Delete trainer_state.json
Browse files- trainer_state.json +0 -276
trainer_state.json
DELETED
@@ -1,276 +0,0 @@
|
|
1 |
-
{
|
2 |
-
"best_metric": null,
|
3 |
-
"best_model_checkpoint": null,
|
4 |
-
"epoch": 2.0,
|
5 |
-
"global_step": 19806,
|
6 |
-
"is_hyper_param_search": false,
|
7 |
-
"is_local_process_zero": true,
|
8 |
-
"is_world_process_zero": true,
|
9 |
-
"log_history": [
|
10 |
-
{
|
11 |
-
"epoch": 0.05,
|
12 |
-
"learning_rate": 4.87377562354842e-05,
|
13 |
-
"loss": 0.5877,
|
14 |
-
"step": 500
|
15 |
-
},
|
16 |
-
{
|
17 |
-
"epoch": 0.1,
|
18 |
-
"learning_rate": 4.74755124709684e-05,
|
19 |
-
"loss": 0.6317,
|
20 |
-
"step": 1000
|
21 |
-
},
|
22 |
-
{
|
23 |
-
"epoch": 0.15,
|
24 |
-
"learning_rate": 4.6213268706452596e-05,
|
25 |
-
"loss": 0.6456,
|
26 |
-
"step": 1500
|
27 |
-
},
|
28 |
-
{
|
29 |
-
"epoch": 0.2,
|
30 |
-
"learning_rate": 4.4951024941936785e-05,
|
31 |
-
"loss": 0.6473,
|
32 |
-
"step": 2000
|
33 |
-
},
|
34 |
-
{
|
35 |
-
"epoch": 0.25,
|
36 |
-
"learning_rate": 4.368878117742099e-05,
|
37 |
-
"loss": 0.6499,
|
38 |
-
"step": 2500
|
39 |
-
},
|
40 |
-
{
|
41 |
-
"epoch": 0.3,
|
42 |
-
"learning_rate": 4.242653741290518e-05,
|
43 |
-
"loss": 0.6602,
|
44 |
-
"step": 3000
|
45 |
-
},
|
46 |
-
{
|
47 |
-
"epoch": 0.35,
|
48 |
-
"learning_rate": 4.116429364838938e-05,
|
49 |
-
"loss": 0.6699,
|
50 |
-
"step": 3500
|
51 |
-
},
|
52 |
-
{
|
53 |
-
"epoch": 0.4,
|
54 |
-
"learning_rate": 3.9902049883873574e-05,
|
55 |
-
"loss": 0.6624,
|
56 |
-
"step": 4000
|
57 |
-
},
|
58 |
-
{
|
59 |
-
"epoch": 0.45,
|
60 |
-
"learning_rate": 3.8639806119357776e-05,
|
61 |
-
"loss": 0.651,
|
62 |
-
"step": 4500
|
63 |
-
},
|
64 |
-
{
|
65 |
-
"epoch": 0.5,
|
66 |
-
"learning_rate": 3.7377562354841965e-05,
|
67 |
-
"loss": 0.6709,
|
68 |
-
"step": 5000
|
69 |
-
},
|
70 |
-
{
|
71 |
-
"epoch": 0.56,
|
72 |
-
"learning_rate": 3.611531859032616e-05,
|
73 |
-
"loss": 0.6582,
|
74 |
-
"step": 5500
|
75 |
-
},
|
76 |
-
{
|
77 |
-
"epoch": 0.61,
|
78 |
-
"learning_rate": 3.485307482581036e-05,
|
79 |
-
"loss": 0.6698,
|
80 |
-
"step": 6000
|
81 |
-
},
|
82 |
-
{
|
83 |
-
"epoch": 0.66,
|
84 |
-
"learning_rate": 3.359083106129456e-05,
|
85 |
-
"loss": 0.6522,
|
86 |
-
"step": 6500
|
87 |
-
},
|
88 |
-
{
|
89 |
-
"epoch": 0.71,
|
90 |
-
"learning_rate": 3.2328587296778754e-05,
|
91 |
-
"loss": 0.6863,
|
92 |
-
"step": 7000
|
93 |
-
},
|
94 |
-
{
|
95 |
-
"epoch": 0.76,
|
96 |
-
"learning_rate": 3.1066343532262956e-05,
|
97 |
-
"loss": 0.6602,
|
98 |
-
"step": 7500
|
99 |
-
},
|
100 |
-
{
|
101 |
-
"epoch": 0.81,
|
102 |
-
"learning_rate": 2.9804099767747152e-05,
|
103 |
-
"loss": 0.6466,
|
104 |
-
"step": 8000
|
105 |
-
},
|
106 |
-
{
|
107 |
-
"epoch": 0.86,
|
108 |
-
"learning_rate": 2.8541856003231344e-05,
|
109 |
-
"loss": 0.7011,
|
110 |
-
"step": 8500
|
111 |
-
},
|
112 |
-
{
|
113 |
-
"epoch": 0.91,
|
114 |
-
"learning_rate": 2.727961223871554e-05,
|
115 |
-
"loss": 0.677,
|
116 |
-
"step": 9000
|
117 |
-
},
|
118 |
-
{
|
119 |
-
"epoch": 0.96,
|
120 |
-
"learning_rate": 2.601736847419974e-05,
|
121 |
-
"loss": 0.6581,
|
122 |
-
"step": 9500
|
123 |
-
},
|
124 |
-
{
|
125 |
-
"epoch": 1.0,
|
126 |
-
"eval_bleu": 30.9028,
|
127 |
-
"eval_chrf++": 57.3001,
|
128 |
-
"eval_gen_len": 27.1266,
|
129 |
-
"eval_loss": 1.247955083847046,
|
130 |
-
"eval_runtime": 1316.7682,
|
131 |
-
"eval_samples_per_second": 3.718,
|
132 |
-
"eval_spbleu": 43.0506,
|
133 |
-
"eval_steps_per_second": 0.744,
|
134 |
-
"eval_ter": 59.7925,
|
135 |
-
"step": 9903
|
136 |
-
},
|
137 |
-
{
|
138 |
-
"epoch": 1.01,
|
139 |
-
"learning_rate": 2.4755124709683934e-05,
|
140 |
-
"loss": 0.6144,
|
141 |
-
"step": 10000
|
142 |
-
},
|
143 |
-
{
|
144 |
-
"epoch": 1.06,
|
145 |
-
"learning_rate": 2.3492880945168133e-05,
|
146 |
-
"loss": 0.4286,
|
147 |
-
"step": 10500
|
148 |
-
},
|
149 |
-
{
|
150 |
-
"epoch": 1.11,
|
151 |
-
"learning_rate": 2.223063718065233e-05,
|
152 |
-
"loss": 0.4329,
|
153 |
-
"step": 11000
|
154 |
-
},
|
155 |
-
{
|
156 |
-
"epoch": 1.16,
|
157 |
-
"learning_rate": 2.0968393416136524e-05,
|
158 |
-
"loss": 0.4324,
|
159 |
-
"step": 11500
|
160 |
-
},
|
161 |
-
{
|
162 |
-
"epoch": 1.21,
|
163 |
-
"learning_rate": 1.9706149651620723e-05,
|
164 |
-
"loss": 0.4403,
|
165 |
-
"step": 12000
|
166 |
-
},
|
167 |
-
{
|
168 |
-
"epoch": 1.26,
|
169 |
-
"learning_rate": 1.844390588710492e-05,
|
170 |
-
"loss": 0.4353,
|
171 |
-
"step": 12500
|
172 |
-
},
|
173 |
-
{
|
174 |
-
"epoch": 1.31,
|
175 |
-
"learning_rate": 1.7181662122589115e-05,
|
176 |
-
"loss": 0.4312,
|
177 |
-
"step": 13000
|
178 |
-
},
|
179 |
-
{
|
180 |
-
"epoch": 1.36,
|
181 |
-
"learning_rate": 1.5919418358073314e-05,
|
182 |
-
"loss": 0.4212,
|
183 |
-
"step": 13500
|
184 |
-
},
|
185 |
-
{
|
186 |
-
"epoch": 1.41,
|
187 |
-
"learning_rate": 1.4657174593557507e-05,
|
188 |
-
"loss": 0.4158,
|
189 |
-
"step": 14000
|
190 |
-
},
|
191 |
-
{
|
192 |
-
"epoch": 1.46,
|
193 |
-
"learning_rate": 1.3394930829041705e-05,
|
194 |
-
"loss": 0.4061,
|
195 |
-
"step": 14500
|
196 |
-
},
|
197 |
-
{
|
198 |
-
"epoch": 1.51,
|
199 |
-
"learning_rate": 1.2132687064525902e-05,
|
200 |
-
"loss": 0.4202,
|
201 |
-
"step": 15000
|
202 |
-
},
|
203 |
-
{
|
204 |
-
"epoch": 1.57,
|
205 |
-
"learning_rate": 1.0870443300010098e-05,
|
206 |
-
"loss": 0.4448,
|
207 |
-
"step": 15500
|
208 |
-
},
|
209 |
-
{
|
210 |
-
"epoch": 1.62,
|
211 |
-
"learning_rate": 9.608199535494297e-06,
|
212 |
-
"loss": 0.4351,
|
213 |
-
"step": 16000
|
214 |
-
},
|
215 |
-
{
|
216 |
-
"epoch": 1.67,
|
217 |
-
"learning_rate": 8.345955770978492e-06,
|
218 |
-
"loss": 0.4072,
|
219 |
-
"step": 16500
|
220 |
-
},
|
221 |
-
{
|
222 |
-
"epoch": 1.72,
|
223 |
-
"learning_rate": 7.083712006462688e-06,
|
224 |
-
"loss": 0.4157,
|
225 |
-
"step": 17000
|
226 |
-
},
|
227 |
-
{
|
228 |
-
"epoch": 1.77,
|
229 |
-
"learning_rate": 5.821468241946885e-06,
|
230 |
-
"loss": 0.4338,
|
231 |
-
"step": 17500
|
232 |
-
},
|
233 |
-
{
|
234 |
-
"epoch": 1.82,
|
235 |
-
"learning_rate": 4.559224477431081e-06,
|
236 |
-
"loss": 0.4212,
|
237 |
-
"step": 18000
|
238 |
-
},
|
239 |
-
{
|
240 |
-
"epoch": 1.87,
|
241 |
-
"learning_rate": 3.2969807129152782e-06,
|
242 |
-
"loss": 0.411,
|
243 |
-
"step": 18500
|
244 |
-
},
|
245 |
-
{
|
246 |
-
"epoch": 1.92,
|
247 |
-
"learning_rate": 2.0347369483994747e-06,
|
248 |
-
"loss": 0.4179,
|
249 |
-
"step": 19000
|
250 |
-
},
|
251 |
-
{
|
252 |
-
"epoch": 1.97,
|
253 |
-
"learning_rate": 7.724931838836716e-07,
|
254 |
-
"loss": 0.4255,
|
255 |
-
"step": 19500
|
256 |
-
},
|
257 |
-
{
|
258 |
-
"epoch": 2.0,
|
259 |
-
"eval_bleu": 31.3408,
|
260 |
-
"eval_chrf++": 57.4539,
|
261 |
-
"eval_gen_len": 27.0633,
|
262 |
-
"eval_loss": 1.3426777124404907,
|
263 |
-
"eval_runtime": 1310.3387,
|
264 |
-
"eval_samples_per_second": 3.736,
|
265 |
-
"eval_spbleu": 43.1979,
|
266 |
-
"eval_steps_per_second": 0.748,
|
267 |
-
"eval_ter": 59.0383,
|
268 |
-
"step": 19806
|
269 |
-
}
|
270 |
-
],
|
271 |
-
"max_steps": 19806,
|
272 |
-
"num_train_epochs": 2,
|
273 |
-
"total_flos": 8.428883485392896e+16,
|
274 |
-
"trial_name": null,
|
275 |
-
"trial_params": null
|
276 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|