README.md

metadata

license: other
tags:
  - generated_from_trainer
datasets:
  - AlekseyKorshuk/dalio-synthetic-io
metrics:
  - accuracy
model-index:
  - name: dalio-synthetic-io-1.3b
    results:
      - task:
          name: Causal Language Modeling
          type: text-generation
        dataset:
          name: AlekseyKorshuk/dalio-synthetic-io
          type: AlekseyKorshuk/dalio-synthetic-io
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.06235550115599075

dalio-synthetic-io-1.3b

This model is a fine-tuned version of facebook/opt-1.3b on the AlekseyKorshuk/dalio-synthetic-io dataset. It achieves the following results on the evaluation set:

• Loss: 2.5957
• Accuracy: 0.0624

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: 3e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• total_train_batch_size: 16
• total_eval_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.6941	0.05	1	2.6543	0.0622
2.6914	0.11	2	2.6543	0.0622
2.6003	0.16	3	2.6016	0.0627
2.5603	0.21	4	2.5703	0.0627
2.6072	0.26	5	2.5508	0.0630
2.5444	0.32	6	2.5469	0.0628
2.4467	0.37	7	2.5508	0.0629
2.5452	0.42	8	2.5508	0.0629
2.6128	0.47	9	2.5449	0.0631
2.4568	0.53	10	2.5391	0.0627
2.5098	0.58	11	2.5352	0.0628
2.6047	0.63	12	2.5234	0.0631
2.5022	0.68	13	2.5156	0.0630
2.605	0.74	14	2.5078	0.0633
2.6055	0.79	15	2.5020	0.0634
2.5061	0.84	16	2.4961	0.0632
2.4348	0.89	17	2.4902	0.0631
2.6284	0.95	18	2.4883	0.0632
2.5574	1.0	19	2.4863	0.0631
2.0814	1.05	20	2.4844	0.0633
2.0636	1.11	21	2.4844	0.0635
1.9459	1.16	22	2.4844	0.0635
2.0527	1.21	23	2.4883	0.0634
1.8881	1.26	24	2.4961	0.0635
1.8668	1.32	25	2.5117	0.0636
2.0375	1.37	26	2.5293	0.0636
1.9402	1.42	27	2.5449	0.0632
1.6086	1.47	28	2.5586	0.0633
1.8185	1.53	29	2.5645	0.0632
1.7324	1.58	30	2.5605	0.0630
1.9285	1.63	31	2.5527	0.0628
1.8031	1.68	32	2.5449	0.0631
1.7321	1.74	33	2.5352	0.0630
1.7802	1.79	34	2.5254	0.0631
2.0637	1.84	35	2.5156	0.0632
1.8159	1.89	36	2.5078	0.0633
1.7142	1.95	37	2.5039	0.0634
1.8793	2.0	38	2.5	0.0634
1.6914	2.05	39	2.5020	0.0636
1.411	2.11	40	2.5039	0.0639
1.4182	2.16	41	2.5098	0.0639
1.6223	2.21	42	2.5176	0.0638
1.623	2.26	43	2.5273	0.0634
1.5748	2.32	44	2.5371	0.0634
1.7166	2.37	45	2.5469	0.0631
1.3432	2.42	46	2.5566	0.0630
1.5325	2.47	47	2.5645	0.0631
1.5076	2.53	48	2.5723	0.0629
1.6636	2.58	49	2.5781	0.0627
1.2897	2.63	50	2.5840	0.0627
1.4559	2.68	51	2.5879	0.0627
1.3904	2.74	52	2.5898	0.0627
1.4961	2.79	53	2.5918	0.0626
1.5276	2.84	54	2.5938	0.0625
1.3479	2.89	55	2.5957	0.0625
1.4094	2.95	56	2.5957	0.0624
1.5486	3.0	57	2.5957	0.0624

Framework versions

• Transformers 4.25.0.dev0
• Pytorch 1.12.1+cu113
• Datasets 2.3.2
• Tokenizers 0.12.1