README.md

metadata

license: other
tags:
  - generated_from_trainer
datasets:
  - AlekseyKorshuk/dalio-all-io
metrics:
  - accuracy
model-index:
  - name: dalio-all-io-125m-3-epoch
    results:
      - task:
          name: Causal Language Modeling
          type: text-generation
        dataset:
          name: AlekseyKorshuk/dalio-all-io
          type: AlekseyKorshuk/dalio-all-io
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.049654305468258955

dalio-all-io-125m-3-epoch

This model is a fine-tuned version of facebook/opt-125m on the AlekseyKorshuk/dalio-all-io dataset. It achieves the following results on the evaluation set:

• Loss: 2.7656
• Accuracy: 0.0497

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
3.1406	0.03	1	3.0762	0.0451
3.074	0.07	2	3.0762	0.0451
3.0557	0.1	3	3.0762	0.0451
3.2166	0.14	4	3.0176	0.0457
3.0989	0.17	5	2.9922	0.0460
3.0732	0.21	6	2.9746	0.0464
3.0867	0.24	7	2.9629	0.0463
2.979	0.28	8	2.9512	0.0467
3.1838	0.31	9	2.9414	0.0467
2.9399	0.34	10	2.9336	0.0467
2.926	0.38	11	2.9258	0.0471
3.2144	0.41	12	2.9199	0.0473
2.978	0.45	13	2.9141	0.0474
3.0076	0.48	14	2.9082	0.0476
2.9897	0.52	15	2.9023	0.0477
2.8831	0.55	16	2.8945	0.0479
2.9749	0.59	17	2.8867	0.0479
2.9431	0.62	18	2.8828	0.0478
3.0498	0.66	19	2.8770	0.0479
2.9409	0.69	20	2.8711	0.0479
2.96	0.72	21	2.8672	0.0480
3.0767	0.76	22	2.8633	0.0478
2.772	0.79	23	2.8594	0.0479
3.0574	0.83	24	2.8535	0.0480
2.8137	0.86	25	2.8496	0.0480
2.8872	0.9	26	2.8438	0.0483
3.0085	0.93	27	2.8398	0.0484
2.9165	0.97	28	2.8359	0.0485
2.8525	1.0	29	2.8340	0.0486
2.7759	1.03	30	2.8301	0.0485
2.7312	1.07	31	2.8281	0.0485
2.6641	1.1	32	2.8262	0.0487
2.7896	1.14	33	2.8242	0.0486
2.7878	1.17	34	2.8223	0.0487
2.4028	1.21	35	2.8203	0.0487
2.5618	1.24	36	2.8184	0.0488
2.6697	1.28	37	2.8164	0.0488
2.6333	1.31	38	2.8145	0.0487
2.4897	1.34	39	2.8125	0.0486
2.4908	1.38	40	2.8105	0.0487
2.6926	1.41	41	2.8086	0.0488
2.6602	1.45	42	2.8066	0.0489
2.8054	1.48	43	2.8047	0.0489
2.5532	1.52	44	2.8047	0.0490
2.4756	1.55	45	2.8027	0.0491
2.6123	1.59	46	2.8008	0.0491
2.5117	1.62	47	2.7988	0.0490
2.5552	1.66	48	2.7969	0.0490
2.5122	1.69	49	2.7949	0.0490
2.5593	1.72	50	2.7930	0.0491
2.5759	1.76	51	2.7910	0.0491
2.5535	1.79	52	2.7891	0.0493
2.6531	1.83	53	2.7871	0.0494
2.5701	1.86	54	2.7852	0.0495
2.6621	1.9	55	2.7832	0.0497
2.532	1.93	56	2.7812	0.0496
2.5928	1.97	57	2.7793	0.0497
2.5486	2.0	58	2.7754	0.0497
2.5009	2.03	59	2.7734	0.0497
2.4346	2.07	60	2.7734	0.0498
2.3259	2.1	61	2.7715	0.0497
2.3569	2.14	62	2.7695	0.0498
2.5898	2.17	63	2.7695	0.0498
2.3657	2.21	64	2.7676	0.0498
2.4875	2.24	65	2.7676	0.0498
2.4392	2.28	66	2.7676	0.0497
2.3595	2.31	67	2.7656	0.0497
2.4757	2.34	68	2.7656	0.0498
2.4617	2.38	69	2.7656	0.0498
2.3376	2.41	70	2.7656	0.0499
2.3129	2.45	71	2.7656	0.0498
2.5703	2.48	72	2.7656	0.0498
2.3491	2.52	73	2.7656	0.0498
2.3484	2.55	74	2.7656	0.0498
2.3782	2.59	75	2.7656	0.0497
2.4033	2.62	76	2.7656	0.0498
2.3821	2.66	77	2.7656	0.0498
2.39	2.69	78	2.7656	0.0498
2.3984	2.72	79	2.7656	0.0497
2.3936	2.76	80	2.7656	0.0498
2.4414	2.79	81	2.7656	0.0497
2.4727	2.83	82	2.7656	0.0497
2.3192	2.86	83	2.7656	0.0497
2.4365	2.9	84	2.7656	0.0497
2.5042	2.93	85	2.7656	0.0497
2.4746	2.97	86	2.7656	0.0497
2.5383	3.0	87	2.7656	0.0497

Framework versions

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