Mistral-7B-v0.1_cola_switrelu_log

This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.5679
• Accuracy: 0.7512

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: 1e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 2
• distributed_type: multi-GPU
• num_devices: 2
• gradient_accumulation_steps: 2
• total_train_batch_size: 256
• total_eval_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• training_steps: 750

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
5.6153	0.33	10	5.3822	0.3231
1.482	0.66	20	1.5553	0.6347
1.1295	0.98	30	1.0969	0.6251
0.9153	1.31	40	0.9182	0.6290
0.7577	1.64	50	0.8733	0.6500
0.8076	1.97	60	0.9483	0.6731
0.7947	2.3	70	0.9969	0.6424
0.9011	2.62	80	1.2108	0.6059
0.9677	2.95	90	1.4661	0.6261
0.8283	3.28	100	1.2659	0.5944
0.7966	3.61	110	0.7540	0.6673
0.7234	3.93	120	0.6601	0.6347
0.6215	4.26	130	0.6336	0.6635
0.6291	4.59	140	0.6215	0.6826
0.5683	4.92	150	0.6205	0.6663
0.636	5.25	160	0.6146	0.6913
0.6094	5.57	170	0.6047	0.6779
0.6069	5.9	180	0.6167	0.6481
0.5531	6.23	190	0.6094	0.6874
0.6018	6.56	200	0.6195	0.6711
0.5861	6.89	210	0.5953	0.6807
0.5658	7.21	220	0.6063	0.6865
0.6259	7.54	230	0.6302	0.6903
0.5858	7.87	240	0.6422	0.6894
0.5213	8.2	250	0.5937	0.6922
0.5589	8.52	260	0.5871	0.6961
0.5416	8.85	270	0.5912	0.6884
0.5449	9.18	280	0.5917	0.7028
0.6001	9.51	290	0.5889	0.6826
0.5101	9.84	300	0.5827	0.6913
0.5183	10.16	310	0.5768	0.7124
0.4761	10.49	320	0.5882	0.7143
0.4813	10.82	330	0.5971	0.6817
0.5052	11.15	340	0.5936	0.7028
0.5235	11.48	350	0.5949	0.6740
0.5136	11.8	360	0.5926	0.6894
0.5185	12.13	370	0.5784	0.7172
0.4815	12.46	380	0.5813	0.7085
0.5257	12.79	390	0.5880	0.7124
0.3868	13.11	400	0.6054	0.6942
0.468	13.44	410	0.5956	0.7028
0.4823	13.77	420	0.5949	0.7114
0.3725	14.1	430	0.5807	0.7095
0.3859	14.43	440	0.6012	0.7124
0.3893	14.75	450	0.5994	0.7239
0.4392	15.08	460	0.5814	0.7162
0.4174	15.41	470	0.6529	0.7181
0.3899	15.74	480	0.5867	0.7172
0.4336	16.07	490	0.6386	0.6884
0.3928	16.39	500	0.6296	0.7152
0.4384	16.72	510	0.6195	0.7028
0.3227	17.05	520	0.6315	0.7114
0.367	17.38	530	0.6490	0.7210
0.3692	17.7	540	0.6184	0.7152
0.3205	18.03	550	0.6415	0.7124
0.3596	18.36	560	0.6559	0.7191
0.37	18.69	570	0.6479	0.7133

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.1+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0