Commit
·
f89a45a
1
Parent(s):
46fe74c
Delete trainer_state.json
Browse files- trainer_state.json +0 -274
trainer_state.json
DELETED
@@ -1,274 +0,0 @@
|
|
1 |
-
{
|
2 |
-
"best_metric": null,
|
3 |
-
"best_model_checkpoint": null,
|
4 |
-
"epoch": 0.9868723033140548,
|
5 |
-
"global_step": 21500,
|
6 |
-
"is_hyper_param_search": false,
|
7 |
-
"is_local_process_zero": true,
|
8 |
-
"is_world_process_zero": true,
|
9 |
-
"log_history": [
|
10 |
-
{
|
11 |
-
"epoch": 0.02,
|
12 |
-
"learning_rate": 1.9540989626365558e-05,
|
13 |
-
"loss": 0.1022,
|
14 |
-
"step": 500
|
15 |
-
},
|
16 |
-
{
|
17 |
-
"epoch": 0.05,
|
18 |
-
"learning_rate": 1.908197925273111e-05,
|
19 |
-
"loss": 0.0998,
|
20 |
-
"step": 1000
|
21 |
-
},
|
22 |
-
{
|
23 |
-
"epoch": 0.07,
|
24 |
-
"learning_rate": 1.862296887909667e-05,
|
25 |
-
"loss": 0.0999,
|
26 |
-
"step": 1500
|
27 |
-
},
|
28 |
-
{
|
29 |
-
"epoch": 0.09,
|
30 |
-
"learning_rate": 1.8163958505462225e-05,
|
31 |
-
"loss": 0.0996,
|
32 |
-
"step": 2000
|
33 |
-
},
|
34 |
-
{
|
35 |
-
"epoch": 0.11,
|
36 |
-
"learning_rate": 1.770494813182778e-05,
|
37 |
-
"loss": 0.0982,
|
38 |
-
"step": 2500
|
39 |
-
},
|
40 |
-
{
|
41 |
-
"epoch": 0.14,
|
42 |
-
"learning_rate": 1.7245937758193335e-05,
|
43 |
-
"loss": 0.0991,
|
44 |
-
"step": 3000
|
45 |
-
},
|
46 |
-
{
|
47 |
-
"epoch": 0.16,
|
48 |
-
"learning_rate": 1.6786927384558895e-05,
|
49 |
-
"loss": 0.0984,
|
50 |
-
"step": 3500
|
51 |
-
},
|
52 |
-
{
|
53 |
-
"epoch": 0.18,
|
54 |
-
"learning_rate": 1.6327917010924448e-05,
|
55 |
-
"loss": 0.0965,
|
56 |
-
"step": 4000
|
57 |
-
},
|
58 |
-
{
|
59 |
-
"epoch": 0.21,
|
60 |
-
"learning_rate": 1.5868906637290005e-05,
|
61 |
-
"loss": 0.0979,
|
62 |
-
"step": 4500
|
63 |
-
},
|
64 |
-
{
|
65 |
-
"epoch": 0.23,
|
66 |
-
"learning_rate": 1.5409896263655558e-05,
|
67 |
-
"loss": 0.0965,
|
68 |
-
"step": 5000
|
69 |
-
},
|
70 |
-
{
|
71 |
-
"epoch": 0.25,
|
72 |
-
"learning_rate": 1.4950885890021116e-05,
|
73 |
-
"loss": 0.0959,
|
74 |
-
"step": 5500
|
75 |
-
},
|
76 |
-
{
|
77 |
-
"epoch": 0.28,
|
78 |
-
"learning_rate": 1.4491875516386671e-05,
|
79 |
-
"loss": 0.0966,
|
80 |
-
"step": 6000
|
81 |
-
},
|
82 |
-
{
|
83 |
-
"epoch": 0.3,
|
84 |
-
"learning_rate": 1.4032865142752228e-05,
|
85 |
-
"loss": 0.0985,
|
86 |
-
"step": 6500
|
87 |
-
},
|
88 |
-
{
|
89 |
-
"epoch": 0.32,
|
90 |
-
"learning_rate": 1.3573854769117783e-05,
|
91 |
-
"loss": 0.0951,
|
92 |
-
"step": 7000
|
93 |
-
},
|
94 |
-
{
|
95 |
-
"epoch": 0.34,
|
96 |
-
"learning_rate": 1.311484439548334e-05,
|
97 |
-
"loss": 0.0956,
|
98 |
-
"step": 7500
|
99 |
-
},
|
100 |
-
{
|
101 |
-
"epoch": 0.37,
|
102 |
-
"learning_rate": 1.2656752042596163e-05,
|
103 |
-
"loss": 0.0949,
|
104 |
-
"step": 8000
|
105 |
-
},
|
106 |
-
{
|
107 |
-
"epoch": 0.39,
|
108 |
-
"learning_rate": 1.219774166896172e-05,
|
109 |
-
"loss": 0.0954,
|
110 |
-
"step": 8500
|
111 |
-
},
|
112 |
-
{
|
113 |
-
"epoch": 0.41,
|
114 |
-
"learning_rate": 1.1738731295327275e-05,
|
115 |
-
"loss": 0.0949,
|
116 |
-
"step": 9000
|
117 |
-
},
|
118 |
-
{
|
119 |
-
"epoch": 0.44,
|
120 |
-
"learning_rate": 1.1279720921692831e-05,
|
121 |
-
"loss": 0.0954,
|
122 |
-
"step": 9500
|
123 |
-
},
|
124 |
-
{
|
125 |
-
"epoch": 0.46,
|
126 |
-
"learning_rate": 1.0820710548058386e-05,
|
127 |
-
"loss": 0.0945,
|
128 |
-
"step": 10000
|
129 |
-
},
|
130 |
-
{
|
131 |
-
"epoch": 0.48,
|
132 |
-
"learning_rate": 1.0361700174423943e-05,
|
133 |
-
"loss": 0.0942,
|
134 |
-
"step": 10500
|
135 |
-
},
|
136 |
-
{
|
137 |
-
"epoch": 0.5,
|
138 |
-
"learning_rate": 9.9026898007895e-06,
|
139 |
-
"loss": 0.0948,
|
140 |
-
"step": 11000
|
141 |
-
},
|
142 |
-
{
|
143 |
-
"epoch": 0.53,
|
144 |
-
"learning_rate": 9.443679427155054e-06,
|
145 |
-
"loss": 0.0933,
|
146 |
-
"step": 11500
|
147 |
-
},
|
148 |
-
{
|
149 |
-
"epoch": 0.55,
|
150 |
-
"learning_rate": 8.98558707426788e-06,
|
151 |
-
"loss": 0.0949,
|
152 |
-
"step": 12000
|
153 |
-
},
|
154 |
-
{
|
155 |
-
"epoch": 0.57,
|
156 |
-
"learning_rate": 8.527494721380703e-06,
|
157 |
-
"loss": 0.0958,
|
158 |
-
"step": 12500
|
159 |
-
},
|
160 |
-
{
|
161 |
-
"epoch": 0.6,
|
162 |
-
"learning_rate": 8.06848434774626e-06,
|
163 |
-
"loss": 0.094,
|
164 |
-
"step": 13000
|
165 |
-
},
|
166 |
-
{
|
167 |
-
"epoch": 0.62,
|
168 |
-
"learning_rate": 7.609473974111815e-06,
|
169 |
-
"loss": 0.0939,
|
170 |
-
"step": 13500
|
171 |
-
},
|
172 |
-
{
|
173 |
-
"epoch": 0.64,
|
174 |
-
"learning_rate": 7.1504636004773705e-06,
|
175 |
-
"loss": 0.0943,
|
176 |
-
"step": 14000
|
177 |
-
},
|
178 |
-
{
|
179 |
-
"epoch": 0.67,
|
180 |
-
"learning_rate": 6.691453226842928e-06,
|
181 |
-
"loss": 0.0921,
|
182 |
-
"step": 14500
|
183 |
-
},
|
184 |
-
{
|
185 |
-
"epoch": 0.69,
|
186 |
-
"learning_rate": 6.232442853208484e-06,
|
187 |
-
"loss": 0.0942,
|
188 |
-
"step": 15000
|
189 |
-
},
|
190 |
-
{
|
191 |
-
"epoch": 0.71,
|
192 |
-
"learning_rate": 5.7734324795740396e-06,
|
193 |
-
"loss": 0.0942,
|
194 |
-
"step": 15500
|
195 |
-
},
|
196 |
-
{
|
197 |
-
"epoch": 0.73,
|
198 |
-
"learning_rate": 5.315340126686863e-06,
|
199 |
-
"loss": 0.0937,
|
200 |
-
"step": 16000
|
201 |
-
},
|
202 |
-
{
|
203 |
-
"epoch": 0.76,
|
204 |
-
"learning_rate": 4.85632975305242e-06,
|
205 |
-
"loss": 0.0942,
|
206 |
-
"step": 16500
|
207 |
-
},
|
208 |
-
{
|
209 |
-
"epoch": 0.78,
|
210 |
-
"learning_rate": 4.3973193794179755e-06,
|
211 |
-
"loss": 0.0933,
|
212 |
-
"step": 17000
|
213 |
-
},
|
214 |
-
{
|
215 |
-
"epoch": 0.8,
|
216 |
-
"learning_rate": 3.938309005783531e-06,
|
217 |
-
"loss": 0.0926,
|
218 |
-
"step": 17500
|
219 |
-
},
|
220 |
-
{
|
221 |
-
"epoch": 0.83,
|
222 |
-
"learning_rate": 3.4792986321490867e-06,
|
223 |
-
"loss": 0.0937,
|
224 |
-
"step": 18000
|
225 |
-
},
|
226 |
-
{
|
227 |
-
"epoch": 0.85,
|
228 |
-
"learning_rate": 3.0202882585146425e-06,
|
229 |
-
"loss": 0.0932,
|
230 |
-
"step": 18500
|
231 |
-
},
|
232 |
-
{
|
233 |
-
"epoch": 0.87,
|
234 |
-
"learning_rate": 2.5612778848801983e-06,
|
235 |
-
"loss": 0.0932,
|
236 |
-
"step": 19000
|
237 |
-
},
|
238 |
-
{
|
239 |
-
"epoch": 0.9,
|
240 |
-
"learning_rate": 2.1022675112457545e-06,
|
241 |
-
"loss": 0.0926,
|
242 |
-
"step": 19500
|
243 |
-
},
|
244 |
-
{
|
245 |
-
"epoch": 0.92,
|
246 |
-
"learning_rate": 1.645093179105848e-06,
|
247 |
-
"loss": 0.0917,
|
248 |
-
"step": 20000
|
249 |
-
},
|
250 |
-
{
|
251 |
-
"epoch": 0.94,
|
252 |
-
"learning_rate": 1.1860828054714039e-06,
|
253 |
-
"loss": 0.095,
|
254 |
-
"step": 20500
|
255 |
-
},
|
256 |
-
{
|
257 |
-
"epoch": 0.96,
|
258 |
-
"learning_rate": 7.270724318369595e-07,
|
259 |
-
"loss": 0.093,
|
260 |
-
"step": 21000
|
261 |
-
},
|
262 |
-
{
|
263 |
-
"epoch": 0.99,
|
264 |
-
"learning_rate": 2.680620582025154e-07,
|
265 |
-
"loss": 0.0942,
|
266 |
-
"step": 21500
|
267 |
-
}
|
268 |
-
],
|
269 |
-
"max_steps": 21786,
|
270 |
-
"num_train_epochs": 1,
|
271 |
-
"total_flos": 4.6557579706368e+16,
|
272 |
-
"trial_name": null,
|
273 |
-
"trial_params": null
|
274 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|