andrealexroom
commited on
Commit
•
a36206c
1
Parent(s):
c9389be
Upload folder using huggingface_hub
Browse files- model-00001-of-00003.safetensors +1 -1
- model-00002-of-00003.safetensors +1 -1
- model-00003-of-00003.safetensors +1 -1
- optimizer.pt +1 -1
- rng_state.pth +1 -1
- scheduler.pt +1 -1
- trainer_state.json +193 -3
model-00001-of-00003.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4943162336
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:59fd3d38352357e03340796aa3d8971284a6300e416f9c9771047098826e6a04
|
3 |
size 4943162336
|
model-00002-of-00003.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4999819336
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:45a24605afd021b013b6ed241e75490003fc0bf576c2ba3489d419e57010c834
|
3 |
size 4999819336
|
model-00003-of-00003.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4540516344
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:deb1c0ca71c6e75cbd07083e5e34b6048f7afe6bd822886bd38a09efbe64a6ab
|
3 |
size 4540516344
|
optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 28967182538
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6963c93c67db8111e9aa28dd7739df7ddccb4daedfd2fbafaa28316997e49e5f
|
3 |
size 28967182538
|
rng_state.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14244
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7c94c375fe5ad2903d244ca6b5cc2a1a6cba4c0c26196f3b9cbd9ddd170bb0b8
|
3 |
size 14244
|
scheduler.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 1000
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4ca4ef70a63953e55eeeb1484c79a338228b2f61b911ccc915369f389165807d
|
3 |
size 1000
|
trainer_state.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
-
"epoch": 0.
|
5 |
"eval_steps": 93,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -89,6 +89,196 @@
|
|
89 |
"eval_validation_altalexprivacy_samples_per_second": 1.259,
|
90 |
"eval_validation_altalexprivacy_steps_per_second": 0.638,
|
91 |
"step": 186
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
92 |
}
|
93 |
],
|
94 |
"logging_steps": 93,
|
@@ -96,7 +286,7 @@
|
|
96 |
"num_input_tokens_seen": 0,
|
97 |
"num_train_epochs": 2,
|
98 |
"save_steps": 93,
|
99 |
-
"total_flos":
|
100 |
"train_batch_size": 1,
|
101 |
"trial_name": null,
|
102 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.450928611628209,
|
5 |
"eval_steps": 93,
|
6 |
+
"global_step": 651,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
89 |
"eval_validation_altalexprivacy_samples_per_second": 1.259,
|
90 |
"eval_validation_altalexprivacy_steps_per_second": 0.638,
|
91 |
"step": 186
|
92 |
+
},
|
93 |
+
{
|
94 |
+
"epoch": 0.19,
|
95 |
+
"learning_rate": 1e-05,
|
96 |
+
"loss": 1.1339,
|
97 |
+
"step": 279
|
98 |
+
},
|
99 |
+
{
|
100 |
+
"epoch": 0.19,
|
101 |
+
"eval_validation_privacy_sources_loss": 0.8503767251968384,
|
102 |
+
"eval_validation_privacy_sources_runtime": 224.5704,
|
103 |
+
"eval_validation_privacy_sources_samples_per_second": 1.26,
|
104 |
+
"eval_validation_privacy_sources_steps_per_second": 0.632,
|
105 |
+
"step": 279
|
106 |
+
},
|
107 |
+
{
|
108 |
+
"epoch": 0.19,
|
109 |
+
"eval_validation_agenda_digitale_loss": 1.4141554832458496,
|
110 |
+
"eval_validation_agenda_digitale_runtime": 201.4979,
|
111 |
+
"eval_validation_agenda_digitale_samples_per_second": 1.261,
|
112 |
+
"eval_validation_agenda_digitale_steps_per_second": 0.63,
|
113 |
+
"step": 279
|
114 |
+
},
|
115 |
+
{
|
116 |
+
"epoch": 0.19,
|
117 |
+
"eval_validation_leggepertutti_loss": 1.2566713094711304,
|
118 |
+
"eval_validation_leggepertutti_runtime": 37.3026,
|
119 |
+
"eval_validation_leggepertutti_samples_per_second": 1.26,
|
120 |
+
"eval_validation_leggepertutti_steps_per_second": 0.643,
|
121 |
+
"step": 279
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"epoch": 0.19,
|
125 |
+
"eval_validation_altalexprivacy_loss": 1.2368062734603882,
|
126 |
+
"eval_validation_altalexprivacy_runtime": 62.6933,
|
127 |
+
"eval_validation_altalexprivacy_samples_per_second": 1.26,
|
128 |
+
"eval_validation_altalexprivacy_steps_per_second": 0.638,
|
129 |
+
"step": 279
|
130 |
+
},
|
131 |
+
{
|
132 |
+
"epoch": 0.26,
|
133 |
+
"learning_rate": 1e-05,
|
134 |
+
"loss": 1.1007,
|
135 |
+
"step": 372
|
136 |
+
},
|
137 |
+
{
|
138 |
+
"epoch": 0.26,
|
139 |
+
"eval_validation_privacy_sources_loss": 0.8314403295516968,
|
140 |
+
"eval_validation_privacy_sources_runtime": 224.5386,
|
141 |
+
"eval_validation_privacy_sources_samples_per_second": 1.26,
|
142 |
+
"eval_validation_privacy_sources_steps_per_second": 0.632,
|
143 |
+
"step": 372
|
144 |
+
},
|
145 |
+
{
|
146 |
+
"epoch": 0.26,
|
147 |
+
"eval_validation_agenda_digitale_loss": 1.3996949195861816,
|
148 |
+
"eval_validation_agenda_digitale_runtime": 201.5206,
|
149 |
+
"eval_validation_agenda_digitale_samples_per_second": 1.26,
|
150 |
+
"eval_validation_agenda_digitale_steps_per_second": 0.63,
|
151 |
+
"step": 372
|
152 |
+
},
|
153 |
+
{
|
154 |
+
"epoch": 0.26,
|
155 |
+
"eval_validation_leggepertutti_loss": 1.2466578483581543,
|
156 |
+
"eval_validation_leggepertutti_runtime": 37.3026,
|
157 |
+
"eval_validation_leggepertutti_samples_per_second": 1.26,
|
158 |
+
"eval_validation_leggepertutti_steps_per_second": 0.643,
|
159 |
+
"step": 372
|
160 |
+
},
|
161 |
+
{
|
162 |
+
"epoch": 0.26,
|
163 |
+
"eval_validation_altalexprivacy_loss": 1.2185348272323608,
|
164 |
+
"eval_validation_altalexprivacy_runtime": 62.6898,
|
165 |
+
"eval_validation_altalexprivacy_samples_per_second": 1.26,
|
166 |
+
"eval_validation_altalexprivacy_steps_per_second": 0.638,
|
167 |
+
"step": 372
|
168 |
+
},
|
169 |
+
{
|
170 |
+
"epoch": 0.32,
|
171 |
+
"learning_rate": 1e-05,
|
172 |
+
"loss": 1.0747,
|
173 |
+
"step": 465
|
174 |
+
},
|
175 |
+
{
|
176 |
+
"epoch": 0.32,
|
177 |
+
"eval_validation_privacy_sources_loss": 0.8157272338867188,
|
178 |
+
"eval_validation_privacy_sources_runtime": 224.5077,
|
179 |
+
"eval_validation_privacy_sources_samples_per_second": 1.261,
|
180 |
+
"eval_validation_privacy_sources_steps_per_second": 0.632,
|
181 |
+
"step": 465
|
182 |
+
},
|
183 |
+
{
|
184 |
+
"epoch": 0.32,
|
185 |
+
"eval_validation_agenda_digitale_loss": 1.3890674114227295,
|
186 |
+
"eval_validation_agenda_digitale_runtime": 201.5006,
|
187 |
+
"eval_validation_agenda_digitale_samples_per_second": 1.261,
|
188 |
+
"eval_validation_agenda_digitale_steps_per_second": 0.63,
|
189 |
+
"step": 465
|
190 |
+
},
|
191 |
+
{
|
192 |
+
"epoch": 0.32,
|
193 |
+
"eval_validation_leggepertutti_loss": 1.2424800395965576,
|
194 |
+
"eval_validation_leggepertutti_runtime": 37.28,
|
195 |
+
"eval_validation_leggepertutti_samples_per_second": 1.261,
|
196 |
+
"eval_validation_leggepertutti_steps_per_second": 0.644,
|
197 |
+
"step": 465
|
198 |
+
},
|
199 |
+
{
|
200 |
+
"epoch": 0.32,
|
201 |
+
"eval_validation_altalexprivacy_loss": 1.203458547592163,
|
202 |
+
"eval_validation_altalexprivacy_runtime": 62.6722,
|
203 |
+
"eval_validation_altalexprivacy_samples_per_second": 1.261,
|
204 |
+
"eval_validation_altalexprivacy_steps_per_second": 0.638,
|
205 |
+
"step": 465
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 0.39,
|
209 |
+
"learning_rate": 1e-05,
|
210 |
+
"loss": 1.0703,
|
211 |
+
"step": 558
|
212 |
+
},
|
213 |
+
{
|
214 |
+
"epoch": 0.39,
|
215 |
+
"eval_validation_privacy_sources_loss": 0.8016490936279297,
|
216 |
+
"eval_validation_privacy_sources_runtime": 224.4789,
|
217 |
+
"eval_validation_privacy_sources_samples_per_second": 1.261,
|
218 |
+
"eval_validation_privacy_sources_steps_per_second": 0.633,
|
219 |
+
"step": 558
|
220 |
+
},
|
221 |
+
{
|
222 |
+
"epoch": 0.39,
|
223 |
+
"eval_validation_agenda_digitale_loss": 1.3786152601242065,
|
224 |
+
"eval_validation_agenda_digitale_runtime": 201.4702,
|
225 |
+
"eval_validation_agenda_digitale_samples_per_second": 1.261,
|
226 |
+
"eval_validation_agenda_digitale_steps_per_second": 0.63,
|
227 |
+
"step": 558
|
228 |
+
},
|
229 |
+
{
|
230 |
+
"epoch": 0.39,
|
231 |
+
"eval_validation_leggepertutti_loss": 1.23654043674469,
|
232 |
+
"eval_validation_leggepertutti_runtime": 37.2956,
|
233 |
+
"eval_validation_leggepertutti_samples_per_second": 1.26,
|
234 |
+
"eval_validation_leggepertutti_steps_per_second": 0.644,
|
235 |
+
"step": 558
|
236 |
+
},
|
237 |
+
{
|
238 |
+
"epoch": 0.39,
|
239 |
+
"eval_validation_altalexprivacy_loss": 1.1914178133010864,
|
240 |
+
"eval_validation_altalexprivacy_runtime": 62.6741,
|
241 |
+
"eval_validation_altalexprivacy_samples_per_second": 1.26,
|
242 |
+
"eval_validation_altalexprivacy_steps_per_second": 0.638,
|
243 |
+
"step": 558
|
244 |
+
},
|
245 |
+
{
|
246 |
+
"epoch": 0.45,
|
247 |
+
"learning_rate": 1e-05,
|
248 |
+
"loss": 1.0514,
|
249 |
+
"step": 651
|
250 |
+
},
|
251 |
+
{
|
252 |
+
"epoch": 0.45,
|
253 |
+
"eval_validation_privacy_sources_loss": 0.7911774516105652,
|
254 |
+
"eval_validation_privacy_sources_runtime": 224.3988,
|
255 |
+
"eval_validation_privacy_sources_samples_per_second": 1.261,
|
256 |
+
"eval_validation_privacy_sources_steps_per_second": 0.633,
|
257 |
+
"step": 651
|
258 |
+
},
|
259 |
+
{
|
260 |
+
"epoch": 0.45,
|
261 |
+
"eval_validation_agenda_digitale_loss": 1.370524525642395,
|
262 |
+
"eval_validation_agenda_digitale_runtime": 201.3956,
|
263 |
+
"eval_validation_agenda_digitale_samples_per_second": 1.261,
|
264 |
+
"eval_validation_agenda_digitale_steps_per_second": 0.631,
|
265 |
+
"step": 651
|
266 |
+
},
|
267 |
+
{
|
268 |
+
"epoch": 0.45,
|
269 |
+
"eval_validation_leggepertutti_loss": 1.2246781587600708,
|
270 |
+
"eval_validation_leggepertutti_runtime": 37.2851,
|
271 |
+
"eval_validation_leggepertutti_samples_per_second": 1.261,
|
272 |
+
"eval_validation_leggepertutti_steps_per_second": 0.644,
|
273 |
+
"step": 651
|
274 |
+
},
|
275 |
+
{
|
276 |
+
"epoch": 0.45,
|
277 |
+
"eval_validation_altalexprivacy_loss": 1.1746492385864258,
|
278 |
+
"eval_validation_altalexprivacy_runtime": 62.6484,
|
279 |
+
"eval_validation_altalexprivacy_samples_per_second": 1.261,
|
280 |
+
"eval_validation_altalexprivacy_steps_per_second": 0.638,
|
281 |
+
"step": 651
|
282 |
}
|
283 |
],
|
284 |
"logging_steps": 93,
|
|
|
286 |
"num_input_tokens_seen": 0,
|
287 |
"num_train_epochs": 2,
|
288 |
"save_steps": 93,
|
289 |
+
"total_flos": 7.280849934083949e+18,
|
290 |
"train_batch_size": 1,
|
291 |
"trial_name": null,
|
292 |
"trial_params": null
|