	1.5641
1.5875	1.42	37500	1.5635
1.7064	1.44	38000	1.5632
1.6517	1.45	38500	1.5629
1.5637	1.47	39000	1.5630
1.5557	1.49	39500	1.5632
1.6615	1.51	40000	1.5626
1.5869	1.53	40500	1.5629
1.6263	1.55	41000	1.5622
1.5958	1.57	41500	1.5624
1.5646	1.59	42000	1.5620
1.5605	1.61	42500	1.5620
1.5753	1.62	43000	1.5621
1.6315	1.64	43500	1.5618
1.6351	1.66	44000	1.5616
1.4516	1.68	44500	1.5615
1.6654	1.7	45000	1.5616
1.4796	1.72	45500	1.5613
1.7079	1.74	46000	1.5613
1.6877	1.76	46500	1.5613
1.5899	1.78	47000	1.5612
1.5419	1.79	47500	1.5609
1.5972	1.81	48000	1.5611
1.6402	1.83	48500	1.5609
1.6036	1.85	49000	1.5607
1.5839	1.87	49500	1.5607
1.6727	1.89	50000	1.5608
1.5385	1.91	50500	1.5605
1.5856	1.93	51000	1.5608
1.6168	1.95	51500	1.5604
1.5426	1.96	52000	1.5605
1.5768	1.98	52500	1.5603
1.519	2.0	53000	1.5606
1.615	2.02	53500	1.5607
1.6096	2.04	54000	1.5606
1.5881	2.06	54500	1.5604
1.5782	2.08	55000	1.5604
1.6988	2.1	55500	1.5604
1.6284	2.12	56000	1.5604
1.6219	2.13	56500	1.5605
1.5288	2.15	57000	1.5604
1.57	2.17	57500	1.5603
1.6524	2.19	58000	1.5605
1.5774	2.21	58500	1.5602
1.5434	2.23	59000	1.5601
1.4985	2.25	59500	1.5602
1.4937	2.27	60000	1.5602
1.5134	2.29	60500	1.5601
1.5064	2.3	61000	1.5601
1.6091	2.32	61500	1.5601
1.6257	2.34	62000	1.5600
1.6497	2.36	62500	1.5601
1.5469	2.38	63000	1.5599
1.5453	2.4	63500	1.5600
1.5256	2.42	64000	1.5599
1.5616	2.44	64500	1.5600
1.6449	2.46	65000	1.5600
1.6298	2.47	65500	1.5598
1.697	2.49	66000	1.5599
1.5351	2.51	66500	1.5598
1.5463	2.53	67000	1.5599
1.6256	2.55	67500	1.5598
1.5567	2.57	68000	1.5598
1.6036	2.59	68500	1.5599
1.5113	2.61	69000	1.5598
1.6975	2.63	69500	1.5598
1.69	2.64	70000	1.5599
1.5828	2.66	70500	1.5598
1.6462	2.68	71000	1.5598
1.5645	2.7	71500	1.5598
1.5385	2.72	72000	1.5599
1.6244	2.74	72500	1.5599
1.5805	2.76	73000	1.5599
1.6334	2.78	73500	1.5599
1.5254	2.8	74000	1.5598
1.5892	2.81	74500	1.5599
1.68	2.83	75000	1.5599
1.5866	2.85	75500	1.5598
1.5692	2.87	76000	1.5598
1.4843	2.89	76500	1.5598
1.633	2.91	77000	1.5598
1.6205	2.93	77500	1.5598
1.5802	2.95	78000	1.5598
1.5723	2.97	78500	1.5598
1.6153	2.98	79000	1.5598

Framework versions

PEFT 0.8.2
Transformers 4.38.1
Pytorch 2.2.0+cu118
Datasets 2.17.1
Tokenizers 0.15.2

Library: peft
License: other
Created: Apr 9, 2024
Updated: Apr 10, 2024