Upload 13 files

Browse files

Files changed (13) hide show

all_results.json +14 -0
config.json +38 -0
eval_results.json +9 -0
predict_results_valid.txt +1120 -0
pytorch_model.bin +3 -0
special_tokens_map.json +9 -0
spm.model +3 -0
tokenizer.json +0 -0
tokenizer.py +26 -0
tokenizer_config.json +23 -0
train_results.json +8 -0
trainer_state.json +76 -0
training_args.bin +3 -0

all_results.json ADDED Viewed

	@@ -0,0 +1,14 @@

+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.8650580644607544,
+    "eval_loss": 0.5917113423347473,
+    "eval_runtime": 12.8425,
+    "eval_samples": 1119,
+    "eval_samples_per_second": 87.132,
+    "eval_steps_per_second": 10.901,
+    "train_loss": 0.3305379035488871,
+    "train_runtime": 996.8011,
+    "train_samples": 8939,
+    "train_samples_per_second": 26.903,
+    "train_steps_per_second": 2.242
+}

config.json ADDED Viewed

	@@ -0,0 +1,38 @@

+{
+  "_name_or_path": "deberta-v2-base-japanese",
+  "architectures": [
+    "DebertaV2ForMultipleChoice"
+  ],
+  "attention_head_size": 64,
+  "attention_probs_dropout_prob": 0.1,
+  "conv_act": "gelu",
+  "conv_kernel_size": 3,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "layer_norm_eps": 1e-07,
+  "max_position_embeddings": 512,
+  "max_relative_positions": -1,
+  "model_type": "deberta-v2",
+  "norm_rel_ebd": "layer_norm",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 0,
+  "pooler_dropout": 0,
+  "pooler_hidden_act": "gelu",
+  "pooler_hidden_size": 768,
+  "pos_att_type": [
+    "p2c",
+    "c2p"
+  ],
+  "position_biased_input": false,
+  "position_buckets": 256,
+  "relative_attention": true,
+  "share_att_key": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.26.0",
+  "type_vocab_size": 0,
+  "vocab_size": 32000
+}

eval_results.json ADDED Viewed

	@@ -0,0 +1,9 @@

+{
+    "epoch": 3.0,
+    "eval_accuracy": 0.8650580644607544,
+    "eval_loss": 0.5917113423347473,
+    "eval_runtime": 12.8425,
+    "eval_samples": 1119,
+    "eval_samples_per_second": 87.132,
+    "eval_steps_per_second": 10.901
+}

predict_results_valid.txt ADDED Viewed

	@@ -0,0 +1,1120 @@

+index	prediction
+0	2
+1	2
+2	0
+3	0
+4	0
+5	4
+6	4
+7	0
+8	2
+9	4
+10	4
+11	0
+12	0
+13	0
+14	4
+15	2
+16	0
+17	0
+18	1
+19	2
+20	2
+21	4
+22	0
+23	4
+24	3
+25	0
+26	1
+27	3
+28	3
+29	3
+30	4
+31	0
+32	3
+33	2
+34	0
+35	3
+36	3
+37	1
+38	0
+39	2
+40	4
+41	0
+42	0
+43	1
+44	3
+45	0
+46	4
+47	1
+48	0
+49	3
+50	0
+51	0
+52	2
+53	3
+54	2
+55	4
+56	4
+57	1
+58	1
+59	1
+60	0
+61	0
+62	3
+63	1
+64	0
+65	1
+66	4
+67	0
+68	3
+69	2
+70	1
+71	2
+72	1
+73	4
+74	1
+75	3
+76	2
+77	1
+78	2
+79	2
+80	4
+81	2
+82	3
+83	0
+84	2
+85	2
+86	0
+87	3
+88	1
+89	3
+90	1
+91	0
+92	4
+93	3
+94	3
+95	4
+96	4
+97	2
+98	0
+99	1
+100	1
+101	2
+102	3
+103	0
+104	3
+105	2
+106	1
+107	4
+108	0
+109	3
+110	3
+111	3
+112	3
+113	0
+114	2
+115	0
+116	0
+117	4
+118	0
+119	2
+120	2
+121	2
+122	2
+123	1
+124	3
+125	0
+126	4
+127	2
+128	3
+129	3
+130	1
+131	1
+132	4
+133	1
+134	1
+135	0
+136	2
+137	1
+138	3
+139	1
+140	3
+141	4
+142	3
+143	4
+144	2
+145	1
+146	1
+147	2
+148	3
+149	0
+150	4
+151	1
+152	3
+153	0
+154	2
+155	4
+156	3
+157	1
+158	4
+159	3
+160	0
+161	3
+162	4
+163	2
+164	2
+165	2
+166	1
+167	0
+168	3
+169	0
+170	0
+171	0
+172	2
+173	3
+174	3
+175	3
+176	0
+177	1
+178	0
+179	0
+180	0
+181	2
+182	2
+183	0
+184	2
+185	3
+186	4
+187	4
+188	1
+189	2
+190	4
+191	0
+192	3
+193	3
+194	1
+195	3
+196	0
+197	1
+198	0
+199	4
+200	2
+201	3
+202	0
+203	1
+204	1
+205	2
+206	4
+207	3
+208	1
+209	2
+210	3
+211	3
+212	2
+213	3
+214	2
+215	1
+216	1
+217	2
+218	2
+219	4
+220	4
+221	3
+222	1
+223	2
+224	3
+225	1
+226	4
+227	2
+228	4
+229	2
+230	0
+231	2
+232	0
+233	2
+234	4
+235	1
+236	0
+237	0
+238	0
+239	4
+240	4
+241	1
+242	3
+243	3
+244	1
+245	0
+246	4
+247	3
+248	2
+249	3
+250	2
+251	0
+252	4
+253	0
+254	2
+255	3
+256	3
+257	2
+258	2
+259	1
+260	2
+261	2
+262	2
+263	4
+264	3
+265	3
+266	3
+267	3
+268	0
+269	2
+270	1
+271	0
+272	0
+273	3
+274	3
+275	3
+276	4
+277	4
+278	0
+279	0
+280	0
+281	3
+282	4
+283	0
+284	0
+285	0
+286	0
+287	2
+288	1
+289	1
+290	2
+291	3
+292	4
+293	4
+294	4
+295	0
+296	0
+297	1
+298	0
+299	2
+300	4
+301	0
+302	1
+303	1
+304	3
+305	0
+306	0
+307	4
+308	2
+309	3
+310	2
+311	4
+312	4
+313	3
+314	4
+315	4
+316	1
+317	0
+318	2
+319	2
+320	1
+321	2
+322	2
+323	2
+324	1
+325	3
+326	3
+327	4
+328	0
+329	2
+330	2
+331	1
+332	1
+333	0
+334	2
+335	4
+336	0
+337	2
+338	1
+339	4
+340	1
+341	1
+342	2
+343	0
+344	1
+345	3
+346	0
+347	3
+348	2
+349	4
+350	1
+351	4
+352	4
+353	1
+354	0
+355	4
+356	3
+357	1
+358	4
+359	1
+360	2
+361	3
+362	3
+363	2
+364	1
+365	1
+366	4
+367	3
+368	3
+369	0
+370	3
+371	4
+372	0
+373	0
+374	0
+375	3
+376	0
+377	2
+378	4
+379	3
+380	2
+381	1
+382	1
+383	1
+384	2
+385	0
+386	4
+387	1
+388	0
+389	0
+390	2
+391	4
+392	3
+393	1
+394	0
+395	2
+396	1
+397	2
+398	1
+399	4
+400	3
+401	2
+402	1
+403	4
+404	4
+405	0
+406	0
+407	1
+408	2
+409	3
+410	1
+411	3
+412	3
+413	4
+414	0
+415	1
+416	2
+417	2
+418	4
+419	2
+420	3
+421	3
+422	0
+423	4
+424	2
+425	1
+426	2
+427	1
+428	3
+429	4
+430	1
+431	4
+432	2
+433	3
+434	2
+435	0
+436	3
+437	2
+438	3
+439	4
+440	4
+441	1
+442	3
+443	3
+444	4
+445	0
+446	4
+447	0
+448	0
+449	1
+450	1
+451	0
+452	3
+453	4
+454	2
+455	3
+456	0
+457	3
+458	4
+459	3
+460	2
+461	3
+462	2
+463	3
+464	0
+465	2
+466	3
+467	1
+468	1
+469	1
+470	0
+471	3
+472	4
+473	0
+474	4
+475	4
+476	3
+477	4
+478	1
+479	2
+480	2
+481	0
+482	4
+483	1
+484	4
+485	2
+486	1
+487	3
+488	1
+489	2
+490	1
+491	2
+492	2
+493	4
+494	3
+495	3
+496	3
+497	4
+498	4
+499	0
+500	1
+501	1
+502	2
+503	4
+504	0
+505	1
+506	0
+507	3
+508	3
+509	0
+510	4
+511	0
+512	0
+513	3
+514	0
+515	2
+516	4
+517	1
+518	2
+519	2
+520	1
+521	1
+522	1
+523	1
+524	3
+525	0
+526	3
+527	0
+528	3
+529	4
+530	4
+531	4
+532	3
+533	0
+534	1
+535	4
+536	0
+537	2
+538	1
+539	1
+540	2
+541	2
+542	4
+543	0
+544	4
+545	4
+546	3
+547	2
+548	4
+549	0
+550	0
+551	0
+552	0
+553	2
+554	3
+555	2
+556	1
+557	2
+558	2
+559	3
+560	0
+561	2
+562	1
+563	2
+564	2
+565	4
+566	4
+567	1
+568	1
+569	2
+570	1
+571	1
+572	2
+573	3
+574	3
+575	1
+576	0
+577	4
+578	3
+579	3
+580	2
+581	2
+582	2
+583	4
+584	4
+585	0
+586	3
+587	1
+588	1
+589	0
+590	1
+591	4
+592	2
+593	4
+594	1
+595	4
+596	4
+597	4
+598	1
+599	3
+600	3
+601	0
+602	2
+603	3
+604	4
+605	3
+606	3
+607	2
+608	2
+609	1
+610	4
+611	0
+612	0
+613	0
+614	0
+615	2
+616	1
+617	4
+618	2
+619	0
+620	2
+621	2
+622	1
+623	2
+624	0
+625	1
+626	2
+627	0
+628	1
+629	0
+630	2
+631	3
+632	4
+633	0
+634	4
+635	1
+636	2
+637	0
+638	3
+639	1
+640	1
+641	2
+642	0
+643	3
+644	3
+645	4
+646	4
+647	2
+648	2
+649	2
+650	1
+651	3
+652	0
+653	4
+654	1
+655	3
+656	3
+657	1
+658	3
+659	4
+660	1
+661	0
+662	1
+663	2
+664	4
+665	1
+666	1
+667	4
+668	0
+669	0
+670	4
+671	1
+672	0
+673	2
+674	1
+675	1
+676	2
+677	3
+678	3
+679	4
+680	2
+681	3
+682	0
+683	0
+684	3
+685	0
+686	3
+687	3
+688	2
+689	2
+690	1
+691	2
+692	4
+693	0
+694	4
+695	1
+696	4
+697	4
+698	3
+699	4
+700	2
+701	3
+702	0
+703	0
+704	4
+705	3
+706	2
+707	3
+708	3
+709	4
+710	1
+711	3
+712	2
+713	1
+714	1
+715	2
+716	1
+717	2
+718	1
+719	0
+720	0
+721	1
+722	4
+723	3
+724	1
+725	4
+726	4
+727	2
+728	0
+729	3
+730	1
+731	3
+732	1
+733	4
+734	0
+735	1
+736	3
+737	2
+738	0
+739	2
+740	1
+741	1
+742	1
+743	3
+744	2
+745	2
+746	1
+747	3
+748	4
+749	0
+750	4
+751	1
+752	2
+753	4
+754	4
+755	4
+756	1
+757	0
+758	4
+759	4
+760	2
+761	4
+762	0
+763	4
+764	4
+765	0
+766	4
+767	0
+768	1
+769	1
+770	0
+771	3
+772	1
+773	0
+774	0
+775	2
+776	4
+777	0
+778	1
+779	3
+780	3
+781	1
+782	4
+783	3
+784	1
+785	4
+786	0
+787	3
+788	1
+789	2
+790	1
+791	0
+792	0
+793	2
+794	4
+795	1
+796	4
+797	1
+798	1
+799	3
+800	1
+801	4
+802	0
+803	1
+804	3
+805	4
+806	3
+807	0
+808	2
+809	1
+810	3
+811	2
+812	3
+813	3
+814	3
+815	1
+816	3
+817	1
+818	0
+819	3
+820	0
+821	1
+822	3
+823	2
+824	0
+825	0
+826	2
+827	2
+828	4
+829	1
+830	4
+831	3
+832	1
+833	1
+834	4
+835	3
+836	4
+837	3
+838	1
+839	3
+840	3
+841	1
+842	4
+843	3
+844	1
+845	0
+846	2
+847	1
+848	0
+849	1
+850	3
+851	2
+852	4
+853	2
+854	1
+855	4
+856	2
+857	4
+858	0
+859	0
+860	3
+861	1
+862	4
+863	2
+864	1
+865	4
+866	1
+867	1
+868	0
+869	4
+870	4
+871	3
+872	1
+873	3
+874	0
+875	4
+876	1
+877	1
+878	1
+879	0
+880	4
+881	2
+882	2
+883	2
+884	4
+885	1
+886	4
+887	1
+888	2
+889	2
+890	1
+891	4
+892	2
+893	1
+894	3
+895	4
+896	1
+897	0
+898	3
+899	2
+900	1
+901	0
+902	1
+903	3
+904	3
+905	0
+906	3
+907	0
+908	4
+909	4
+910	0
+911	1
+912	4
+913	2
+914	0
+915	2
+916	1
+917	0
+918	1
+919	4
+920	1
+921	4
+922	2
+923	3
+924	2
+925	3
+926	3
+927	4
+928	2
+929	2
+930	0
+931	0
+932	1
+933	4
+934	0
+935	0
+936	1
+937	0
+938	0
+939	4
+940	2
+941	3
+942	1
+943	3
+944	4
+945	3
+946	2
+947	1
+948	4
+949	3
+950	2
+951	3
+952	4
+953	1
+954	3
+955	3
+956	1
+957	2
+958	1
+959	0
+960	4
+961	0
+962	1
+963	4
+964	1
+965	3
+966	0
+967	4
+968	3
+969	4
+970	2
+971	4
+972	2
+973	4
+974	0
+975	1
+976	0
+977	4
+978	1
+979	1
+980	3
+981	3
+982	4
+983	0
+984	3
+985	3
+986	0
+987	1
+988	3
+989	0
+990	1
+991	3
+992	2
+993	3
+994	3
+995	3
+996	0
+997	4
+998	2
+999	4
+1000	2
+1001	1
+1002	2
+1003	0
+1004	2
+1005	3
+1006	1
+1007	1
+1008	2
+1009	4
+1010	3
+1011	4
+1012	4
+1013	0
+1014	3
+1015	0
+1016	0
+1017	0
+1018	1
+1019	0
+1020	4
+1021	2
+1022	0
+1023	3
+1024	1
+1025	1
+1026	3
+1027	0
+1028	4
+1029	3
+1030	1
+1031	3
+1032	1
+1033	4
+1034	3
+1035	4
+1036	4
+1037	0
+1038	0
+1039	3
+1040	3
+1041	0
+1042	2
+1043	2
+1044	2
+1045	2
+1046	2
+1047	2
+1048	2
+1049	0
+1050	2
+1051	4
+1052	4
+1053	1
+1054	0
+1055	2
+1056	1
+1057	2
+1058	0
+1059	1
+1060	4
+1061	2
+1062	3
+1063	3
+1064	0
+1065	4
+1066	2
+1067	3
+1068	1
+1069	2
+1070	2
+1071	2
+1072	4
+1073	0
+1074	1
+1075	2
+1076	0
+1077	3
+1078	2
+1079	4
+1080	0
+1081	1
+1082	1
+1083	2
+1084	1
+1085	3
+1086	1
+1087	1
+1088	3
+1089	1
+1090	0
+1091	4
+1092	2
+1093	1
+1094	0
+1095	0
+1096	2
+1097	0
+1098	3
+1099	2
+1100	0
+1101	1
+1102	2
+1103	2
+1104	1
+1105	4
+1106	4
+1107	2
+1108	3
+1109	4
+1110	0
+1111	1
+1112	1
+1113	2
+1114	3
+1115	3
+1116	2
+1117	1
+1118	2

pytorch_model.bin ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:05c600c56d5a92c2dd79a9c95971c92fad8b686e61e9a3b443485a5bf869dd40
+size 449634953

special_tokens_map.json ADDED Viewed

	@@ -0,0 +1,9 @@

+{
+  "bos_token": "[CLS]",
+  "cls_token": "[CLS]",
+  "eos_token": "[SEP]",
+  "mask_token": "[MASK]",
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "unk_token": "[UNK]"
+}

spm.model ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:6c111c16e2e52366dcac46b886e40650bb843fe2938a65f5970271fc5697a127
+size 805061

tokenizer.json ADDED Viewed

The diff for this file is too large to render. See raw diff

tokenizer.py ADDED Viewed

	@@ -0,0 +1,26 @@

+#! /usr/bin/python3
+from transformers import DebertaV2TokenizerFast
+from transformers.models.bert_japanese.tokenization_bert_japanese import JumanppTokenizer
+class JumanppPreTokenizer(JumanppTokenizer):
+  def jumanpp_split(self,i,normalized_string):
+    import textspan
+    t=str(normalized_string)
+    k=self.tokenize(t)
+    return [normalized_string[s:e] for c in textspan.get_original_spans(k,t) for s,e in c]
+  def pre_tokenize(self,pretok):
+    pretok.split(self.jumanpp_split)
+class JumanppDebertaV2TokenizerFast(DebertaV2TokenizerFast):
+  def __init__(self,**kwargs):
+    from tokenizers.pre_tokenizers import PreTokenizer,Metaspace,Sequence
+    super().__init__(**kwargs)
+    self._tokenizer.pre_tokenizer=Sequence([PreTokenizer.custom(JumanppPreTokenizer()),Metaspace()])
+  def save_pretrained(self,save_directory,**kwargs):
+    import os
+    import shutil
+    from tokenizers.pre_tokenizers import PreTokenizer,Metaspace,Sequence
+    self._auto_map={"AutoTokenizer":[None,"tokenizer.JumanppDebertaV2TokenizerFast"]}
+    self._tokenizer.pre_tokenizer=Metaspace()
+    super().save_pretrained(save_directory,**kwargs)
+    self._tokenizer.pre_tokenizer=Sequence([PreTokenizer.custom(JumanppPreTokenizer()),Metaspace()])
+    shutil.copy(os.path.abspath(__file__),os.path.join(save_directory,"tokenizer.py"))

tokenizer_config.json ADDED Viewed

	@@ -0,0 +1,23 @@

+{
+  "auto_map": {
+    "AutoTokenizer": [
+      "tokenization_deberta_v2.DebertaV2Tokenizer",
+      "tokenizer.JumanppDebertaV2TokenizerFast"
+    ]
+  },
+  "bos_token": "[CLS]",
+  "cls_token": "[CLS]",
+  "do_lower_case": false,
+  "eos_token": "[SEP]",
+  "keep_accents": true,
+  "mask_token": "[MASK]",
+  "model_max_length": 1000000000000000019884624838656,
+  "name_or_path": "deberta-v2-base-japanese",
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "sp_model_kwargs": {},
+  "special_tokens_map_file": null,
+  "split_by_punct": false,
+  "tokenizer_class": "JumanppDebertaV2Tokenizer",
+  "unk_token": "[UNK]"
+}

train_results.json ADDED Viewed

	@@ -0,0 +1,8 @@

+{
+    "epoch": 3.0,
+    "train_loss": 0.3305379035488871,
+    "train_runtime": 996.8011,
+    "train_samples": 8939,
+    "train_samples_per_second": 26.903,
+    "train_steps_per_second": 2.242
+}

trainer_state.json ADDED Viewed

	@@ -0,0 +1,76 @@

+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "global_step": 2235,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.67,
+      "learning_rate": 3.881431767337807e-05,
+      "loss": 0.699,
+      "step": 500
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.8391420841217041,
+      "eval_loss": 0.4257535934448242,
+      "eval_runtime": 12.6758,
+      "eval_samples_per_second": 88.278,
+      "eval_steps_per_second": 11.045,
+      "step": 745
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 2.7628635346756153e-05,
+      "loss": 0.4032,
+      "step": 1000
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.8570151925086975,
+      "eval_loss": 0.49252861738204956,
+      "eval_runtime": 12.6658,
+      "eval_samples_per_second": 88.348,
+      "eval_steps_per_second": 11.053,
+      "step": 1490
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.644295302013423e-05,
+      "loss": 0.2427,
+      "step": 1500
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 5.257270693512305e-06,
+      "loss": 0.0903,
+      "step": 2000
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.8650580644607544,
+      "eval_loss": 0.5917113423347473,
+      "eval_runtime": 12.628,
+      "eval_samples_per_second": 88.613,
+      "eval_steps_per_second": 11.086,
+      "step": 2235
+    },
+    {
+      "epoch": 3.0,
+      "step": 2235,
+      "total_flos": 1862975629346310.0,
+      "train_loss": 0.3305379035488871,
+      "train_runtime": 996.8011,
+      "train_samples_per_second": 26.903,
+      "train_steps_per_second": 2.242
+    }
+  ],
+  "max_steps": 2235,
+  "num_train_epochs": 3,
+  "total_flos": 1862975629346310.0,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:f5d9c05feb618e069ccf9aa1d453dabdec6c6ee6e634d834c0bb1df9062d40c9
+size 3579