grandpythia-200k-70m

This model is a fine-tuned version of EleutherAI/pythia-70m-deduped on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.8419

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: 64
eval_batch_size: 64
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss
1.1766	0.01	68	1.2007
1.0903	0.02	136	1.1284
1.0809	0.03	204	1.0993
1.0928	0.04	272	1.0712
0.989	0.05	340	1.0473
1.0044	0.06	408	1.0373
0.985	0.07	476	1.0241
1.0272	0.08	544	1.0130
1.0295	0.09	612	1.0036
1.0172	0.1	680	0.9985
0.9582	0.11	748	0.9924
1.0342	0.12	816	0.9916
1.0053	0.13	884	0.9844
0.9321	0.14	952	0.9798
0.9473	0.15	1020	0.9727
0.9197	0.16	1088	0.9688
0.9827	0.17	1156	0.9632
0.9423	0.18	1224	0.9613
0.9662	0.19	1292	0.9578
0.9417	0.2	1360	0.9549
0.9501	0.21	1428	0.9461
0.9744	0.22	1496	0.9466
0.8693	0.23	1564	0.9394
0.9467	0.24	1632	0.9393
0.9274	0.25	1700	0.9362
0.8793	0.26	1768	0.9338
0.99	0.27	1836	0.9276
0.8983	0.28	1904	0.9291
0.9177	0.29	1972	0.9246
0.9586	0.3	2040	0.9224
0.9364	0.31	2108	0.9178
0.9248	0.32	2176	0.9175
0.9294	0.33	2244	0.9171
0.9142	0.34	2312	0.9136
0.9533	0.35	2380	0.9102
0.9193	0.36	2448	0.9094
0.9072	0.37	2516	0.9075
0.8927	0.38	2584	0.9043
0.9055	0.39	2652	0.9032
0.9276	0.4	2720	0.9030
0.8847	0.41	2788	0.8966
0.9449	0.42	2856	0.8963
0.8754	0.43	2924	0.8971
0.8612	0.44	2992	0.8935
0.9028	0.45	3060	0.8895
0.8641	0.46	3128	0.8925
0.8668	0.47	3196	0.8887
0.8935	0.48	3264	0.8863
0.8889	0.49	3332	0.8837
0.8854	0.5	3400	0.8849
0.8725	0.51	3468	0.8831
0.9425	0.52	3536	0.8796
0.8577	0.53	3604	0.8780
0.8281	0.54	3672	0.8747
0.9141	0.55	3740	0.8736
0.8684	0.56	3808	0.8738
0.8476	0.57	3876	0.8718
0.8761	0.58	3944	0.8735
0.8464	0.59	4012	0.8708
0.8732	0.6	4080	0.8681
0.9441	0.61	4148	0.8669
0.881	0.62	4216	0.8657
0.8635	0.63	4284	0.8640
0.827	0.64	4352	0.8625
0.9123	0.65	4420	0.8628
0.8557	0.66	4488	0.8605
0.8157	0.67	4556	0.8591
0.9008	0.68	4624	0.8580
0.8574	0.69	4692	0.8580
0.8374	0.7	4760	0.8563
0.8698	0.71	4828	0.8554
0.8817	0.72	4896	0.8545
0.8375	0.73	4964	0.8532
0.8504	0.74	5032	0.8524
0.8526	0.75	5100	0.8516
0.9306	0.76	5168	0.8511
0.7999	0.77	5236	0.8502
0.8337	0.78	5304	0.8495
0.7934	0.79	5372	0.8488
0.8159	0.8	5440	0.8480
0.7997	0.81	5508	0.8473
0.8909	0.82	5576	0.8470
0.852	0.83	5644	0.8461
0.8285	0.84	5712	0.8455
0.8437	0.85	5780	0.8448
0.8784	0.86	5848	0.8444
0.8123	0.87	5916	0.8440
0.8439	0.88	5984	0.8436
0.8847	0.89	6052	0.8433
0.8165	0.9	6120	0.8429
0.8405	0.91	6188	0.8427
0.8641	0.92	6256	0.8425
0.8536	0.93	6324	0.8424
0.8426	0.94	6392	0.8421
0.8547	0.95	6460	0.8421
0.8144	0.96	6528	0.8419
0.8475	0.97	6596	0.8419
0.8063	0.98	6664	0.8419
0.7943	0.99	6732	0.8419

Framework versions

Transformers 4.38.2
Pytorch 2.1.0+cu121
Datasets 2.18.0
Tokenizers 0.15.2

mlabonne/grandpythia-200k-70m

grandpythia-200k-70m

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

Seed this model