Upload model files
Browse files- config.json +264 -0
- merges.txt +0 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +15 -0
- tokenizer.json +0 -0
- tokenizer_config.json +15 -0
- trainer_state.json +234 -0
- training_args.bin +3 -0
- vocab.json +0 -0
config.json
ADDED
@@ -0,0 +1,264 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "roberta-base",
|
3 |
+
"architectures": [
|
4 |
+
"RobertaForSequenceClassification"
|
5 |
+
],
|
6 |
+
"attention_probs_dropout_prob": 0.1,
|
7 |
+
"bos_token_id": 0,
|
8 |
+
"classifier_dropout": null,
|
9 |
+
"eos_token_id": 2,
|
10 |
+
"hidden_act": "gelu",
|
11 |
+
"hidden_dropout_prob": 0.1,
|
12 |
+
"hidden_size": 768,
|
13 |
+
"id2label": {
|
14 |
+
"0": "acceptagreementcontractceasefire",
|
15 |
+
"1": "accidentcrash",
|
16 |
+
"2": "agreements",
|
17 |
+
"3": "airstrikemissilestrike",
|
18 |
+
"4": "arrestjaildetain",
|
19 |
+
"5": "artifact",
|
20 |
+
"6": "artifactexistence",
|
21 |
+
"7": "attack",
|
22 |
+
"8": "biologicalchemicalpoisonattack",
|
23 |
+
"9": "bombing",
|
24 |
+
"10": "borrowlend",
|
25 |
+
"11": "bringcarryunload",
|
26 |
+
"12": "broadcast",
|
27 |
+
"13": "build",
|
28 |
+
"14": "castvote",
|
29 |
+
"15": "chargeindict",
|
30 |
+
"16": "collaborate",
|
31 |
+
"17": "commandorder",
|
32 |
+
"18": "commitmentpromiseexpressintent",
|
33 |
+
"19": "conflict",
|
34 |
+
"20": "contact",
|
35 |
+
"21": "convict",
|
36 |
+
"22": "correspondence",
|
37 |
+
"23": "createintellectualproperty",
|
38 |
+
"24": "createmanufacture",
|
39 |
+
"25": "damage",
|
40 |
+
"26": "damagedestroy",
|
41 |
+
"27": "deathcausedbyviolentevents",
|
42 |
+
"28": "demonstrate",
|
43 |
+
"29": "destroy",
|
44 |
+
"30": "die",
|
45 |
+
"31": "disaster",
|
46 |
+
"32": "discussion",
|
47 |
+
"33": "disperseseparate",
|
48 |
+
"34": "elect",
|
49 |
+
"35": "embargosanction",
|
50 |
+
"36": "endposition",
|
51 |
+
"37": "evacuationrescue",
|
52 |
+
"38": "execute",
|
53 |
+
"39": "extradite",
|
54 |
+
"40": "fall",
|
55 |
+
"41": "firearmattack",
|
56 |
+
"42": "fireexplosion",
|
57 |
+
"43": "firinglayoff",
|
58 |
+
"44": "formation",
|
59 |
+
"45": "funeralvigil",
|
60 |
+
"46": "giftgrantprovideaid",
|
61 |
+
"47": "government",
|
62 |
+
"48": "grantentry",
|
63 |
+
"49": "grantentryasylum",
|
64 |
+
"50": "hanging",
|
65 |
+
"51": "hide",
|
66 |
+
"52": "hiring",
|
67 |
+
"53": "illnessdegradationhungerthirst",
|
68 |
+
"54": "illnessdegradationphysical",
|
69 |
+
"55": "initiatejudicialprocess",
|
70 |
+
"56": "injure",
|
71 |
+
"57": "injurycausedbyviolentevents",
|
72 |
+
"58": "inspection",
|
73 |
+
"59": "inspectpeopleorganization",
|
74 |
+
"60": "invade",
|
75 |
+
"61": "investigate",
|
76 |
+
"62": "investigatecrime",
|
77 |
+
"63": "judicialconsequences",
|
78 |
+
"64": "justice",
|
79 |
+
"65": "legislate",
|
80 |
+
"66": "life",
|
81 |
+
"67": "manufacture",
|
82 |
+
"68": "marchprotestpoliticalgathering",
|
83 |
+
"69": "mediastatement",
|
84 |
+
"70": "meet",
|
85 |
+
"71": "mergegpe",
|
86 |
+
"72": "monitorelection",
|
87 |
+
"73": "movement",
|
88 |
+
"74": "negotiate",
|
89 |
+
"75": "nonviolentdeath",
|
90 |
+
"76": "nonviolentthrowlaunch",
|
91 |
+
"77": "payforservice",
|
92 |
+
"78": "personnel",
|
93 |
+
"79": "physicalinvestigateinspect",
|
94 |
+
"80": "prevarication",
|
95 |
+
"81": "prevententry",
|
96 |
+
"82": "preventexit",
|
97 |
+
"83": "publicstatementinperson",
|
98 |
+
"84": "purchase",
|
99 |
+
"85": "quitretire",
|
100 |
+
"86": "receiveimport",
|
101 |
+
"87": "rejectnullifyagreementcontractceasefire",
|
102 |
+
"88": "requestadvise",
|
103 |
+
"89": "retreat",
|
104 |
+
"90": "selfdirectedbattle",
|
105 |
+
"91": "selfmotion",
|
106 |
+
"92": "sendsupplyexport",
|
107 |
+
"93": "sensoryobserve",
|
108 |
+
"94": "setfire",
|
109 |
+
"95": "smuggleextract",
|
110 |
+
"96": "spy",
|
111 |
+
"97": "stabbing",
|
112 |
+
"98": "startgpe",
|
113 |
+
"99": "startposition",
|
114 |
+
"100": "stealrobhijack",
|
115 |
+
"101": "strangling",
|
116 |
+
"102": "surrender",
|
117 |
+
"103": "threatencoerce",
|
118 |
+
"104": "transaction",
|
119 |
+
"105": "transfercontrol",
|
120 |
+
"106": "transfermoney",
|
121 |
+
"107": "transferownership",
|
122 |
+
"108": "transportartifact",
|
123 |
+
"109": "transportperson",
|
124 |
+
"110": "trialhearing",
|
125 |
+
"111": "violateagreement",
|
126 |
+
"112": "violationspreventvote",
|
127 |
+
"113": "vote",
|
128 |
+
"114": "winelection",
|
129 |
+
"115": "yield"
|
130 |
+
},
|
131 |
+
"initializer_range": 0.02,
|
132 |
+
"intermediate_size": 3072,
|
133 |
+
"label2id": {
|
134 |
+
"acceptagreementcontractceasefire": 0,
|
135 |
+
"accidentcrash": 1,
|
136 |
+
"agreements": 2,
|
137 |
+
"airstrikemissilestrike": 3,
|
138 |
+
"arrestjaildetain": 4,
|
139 |
+
"artifact": 5,
|
140 |
+
"artifactexistence": 6,
|
141 |
+
"attack": 7,
|
142 |
+
"biologicalchemicalpoisonattack": 8,
|
143 |
+
"bombing": 9,
|
144 |
+
"borrowlend": 10,
|
145 |
+
"bringcarryunload": 11,
|
146 |
+
"broadcast": 12,
|
147 |
+
"build": 13,
|
148 |
+
"castvote": 14,
|
149 |
+
"chargeindict": 15,
|
150 |
+
"collaborate": 16,
|
151 |
+
"commandorder": 17,
|
152 |
+
"commitmentpromiseexpressintent": 18,
|
153 |
+
"conflict": 19,
|
154 |
+
"contact": 20,
|
155 |
+
"convict": 21,
|
156 |
+
"correspondence": 22,
|
157 |
+
"createintellectualproperty": 23,
|
158 |
+
"createmanufacture": 24,
|
159 |
+
"damage": 25,
|
160 |
+
"damagedestroy": 26,
|
161 |
+
"deathcausedbyviolentevents": 27,
|
162 |
+
"demonstrate": 28,
|
163 |
+
"destroy": 29,
|
164 |
+
"die": 30,
|
165 |
+
"disaster": 31,
|
166 |
+
"discussion": 32,
|
167 |
+
"disperseseparate": 33,
|
168 |
+
"elect": 34,
|
169 |
+
"embargosanction": 35,
|
170 |
+
"endposition": 36,
|
171 |
+
"evacuationrescue": 37,
|
172 |
+
"execute": 38,
|
173 |
+
"extradite": 39,
|
174 |
+
"fall": 40,
|
175 |
+
"firearmattack": 41,
|
176 |
+
"fireexplosion": 42,
|
177 |
+
"firinglayoff": 43,
|
178 |
+
"formation": 44,
|
179 |
+
"funeralvigil": 45,
|
180 |
+
"giftgrantprovideaid": 46,
|
181 |
+
"government": 47,
|
182 |
+
"grantentry": 48,
|
183 |
+
"grantentryasylum": 49,
|
184 |
+
"hanging": 50,
|
185 |
+
"hide": 51,
|
186 |
+
"hiring": 52,
|
187 |
+
"illnessdegradationhungerthirst": 53,
|
188 |
+
"illnessdegradationphysical": 54,
|
189 |
+
"initiatejudicialprocess": 55,
|
190 |
+
"injure": 56,
|
191 |
+
"injurycausedbyviolentevents": 57,
|
192 |
+
"inspection": 58,
|
193 |
+
"inspectpeopleorganization": 59,
|
194 |
+
"invade": 60,
|
195 |
+
"investigate": 61,
|
196 |
+
"investigatecrime": 62,
|
197 |
+
"judicialconsequences": 63,
|
198 |
+
"justice": 64,
|
199 |
+
"legislate": 65,
|
200 |
+
"life": 66,
|
201 |
+
"manufacture": 67,
|
202 |
+
"marchprotestpoliticalgathering": 68,
|
203 |
+
"mediastatement": 69,
|
204 |
+
"meet": 70,
|
205 |
+
"mergegpe": 71,
|
206 |
+
"monitorelection": 72,
|
207 |
+
"movement": 73,
|
208 |
+
"negotiate": 74,
|
209 |
+
"nonviolentdeath": 75,
|
210 |
+
"nonviolentthrowlaunch": 76,
|
211 |
+
"payforservice": 77,
|
212 |
+
"personnel": 78,
|
213 |
+
"physicalinvestigateinspect": 79,
|
214 |
+
"prevarication": 80,
|
215 |
+
"prevententry": 81,
|
216 |
+
"preventexit": 82,
|
217 |
+
"publicstatementinperson": 83,
|
218 |
+
"purchase": 84,
|
219 |
+
"quitretire": 85,
|
220 |
+
"receiveimport": 86,
|
221 |
+
"rejectnullifyagreementcontractceasefire": 87,
|
222 |
+
"requestadvise": 88,
|
223 |
+
"retreat": 89,
|
224 |
+
"selfdirectedbattle": 90,
|
225 |
+
"selfmotion": 91,
|
226 |
+
"sendsupplyexport": 92,
|
227 |
+
"sensoryobserve": 93,
|
228 |
+
"setfire": 94,
|
229 |
+
"smuggleextract": 95,
|
230 |
+
"spy": 96,
|
231 |
+
"stabbing": 97,
|
232 |
+
"startgpe": 98,
|
233 |
+
"startposition": 99,
|
234 |
+
"stealrobhijack": 100,
|
235 |
+
"strangling": 101,
|
236 |
+
"surrender": 102,
|
237 |
+
"threatencoerce": 103,
|
238 |
+
"transaction": 104,
|
239 |
+
"transfercontrol": 105,
|
240 |
+
"transfermoney": 106,
|
241 |
+
"transferownership": 107,
|
242 |
+
"transportartifact": 108,
|
243 |
+
"transportperson": 109,
|
244 |
+
"trialhearing": 110,
|
245 |
+
"violateagreement": 111,
|
246 |
+
"violationspreventvote": 112,
|
247 |
+
"vote": 113,
|
248 |
+
"winelection": 114,
|
249 |
+
"yield": 115
|
250 |
+
},
|
251 |
+
"layer_norm_eps": 1e-05,
|
252 |
+
"max_position_embeddings": 514,
|
253 |
+
"model_type": "roberta",
|
254 |
+
"num_attention_heads": 12,
|
255 |
+
"num_hidden_layers": 12,
|
256 |
+
"pad_token_id": 1,
|
257 |
+
"position_embedding_type": "absolute",
|
258 |
+
"problem_type": "multi_label_classification",
|
259 |
+
"torch_dtype": "float32",
|
260 |
+
"transformers_version": "4.30.0",
|
261 |
+
"type_vocab_size": 1,
|
262 |
+
"use_cache": true,
|
263 |
+
"vocab_size": 50265
|
264 |
+
}
|
merges.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f581d2baf77a71ebf12a5010656f02dd5bb17fde4a1df174299a1519171a6782
|
3 |
+
size 685989061
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b23d8d05114cd040f4b2e354f6cf088e22869aa882b6ec70485d648ed9f4a7bd
|
3 |
+
size 499012725
|
rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5222e75d734bccdd7c107bea8376baa60ac7774c6f42faf740863489eab72287
|
3 |
+
size 14575
|
scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e73a459b3df7488e4cc14fe1f2f5d67538056e0c66f5fa32c217682302d83949
|
3 |
+
size 627
|
special_tokens_map.json
ADDED
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": "<s>",
|
3 |
+
"cls_token": "<s>",
|
4 |
+
"eos_token": "</s>",
|
5 |
+
"mask_token": {
|
6 |
+
"content": "<mask>",
|
7 |
+
"lstrip": true,
|
8 |
+
"normalized": false,
|
9 |
+
"rstrip": false,
|
10 |
+
"single_word": false
|
11 |
+
},
|
12 |
+
"pad_token": "<pad>",
|
13 |
+
"sep_token": "</s>",
|
14 |
+
"unk_token": "<unk>"
|
15 |
+
}
|
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
ADDED
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_prefix_space": false,
|
3 |
+
"bos_token": "<s>",
|
4 |
+
"clean_up_tokenization_spaces": true,
|
5 |
+
"cls_token": "<s>",
|
6 |
+
"eos_token": "</s>",
|
7 |
+
"errors": "replace",
|
8 |
+
"mask_token": "<mask>",
|
9 |
+
"model_max_length": 512,
|
10 |
+
"pad_token": "<pad>",
|
11 |
+
"sep_token": "</s>",
|
12 |
+
"tokenizer_class": "RobertaTokenizer",
|
13 |
+
"trim_offsets": true,
|
14 |
+
"unk_token": "<unk>"
|
15 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,234 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.8082437275985663,
|
3 |
+
"best_model_checkpoint": "roberta_finetuned_hier/checkpoint-9170",
|
4 |
+
"epoch": 10.0,
|
5 |
+
"global_step": 9170,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 0.55,
|
12 |
+
"learning_rate": 1.8909487459105783e-05,
|
13 |
+
"loss": 0.1685,
|
14 |
+
"step": 500
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"epoch": 1.0,
|
18 |
+
"eval_accuracy": 0.0,
|
19 |
+
"eval_f1": 0.0,
|
20 |
+
"eval_loss": 0.0928020253777504,
|
21 |
+
"eval_roc_auc": 0.5,
|
22 |
+
"eval_runtime": 17.532,
|
23 |
+
"eval_samples_per_second": 52.704,
|
24 |
+
"eval_steps_per_second": 6.616,
|
25 |
+
"step": 917
|
26 |
+
},
|
27 |
+
{
|
28 |
+
"epoch": 1.09,
|
29 |
+
"learning_rate": 1.781897491821156e-05,
|
30 |
+
"loss": 0.094,
|
31 |
+
"step": 1000
|
32 |
+
},
|
33 |
+
{
|
34 |
+
"epoch": 1.64,
|
35 |
+
"learning_rate": 1.6728462377317342e-05,
|
36 |
+
"loss": 0.093,
|
37 |
+
"step": 1500
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 2.0,
|
41 |
+
"eval_accuracy": 0.0021645021645021645,
|
42 |
+
"eval_f1": 0.2291442892946652,
|
43 |
+
"eval_loss": 0.07939616590738297,
|
44 |
+
"eval_roc_auc": 0.5650002587190003,
|
45 |
+
"eval_runtime": 17.4376,
|
46 |
+
"eval_samples_per_second": 52.989,
|
47 |
+
"eval_steps_per_second": 6.652,
|
48 |
+
"step": 1834
|
49 |
+
},
|
50 |
+
{
|
51 |
+
"epoch": 2.18,
|
52 |
+
"learning_rate": 1.563794983642312e-05,
|
53 |
+
"loss": 0.0845,
|
54 |
+
"step": 2000
|
55 |
+
},
|
56 |
+
{
|
57 |
+
"epoch": 2.73,
|
58 |
+
"learning_rate": 1.45474372955289e-05,
|
59 |
+
"loss": 0.0637,
|
60 |
+
"step": 2500
|
61 |
+
},
|
62 |
+
{
|
63 |
+
"epoch": 3.0,
|
64 |
+
"eval_accuracy": 0.10064935064935066,
|
65 |
+
"eval_f1": 0.6067974578612877,
|
66 |
+
"eval_loss": 0.048662275075912476,
|
67 |
+
"eval_roc_auc": 0.7229654750415772,
|
68 |
+
"eval_runtime": 17.4089,
|
69 |
+
"eval_samples_per_second": 53.076,
|
70 |
+
"eval_steps_per_second": 6.663,
|
71 |
+
"step": 2751
|
72 |
+
},
|
73 |
+
{
|
74 |
+
"epoch": 3.27,
|
75 |
+
"learning_rate": 1.3456924754634679e-05,
|
76 |
+
"loss": 0.0511,
|
77 |
+
"step": 3000
|
78 |
+
},
|
79 |
+
{
|
80 |
+
"epoch": 3.82,
|
81 |
+
"learning_rate": 1.236641221374046e-05,
|
82 |
+
"loss": 0.0429,
|
83 |
+
"step": 3500
|
84 |
+
},
|
85 |
+
{
|
86 |
+
"epoch": 4.0,
|
87 |
+
"eval_accuracy": 0.16017316017316016,
|
88 |
+
"eval_f1": 0.7002300025555839,
|
89 |
+
"eval_loss": 0.03621600940823555,
|
90 |
+
"eval_roc_auc": 0.7781674734201419,
|
91 |
+
"eval_runtime": 17.2108,
|
92 |
+
"eval_samples_per_second": 53.687,
|
93 |
+
"eval_steps_per_second": 6.74,
|
94 |
+
"step": 3668
|
95 |
+
},
|
96 |
+
{
|
97 |
+
"epoch": 4.36,
|
98 |
+
"learning_rate": 1.1275899672846238e-05,
|
99 |
+
"loss": 0.0374,
|
100 |
+
"step": 4000
|
101 |
+
},
|
102 |
+
{
|
103 |
+
"epoch": 4.91,
|
104 |
+
"learning_rate": 1.0185387131952018e-05,
|
105 |
+
"loss": 0.0334,
|
106 |
+
"step": 4500
|
107 |
+
},
|
108 |
+
{
|
109 |
+
"epoch": 5.0,
|
110 |
+
"eval_accuracy": 0.2196969696969697,
|
111 |
+
"eval_f1": 0.7534612581977169,
|
112 |
+
"eval_loss": 0.029505521059036255,
|
113 |
+
"eval_roc_auc": 0.8148612406998286,
|
114 |
+
"eval_runtime": 17.208,
|
115 |
+
"eval_samples_per_second": 53.696,
|
116 |
+
"eval_steps_per_second": 6.741,
|
117 |
+
"step": 4585
|
118 |
+
},
|
119 |
+
{
|
120 |
+
"epoch": 5.45,
|
121 |
+
"learning_rate": 9.094874591057799e-06,
|
122 |
+
"loss": 0.0305,
|
123 |
+
"step": 5000
|
124 |
+
},
|
125 |
+
{
|
126 |
+
"epoch": 6.0,
|
127 |
+
"learning_rate": 8.004362050163578e-06,
|
128 |
+
"loss": 0.0282,
|
129 |
+
"step": 5500
|
130 |
+
},
|
131 |
+
{
|
132 |
+
"epoch": 6.0,
|
133 |
+
"eval_accuracy": 0.23376623376623376,
|
134 |
+
"eval_f1": 0.7756258856872934,
|
135 |
+
"eval_loss": 0.02626137062907219,
|
136 |
+
"eval_roc_auc": 0.8332405627512013,
|
137 |
+
"eval_runtime": 17.2218,
|
138 |
+
"eval_samples_per_second": 53.653,
|
139 |
+
"eval_steps_per_second": 6.736,
|
140 |
+
"step": 5502
|
141 |
+
},
|
142 |
+
{
|
143 |
+
"epoch": 6.54,
|
144 |
+
"learning_rate": 6.913849509269357e-06,
|
145 |
+
"loss": 0.0267,
|
146 |
+
"step": 6000
|
147 |
+
},
|
148 |
+
{
|
149 |
+
"epoch": 7.0,
|
150 |
+
"eval_accuracy": 0.2694805194805195,
|
151 |
+
"eval_f1": 0.7864801864801864,
|
152 |
+
"eval_loss": 0.023826098069548607,
|
153 |
+
"eval_roc_auc": 0.8423381052507725,
|
154 |
+
"eval_runtime": 17.4138,
|
155 |
+
"eval_samples_per_second": 53.061,
|
156 |
+
"eval_steps_per_second": 6.661,
|
157 |
+
"step": 6419
|
158 |
+
},
|
159 |
+
{
|
160 |
+
"epoch": 7.09,
|
161 |
+
"learning_rate": 5.823336968375137e-06,
|
162 |
+
"loss": 0.0252,
|
163 |
+
"step": 6500
|
164 |
+
},
|
165 |
+
{
|
166 |
+
"epoch": 7.63,
|
167 |
+
"learning_rate": 4.732824427480917e-06,
|
168 |
+
"loss": 0.0241,
|
169 |
+
"step": 7000
|
170 |
+
},
|
171 |
+
{
|
172 |
+
"epoch": 8.0,
|
173 |
+
"eval_accuracy": 0.2922077922077922,
|
174 |
+
"eval_f1": 0.7973497829563628,
|
175 |
+
"eval_loss": 0.02243439108133316,
|
176 |
+
"eval_roc_auc": 0.8539930593410131,
|
177 |
+
"eval_runtime": 17.2304,
|
178 |
+
"eval_samples_per_second": 53.626,
|
179 |
+
"eval_steps_per_second": 6.732,
|
180 |
+
"step": 7336
|
181 |
+
},
|
182 |
+
{
|
183 |
+
"epoch": 8.18,
|
184 |
+
"learning_rate": 3.6423118865866965e-06,
|
185 |
+
"loss": 0.0235,
|
186 |
+
"step": 7500
|
187 |
+
},
|
188 |
+
{
|
189 |
+
"epoch": 8.72,
|
190 |
+
"learning_rate": 2.5517993456924756e-06,
|
191 |
+
"loss": 0.0224,
|
192 |
+
"step": 8000
|
193 |
+
},
|
194 |
+
{
|
195 |
+
"epoch": 9.0,
|
196 |
+
"eval_accuracy": 0.28679653679653677,
|
197 |
+
"eval_f1": 0.7981900452488687,
|
198 |
+
"eval_loss": 0.021712390705943108,
|
199 |
+
"eval_roc_auc": 0.8577418326100689,
|
200 |
+
"eval_runtime": 18.0929,
|
201 |
+
"eval_samples_per_second": 51.07,
|
202 |
+
"eval_steps_per_second": 6.411,
|
203 |
+
"step": 8253
|
204 |
+
},
|
205 |
+
{
|
206 |
+
"epoch": 9.27,
|
207 |
+
"learning_rate": 1.4612868047982554e-06,
|
208 |
+
"loss": 0.0223,
|
209 |
+
"step": 8500
|
210 |
+
},
|
211 |
+
{
|
212 |
+
"epoch": 9.81,
|
213 |
+
"learning_rate": 3.7077426390403497e-07,
|
214 |
+
"loss": 0.0218,
|
215 |
+
"step": 9000
|
216 |
+
},
|
217 |
+
{
|
218 |
+
"epoch": 10.0,
|
219 |
+
"eval_accuracy": 0.3106060606060606,
|
220 |
+
"eval_f1": 0.8082437275985663,
|
221 |
+
"eval_loss": 0.021442167460918427,
|
222 |
+
"eval_roc_auc": 0.8658561225293185,
|
223 |
+
"eval_runtime": 17.28,
|
224 |
+
"eval_samples_per_second": 53.472,
|
225 |
+
"eval_steps_per_second": 6.713,
|
226 |
+
"step": 9170
|
227 |
+
}
|
228 |
+
],
|
229 |
+
"max_steps": 9170,
|
230 |
+
"num_train_epochs": 10,
|
231 |
+
"total_flos": 1.930314697371648e+16,
|
232 |
+
"trial_name": null,
|
233 |
+
"trial_params": null
|
234 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b067917048e0f73b7e123db97e5b3b3f3cf9ce46b7963373d25007ee56bc1a94
|
3 |
+
size 3899
|
vocab.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|