fine-tuning-Phi2-with-webglm-qa-with-lora_2

Gunslinger3D/fine-tuning-Phi2-with-webglm-qa-with-lora_2

fine-tuning-Phi2-with-webglm-qa-with-lora_2 is a machine learning model from Gunslinger3D. 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 mit.

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

Downloads · 30 days

5% of all-time downloads

All-time downloads

Public

Repo size

94.4 MB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors94.4 MB · 97%

At a glance

Library: peft
License: mit
Access: Public
Created: Mar 5, 2024
Updated: Mar 5, 2024
SHA: 30062cc3

Base models

microsoft/phi-2

Model card

From the Hugging Face model README

fine-tuning-Phi2-with-webglm-qa-with-lora_2

This model is a fine-tuned version of microsoft/phi-2 on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.0749

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: 5e-05
train_batch_size: 2
eval_batch_size: 2
seed: 42
gradient_accumulation_steps: 5
total_train_batch_size: 10
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 100
training_steps: 1000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.2	10	8.2244
No log	0.4	20	7.4785
No log	0.6	30	5.5477
No log	0.8	40	2.9270
5.8579	1.0	50	0.6602
5.8579	1.2	60	0.5707
5.8579	1.39	70	0.5056
5.8579	1.59	80	0.4537
5.8579	1.79	90	0.4048
0.4258	1.99	100	0.3539
0.4258	2.19	110	0.3117
0.4258	2.39	120	0.2848
0.4258	2.59	130	0.2589
0.4258	2.79	140	0.2344
0.2253	2.99	150	0.2160
0.2253	3.19	160	0.2037
0.2253	3.39	170	0.1890
0.2253	3.59	180	0.1794
0.2253	3.78	190	0.1695
0.1573	3.98	200	0.1607
0.1573	4.18	210	0.1544
0.1573	4.38	220	0.1518
0.1573	4.58	230	0.1441
0.1573	4.78	240	0.1365
0.1251	4.98	250	0.1302
0.1251	5.18	260	0.1286
0.1251	5.38	270	0.1258
0.1251	5.58	280	0.1228
0.1251	5.78	290	0.1203
0.1059	5.98	300	0.1159
0.1059	6.18	310	0.1116
0.1059	6.37	320	0.1112
0.1059	6.57	330	0.1092
0.1059	6.77	340	0.1046
0.0905	6.97	350	0.1032
0.0905	7.17	360	0.1028
0.0905	7.37	370	0.1005
0.0905	7.57	380	0.1011
0.0905	7.77	390	0.0991
0.0816	7.97	400	0.0973
0.0816	8.17	410	0.0965
0.0816	8.37	420	0.0954
0.0816	8.57	430	0.0949
0.0816	8.76	440	0.0938
0.0722	8.96	450	0.0915
0.0722	9.16	460	0.0907
0.0722	9.36	470	0.0900
0.0722	9.56	480	0.0893
0.0722	9.76	490	0.0877
0.0656	9.96	500	0.0872
0.0656	10.16	510	0.0868
0.0656	10.36	520	0.0867
0.0656	10.56	530	0.0874
0.0656	10.76	540	0.0863
0.0614	10.96	550	0.0849
0.0614	11.16	560	0.0834
0.0614	11.35	570	0.0829
0.0614	11.55	580	0.0827
0.0614	11.75	590	0.0817
0.0553	11.95	600	0.0817
0.0553	12.15	610	0.0824
0.0553	12.35	620	0.0826
0.0553	12.55	630	0.0810
0.0553	12.75	640	0.0814
0.053	12.95	650	0.0804
0.053	13.15	660	0.0807
0.053	13.35	670	0.0792
0.053	13.55	680	0.0792
0.053	13.75	690	0.0792
0.0495	13.94	700	0.0788
0.0495	14.14	710	0.0783
0.0495	14.34	720	0.0784
0.0495	14.54	730	0.0779
0.0495	14.74	740	0.0776
0.0477	14.94	750	0.0773
0.0477	15.14	760	0.0787
0.0477	15.34	770	0.0772
0.0477	15.54	780	0.0763
0.0477	15.74	790	0.0759
0.0456	15.94	800	0.0766
0.0456	16.14	810	0.0770
0.0456	16.33	820	0.0774
0.0456	16.53	830	0.0770
0.0456	16.73	840	0.0764
0.0438	16.93	850	0.0754
0.0438	17.13	860	0.0759
0.0438	17.33	870	0.0762
0.0438	17.53	880	0.0758
0.0438	17.73	890	0.0761
0.0415	17.93	900	0.0758
0.0415	18.13	910	0.0754
0.0415	18.33	920	0.0754
0.0415	18.53	930	0.0753
0.0415	18.73	940	0.0751
0.0408	18.92	950	0.0752
0.0408	19.12	960	0.0750
0.0408	19.32	970	0.0749
0.0408	19.52	980	0.0749
0.0408	19.72	990	0.0749
0.0403	19.92	1000	0.0749

Framework versions

PEFT 0.7.1
Transformers 4.36.2
Pytorch 2.0.0
Datasets 2.15.0
Tokenizers 0.15.0

Library: peft
License: mit
Created: Mar 5, 2024
Updated: Mar 5, 2024