scenario-KD-PR-MSV-D2_data-AmazonScience_massive_all_1_144

This model is a fine-tuned version of haryoaw/scenario-MDBT-TCR_data-AmazonScience_massive_all_1_1 on the massive dataset. It achieves the following results on the evaluation set:

• Loss: 1.4509
• Accuracy: 0.8576
• F1: 0.8345

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: 32
• eval_batch_size: 32
• seed: 44
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
1.4358	0.27	5000	1.5925	0.8182	0.7743
1.291	0.53	10000	1.5425	0.8257	0.7947
1.207	0.8	15000	1.4865	0.8367	0.8041
1.0927	1.07	20000	1.4621	0.8439	0.8136
1.0824	1.34	25000	1.4744	0.8388	0.8066
1.0564	1.6	30000	1.4548	0.8454	0.8143
1.0461	1.87	35000	1.4542	0.8452	0.8125
0.9801	2.14	40000	1.4552	0.8461	0.8179
0.9812	2.41	45000	1.4601	0.8467	0.8224
0.9878	2.67	50000	1.4641	0.8454	0.8238
0.9767	2.94	55000	1.4509	0.8480	0.8259
0.9201	3.21	60000	1.4822	0.8429	0.8192
0.9235	3.47	65000	1.4676	0.8451	0.8237
0.9299	3.74	70000	1.4711	0.8456	0.8216
0.9204	4.01	75000	1.4883	0.8425	0.8188
0.8934	4.28	80000	1.4950	0.8425	0.8188
0.8967	4.54	85000	1.4874	0.8444	0.8193
0.9039	4.81	90000	1.4726	0.8480	0.8238
0.8764	5.08	95000	1.4938	0.8446	0.8197
0.8737	5.34	100000	1.4776	0.8484	0.8282
0.8766	5.61	105000	1.4764	0.8488	0.8257
0.8836	5.88	110000	1.4775	0.8464	0.8221
0.861	6.15	115000	1.4765	0.8486	0.8244
0.8595	6.41	120000	1.4726	0.8502	0.8276
0.8687	6.68	125000	1.4702	0.8502	0.8230
0.8641	6.95	130000	1.4841	0.8480	0.8272
0.8474	7.22	135000	1.4871	0.8485	0.8229
0.8503	7.48	140000	1.4910	0.8456	0.8181
0.8548	7.75	145000	1.4877	0.8484	0.8223
0.8404	8.02	150000	1.4682	0.8509	0.8275
0.8481	8.28	155000	1.4994	0.8471	0.8226
0.8427	8.55	160000	1.4933	0.8476	0.8240
0.8418	8.82	165000	1.4812	0.8484	0.8201
0.8323	9.09	170000	1.4808	0.8500	0.8254
0.8353	9.35	175000	1.4962	0.8462	0.8196
0.8331	9.62	180000	1.4947	0.8459	0.8219
0.8384	9.89	185000	1.4916	0.8476	0.8228
0.8202	10.15	190000	1.4909	0.8476	0.8232
0.8308	10.42	195000	1.4897	0.8498	0.8233
0.8283	10.69	200000	1.4882	0.8476	0.8246
0.8303	10.96	205000	1.4837	0.8477	0.8228
0.8154	11.22	210000	1.4930	0.8489	0.8233
0.8269	11.49	215000	1.5044	0.8454	0.8191
0.8172	11.76	220000	1.4946	0.8484	0.8237
0.8071	12.03	225000	1.4824	0.8513	0.8271
0.8124	12.29	230000	1.4778	0.8514	0.8262
0.8187	12.56	235000	1.4866	0.8478	0.8236
0.8153	12.83	240000	1.5002	0.8469	0.8257
0.8117	13.09	245000	1.4883	0.8492	0.8235
0.807	13.36	250000	1.4906	0.8511	0.8293
0.8151	13.63	255000	1.4702	0.8526	0.8295
0.8107	13.9	260000	1.4772	0.8513	0.8237
0.8012	14.16	265000	1.4784	0.8518	0.8245
0.8039	14.43	270000	1.4933	0.8485	0.8258
0.8031	14.7	275000	1.4811	0.8519	0.8252
0.8058	14.96	280000	1.4802	0.8508	0.8263
0.8014	15.23	285000	1.4919	0.8498	0.8240
0.8002	15.5	290000	1.4780	0.8515	0.8247
0.8043	15.77	295000	1.4755	0.8519	0.8237
0.7967	16.03	300000	1.4765	0.8516	0.8296
0.7958	16.3	305000	1.4910	0.8509	0.8260
0.8032	16.57	310000	1.4795	0.8499	0.8221
0.8002	16.84	315000	1.4864	0.8497	0.8235
0.7938	17.1	320000	1.4832	0.8509	0.8280
0.7981	17.37	325000	1.4866	0.8508	0.8285
0.798	17.64	330000	1.4922	0.8496	0.8274
0.8004	17.9	335000	1.4848	0.8505	0.8278
0.7899	18.17	340000	1.4919	0.8490	0.8244
0.7961	18.44	345000	1.4764	0.8522	0.8300
0.793	18.71	350000	1.4795	0.8522	0.8279
0.7933	18.97	355000	1.4845	0.8511	0.8257
0.788	19.24	360000	1.4814	0.8508	0.8272
0.787	19.51	365000	1.4739	0.8524	0.8282
0.7911	19.77	370000	1.4788	0.8514	0.8305
0.7843	20.04	375000	1.4762	0.8533	0.8285
0.7859	20.31	380000	1.4777	0.8525	0.8299
0.7846	20.58	385000	1.4696	0.8524	0.8309
0.7857	20.84	390000	1.4812	0.8519	0.8267
0.7861	21.11	395000	1.4924	0.8491	0.8241
0.7845	21.38	400000	1.4847	0.8511	0.8253
0.7847	21.65	405000	1.4748	0.8527	0.8277
0.785	21.91	410000	1.4794	0.8530	0.8281
0.7816	22.18	415000	1.4764	0.8529	0.8268
0.7845	22.45	420000	1.4738	0.8529	0.8288
0.7807	22.71	425000	1.4864	0.8511	0.8278
0.7809	22.98	430000	1.4782	0.8509	0.8259
0.7761	23.25	435000	1.4829	0.8527	0.8281
0.7815	23.52	440000	1.4627	0.8541	0.8298
0.7833	23.78	445000	1.4756	0.8532	0.8293
0.7786	24.05	450000	1.4612	0.8556	0.8318
0.7826	24.32	455000	1.4582	0.8553	0.8304
0.7821	24.58	460000	1.4614	0.8562	0.8323
0.7793	24.85	465000	1.4600	0.8564	0.8319
0.7763	25.12	470000	1.4650	0.8549	0.8311
0.776	25.39	475000	1.4658	0.8553	0.8285
0.7796	25.65	480000	1.4572	0.8560	0.8308
0.7794	25.92	485000	1.4638	0.8545	0.8302
0.7747	26.19	490000	1.4541	0.8565	0.8329
0.7761	26.46	495000	1.4596	0.8553	0.8283
0.7754	26.72	500000	1.4622	0.8566	0.8325
0.7769	26.99	505000	1.4665	0.8544	0.8303
0.777	27.26	510000	1.4622	0.8557	0.8300
0.7741	27.52	515000	1.4599	0.8562	0.8321
0.7747	27.79	520000	1.4575	0.8558	0.8301
0.7764	28.06	525000	1.4513	0.8565	0.8310
0.7731	28.33	530000	1.4551	0.8562	0.8322
0.7752	28.59	535000	1.4530	0.8568	0.8322
0.7729	28.86	540000	1.4563	0.8562	0.8323
0.7751	29.13	545000	1.4548	0.8567	0.8332
0.7742	29.39	550000	1.4494	0.8569	0.8333
0.7725	29.66	555000	1.4496	0.8576	0.8342
0.7756	29.93	560000	1.4509	0.8576	0.8345

Framework versions

• Transformers 4.33.3
• Pytorch 2.1.1+cu121
• Datasets 2.14.5
• Tokenizers 0.13.3