init
Browse files- checkpoint-4000/config.json +0 -26
- checkpoint-4000/global_step4000/mp_rank_00_model_states.pt +0 -3
- checkpoint-4000/global_step4000/zero_pp_rank_2_mp_rank_00_optim_states.pt +0 -3
- checkpoint-4000/pytorch_model-00001-of-00002.bin +0 -3
- checkpoint-4000/rng_state_1.pth +0 -3
- checkpoint-4000/rng_state_2.pth +0 -3
- checkpoint-4000/rng_state_3.pth +0 -3
- checkpoint-4000/rng_state_4.pth +0 -3
- checkpoint-4000/rng_state_5.pth +0 -3
- checkpoint-4000/rng_state_6.pth +0 -3
- checkpoint-4000/rng_state_7.pth +0 -3
- checkpoint-4000/trainer_state.json +0 -3142
- checkpoint-4000/training_args.bin +0 -3
- checkpoint-4000/zero_to_fp32.py +0 -578
checkpoint-4000/config.json
DELETED
@@ -1,26 +0,0 @@
|
|
1 |
-
{
|
2 |
-
"_name_or_path": "/data/searchgpt/yq/GoGPT/outputs-pt-v1-7b-llama2/ckpt",
|
3 |
-
"architectures": [
|
4 |
-
"LlamaForCausalLM"
|
5 |
-
],
|
6 |
-
"bos_token_id": 1,
|
7 |
-
"eos_token_id": 2,
|
8 |
-
"hidden_act": "silu",
|
9 |
-
"hidden_size": 4096,
|
10 |
-
"initializer_range": 0.02,
|
11 |
-
"intermediate_size": 11008,
|
12 |
-
"max_position_embeddings": 2048,
|
13 |
-
"model_type": "llama",
|
14 |
-
"num_attention_heads": 32,
|
15 |
-
"num_hidden_layers": 32,
|
16 |
-
"num_key_value_heads": 32,
|
17 |
-
"pad_token_id": 0,
|
18 |
-
"pretraining_tp": 1,
|
19 |
-
"rms_norm_eps": 1e-05,
|
20 |
-
"rope_scaling": null,
|
21 |
-
"tie_word_embeddings": false,
|
22 |
-
"torch_dtype": "float16",
|
23 |
-
"transformers_version": "4.29.1",
|
24 |
-
"use_cache": false,
|
25 |
-
"vocab_size": 68419
|
26 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-4000/global_step4000/mp_rank_00_model_states.pt
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:814e8129fe526c42fdbd6fe4ff2c9b21f0c161f6982e3883c3e5fea1260a4cb1
|
3 |
-
size 2607759360
|
|
|
|
|
|
|
|
checkpoint-4000/global_step4000/zero_pp_rank_2_mp_rank_00_optim_states.pt
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:aabd69e33896e791496a8b91b6325d7e6dafe50483e2ed374bf68f6bbc025924
|
3 |
-
size 840736473
|
|
|
|
|
|
|
|
checkpoint-4000/pytorch_model-00001-of-00002.bin
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:bcb73d2c6cc7cec148b9ea0157637718f3414f3a18d493daf4255e47befedae9
|
3 |
-
size 10531361877
|
|
|
|
|
|
|
|
checkpoint-4000/rng_state_1.pth
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:f7f1e51f8b260af2143aa0e342ab171171f7d983fbaaf95e9400d872b4d8c542
|
3 |
-
size 21687
|
|
|
|
|
|
|
|
checkpoint-4000/rng_state_2.pth
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:057591ad3bc33c9f75b1e6556c3c29b4328e0785d67a420f2a5a8bb25e58812e
|
3 |
-
size 21687
|
|
|
|
|
|
|
|
checkpoint-4000/rng_state_3.pth
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:5b595d89f2fe53ca5dccc290b880b858a60eb07a5f694175f05519f2257fafbf
|
3 |
-
size 21687
|
|
|
|
|
|
|
|
checkpoint-4000/rng_state_4.pth
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:8dae996da909b80faa4e9fef9f945126f53fa3c7c6da06367a3f71ae24b0f138
|
3 |
-
size 21687
|
|
|
|
|
|
|
|
checkpoint-4000/rng_state_5.pth
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:8de190b28a07218f4d05a1c7ccfccc4e53d6033af244fe3bd9fe3e20275d59de
|
3 |
-
size 21687
|
|
|
|
|
|
|
|
checkpoint-4000/rng_state_6.pth
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:a4cccb61a1758c3ec09da0ddc7ef0957d32362d5baa800f340839c1fa72376e2
|
3 |
-
size 21687
|
|
|
|
|
|
|
|
checkpoint-4000/rng_state_7.pth
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:ed085eb5cf063033df209f9c7714e3c87ee78a57b9ee629a7258a15bd8cde7ac
|
3 |
-
size 21687
|
|
|
|
|
|
|
|
checkpoint-4000/trainer_state.json
DELETED
@@ -1,3142 +0,0 @@
|
|
1 |
-
{
|
2 |
-
"best_metric": null,
|
3 |
-
"best_model_checkpoint": null,
|
4 |
-
"epoch": 0.8226221079691517,
|
5 |
-
"global_step": 4000,
|
6 |
-
"is_hyper_param_search": false,
|
7 |
-
"is_local_process_zero": true,
|
8 |
-
"is_world_process_zero": true,
|
9 |
-
"log_history": [
|
10 |
-
{
|
11 |
-
"epoch": 0.0,
|
12 |
-
"learning_rate": 0,
|
13 |
-
"loss": 11.981,
|
14 |
-
"step": 1
|
15 |
-
},
|
16 |
-
{
|
17 |
-
"epoch": 0.0,
|
18 |
-
"learning_rate": 3.9970117109141705e-05,
|
19 |
-
"loss": 12.0789,
|
20 |
-
"step": 10
|
21 |
-
},
|
22 |
-
{
|
23 |
-
"epoch": 0.0,
|
24 |
-
"learning_rate": 9.331893267009234e-05,
|
25 |
-
"loss": 10.7133,
|
26 |
-
"step": 20
|
27 |
-
},
|
28 |
-
{
|
29 |
-
"epoch": 0.01,
|
30 |
-
"learning_rate": 0.00011407670594843083,
|
31 |
-
"loss": 8.7339,
|
32 |
-
"step": 30
|
33 |
-
},
|
34 |
-
{
|
35 |
-
"epoch": 0.01,
|
36 |
-
"learning_rate": 0.00012721122651399258,
|
37 |
-
"loss": 8.301,
|
38 |
-
"step": 40
|
39 |
-
},
|
40 |
-
{
|
41 |
-
"epoch": 0.01,
|
42 |
-
"learning_rate": 0.00013684136855727938,
|
43 |
-
"loss": 8.1964,
|
44 |
-
"step": 50
|
45 |
-
},
|
46 |
-
{
|
47 |
-
"epoch": 0.01,
|
48 |
-
"eval_accuracy": 0.10310832025117739,
|
49 |
-
"eval_loss": 8.203125,
|
50 |
-
"eval_runtime": 6.5764,
|
51 |
-
"eval_samples_per_second": 5.322,
|
52 |
-
"eval_steps_per_second": 0.304,
|
53 |
-
"step": 50
|
54 |
-
},
|
55 |
-
{
|
56 |
-
"epoch": 0.01,
|
57 |
-
"learning_rate": 0.00014444862339428802,
|
58 |
-
"loss": 8.0553,
|
59 |
-
"step": 60
|
60 |
-
},
|
61 |
-
{
|
62 |
-
"epoch": 0.01,
|
63 |
-
"learning_rate": 0.00015073705430110066,
|
64 |
-
"loss": 7.9436,
|
65 |
-
"step": 70
|
66 |
-
},
|
67 |
-
{
|
68 |
-
"epoch": 0.02,
|
69 |
-
"learning_rate": 0.00015609707636042195,
|
70 |
-
"loss": 7.8368,
|
71 |
-
"step": 80
|
72 |
-
},
|
73 |
-
{
|
74 |
-
"epoch": 0.02,
|
75 |
-
"learning_rate": 0.00016076788727202945,
|
76 |
-
"loss": 7.7333,
|
77 |
-
"step": 90
|
78 |
-
},
|
79 |
-
{
|
80 |
-
"epoch": 0.02,
|
81 |
-
"learning_rate": 0.00016490670495758757,
|
82 |
-
"loss": 7.6139,
|
83 |
-
"step": 100
|
84 |
-
},
|
85 |
-
{
|
86 |
-
"epoch": 0.02,
|
87 |
-
"eval_accuracy": 0.12459794174079888,
|
88 |
-
"eval_loss": 7.81640625,
|
89 |
-
"eval_runtime": 6.6059,
|
90 |
-
"eval_samples_per_second": 5.298,
|
91 |
-
"eval_steps_per_second": 0.303,
|
92 |
-
"step": 100
|
93 |
-
},
|
94 |
-
{
|
95 |
-
"epoch": 0.02,
|
96 |
-
"learning_rate": 0.0001686224178807056,
|
97 |
-
"loss": 7.4892,
|
98 |
-
"step": 110
|
99 |
-
},
|
100 |
-
{
|
101 |
-
"epoch": 0.02,
|
102 |
-
"learning_rate": 0.000171993565594773,
|
103 |
-
"loss": 7.3256,
|
104 |
-
"step": 120
|
105 |
-
},
|
106 |
-
{
|
107 |
-
"epoch": 0.03,
|
108 |
-
"learning_rate": 0.00017507866443784335,
|
109 |
-
"loss": 7.1827,
|
110 |
-
"step": 130
|
111 |
-
},
|
112 |
-
{
|
113 |
-
"epoch": 0.03,
|
114 |
-
"learning_rate": 0.0001779224840062419,
|
115 |
-
"loss": 6.9698,
|
116 |
-
"step": 140
|
117 |
-
},
|
118 |
-
{
|
119 |
-
"epoch": 0.03,
|
120 |
-
"learning_rate": 0.00018056004207494319,
|
121 |
-
"loss": 6.8162,
|
122 |
-
"step": 150
|
123 |
-
},
|
124 |
-
{
|
125 |
-
"epoch": 0.03,
|
126 |
-
"eval_accuracy": 0.1567556253270539,
|
127 |
-
"eval_loss": 7.08203125,
|
128 |
-
"eval_runtime": 6.5908,
|
129 |
-
"eval_samples_per_second": 5.31,
|
130 |
-
"eval_steps_per_second": 0.303,
|
131 |
-
"step": 150
|
132 |
-
},
|
133 |
-
{
|
134 |
-
"epoch": 0.03,
|
135 |
-
"learning_rate": 0.00018301924610008189,
|
136 |
-
"loss": 6.6293,
|
137 |
-
"step": 160
|
138 |
-
},
|
139 |
-
{
|
140 |
-
"epoch": 0.03,
|
141 |
-
"learning_rate": 0.00018532269677939782,
|
142 |
-
"loss": 6.4114,
|
143 |
-
"step": 170
|
144 |
-
},
|
145 |
-
{
|
146 |
-
"epoch": 0.04,
|
147 |
-
"learning_rate": 0.00018748895370481112,
|
148 |
-
"loss": 6.2911,
|
149 |
-
"step": 180
|
150 |
-
},
|
151 |
-
{
|
152 |
-
"epoch": 0.04,
|
153 |
-
"learning_rate": 0.00018953344483335556,
|
154 |
-
"loss": 6.1047,
|
155 |
-
"step": 190
|
156 |
-
},
|
157 |
-
{
|
158 |
-
"epoch": 0.04,
|
159 |
-
"learning_rate": 0.00019146913367833817,
|
160 |
-
"loss": 5.9957,
|
161 |
-
"step": 200
|
162 |
-
},
|
163 |
-
{
|
164 |
-
"epoch": 0.04,
|
165 |
-
"eval_accuracy": 0.19811616954474098,
|
166 |
-
"eval_loss": 6.4296875,
|
167 |
-
"eval_runtime": 6.5897,
|
168 |
-
"eval_samples_per_second": 5.311,
|
169 |
-
"eval_steps_per_second": 0.304,
|
170 |
-
"step": 200
|
171 |
-
},
|
172 |
-
{
|
173 |
-
"epoch": 0.04,
|
174 |
-
"learning_rate": 0.00019330701776944063,
|
175 |
-
"loss": 5.8281,
|
176 |
-
"step": 210
|
177 |
-
},
|
178 |
-
{
|
179 |
-
"epoch": 0.05,
|
180 |
-
"learning_rate": 0.00019505650713185044,
|
181 |
-
"loss": 5.6927,
|
182 |
-
"step": 220
|
183 |
-
},
|
184 |
-
{
|
185 |
-
"epoch": 0.05,
|
186 |
-
"learning_rate": 0.00019672571585424665,
|
187 |
-
"loss": 5.5564,
|
188 |
-
"step": 230
|
189 |
-
},
|
190 |
-
{
|
191 |
-
"epoch": 0.05,
|
192 |
-
"learning_rate": 0.00019832168964685297,
|
193 |
-
"loss": 5.3813,
|
194 |
-
"step": 240
|
195 |
-
},
|
196 |
-
{
|
197 |
-
"epoch": 0.05,
|
198 |
-
"learning_rate": 0.0001998505855457085,
|
199 |
-
"loss": 5.2496,
|
200 |
-
"step": 250
|
201 |
-
},
|
202 |
-
{
|
203 |
-
"epoch": 0.05,
|
204 |
-
"eval_accuracy": 0.24055817198674342,
|
205 |
-
"eval_loss": 5.8203125,
|
206 |
-
"eval_runtime": 6.6013,
|
207 |
-
"eval_samples_per_second": 5.302,
|
208 |
-
"eval_steps_per_second": 0.303,
|
209 |
-
"step": 250
|
210 |
-
},
|
211 |
-
{
|
212 |
-
"epoch": 0.05,
|
213 |
-
"learning_rate": 0.0001996535296665223,
|
214 |
-
"loss": 5.13,
|
215 |
-
"step": 260
|
216 |
-
},
|
217 |
-
{
|
218 |
-
"epoch": 0.06,
|
219 |
-
"learning_rate": 0.0001992204417496752,
|
220 |
-
"loss": 5.0354,
|
221 |
-
"step": 270
|
222 |
-
},
|
223 |
-
{
|
224 |
-
"epoch": 0.06,
|
225 |
-
"learning_rate": 0.00019878735383282807,
|
226 |
-
"loss": 4.9021,
|
227 |
-
"step": 280
|
228 |
-
},
|
229 |
-
{
|
230 |
-
"epoch": 0.06,
|
231 |
-
"learning_rate": 0.00019835426591598097,
|
232 |
-
"loss": 4.8181,
|
233 |
-
"step": 290
|
234 |
-
},
|
235 |
-
{
|
236 |
-
"epoch": 0.06,
|
237 |
-
"learning_rate": 0.00019792117799913384,
|
238 |
-
"loss": 4.6993,
|
239 |
-
"step": 300
|
240 |
-
},
|
241 |
-
{
|
242 |
-
"epoch": 0.06,
|
243 |
-
"eval_accuracy": 0.2641758241758242,
|
244 |
-
"eval_loss": 5.41796875,
|
245 |
-
"eval_runtime": 6.6009,
|
246 |
-
"eval_samples_per_second": 5.302,
|
247 |
-
"eval_steps_per_second": 0.303,
|
248 |
-
"step": 300
|
249 |
-
},
|
250 |
-
{
|
251 |
-
"epoch": 0.06,
|
252 |
-
"learning_rate": 0.0001974880900822867,
|
253 |
-
"loss": 4.6761,
|
254 |
-
"step": 310
|
255 |
-
},
|
256 |
-
{
|
257 |
-
"epoch": 0.07,
|
258 |
-
"learning_rate": 0.00019705500216543958,
|
259 |
-
"loss": 4.5908,
|
260 |
-
"step": 320
|
261 |
-
},
|
262 |
-
{
|
263 |
-
"epoch": 0.07,
|
264 |
-
"learning_rate": 0.00019662191424859245,
|
265 |
-
"loss": 4.5301,
|
266 |
-
"step": 330
|
267 |
-
},
|
268 |
-
{
|
269 |
-
"epoch": 0.07,
|
270 |
-
"learning_rate": 0.00019618882633174535,
|
271 |
-
"loss": 4.4729,
|
272 |
-
"step": 340
|
273 |
-
},
|
274 |
-
{
|
275 |
-
"epoch": 0.07,
|
276 |
-
"learning_rate": 0.00019575573841489822,
|
277 |
-
"loss": 4.3928,
|
278 |
-
"step": 350
|
279 |
-
},
|
280 |
-
{
|
281 |
-
"epoch": 0.07,
|
282 |
-
"eval_accuracy": 0.27927437641723357,
|
283 |
-
"eval_loss": 5.14453125,
|
284 |
-
"eval_runtime": 6.6101,
|
285 |
-
"eval_samples_per_second": 5.295,
|
286 |
-
"eval_steps_per_second": 0.303,
|
287 |
-
"step": 350
|
288 |
-
},
|
289 |
-
{
|
290 |
-
"epoch": 0.07,
|
291 |
-
"learning_rate": 0.00019532265049805112,
|
292 |
-
"loss": 4.4083,
|
293 |
-
"step": 360
|
294 |
-
},
|
295 |
-
{
|
296 |
-
"epoch": 0.08,
|
297 |
-
"learning_rate": 0.000194889562581204,
|
298 |
-
"loss": 4.336,
|
299 |
-
"step": 370
|
300 |
-
},
|
301 |
-
{
|
302 |
-
"epoch": 0.08,
|
303 |
-
"learning_rate": 0.0001944564746643569,
|
304 |
-
"loss": 4.2714,
|
305 |
-
"step": 380
|
306 |
-
},
|
307 |
-
{
|
308 |
-
"epoch": 0.08,
|
309 |
-
"learning_rate": 0.00019402338674750976,
|
310 |
-
"loss": 4.2124,
|
311 |
-
"step": 390
|
312 |
-
},
|
313 |
-
{
|
314 |
-
"epoch": 0.08,
|
315 |
-
"learning_rate": 0.00019359029883066263,
|
316 |
-
"loss": 4.2395,
|
317 |
-
"step": 400
|
318 |
-
},
|
319 |
-
{
|
320 |
-
"epoch": 0.08,
|
321 |
-
"eval_accuracy": 0.2894470608756323,
|
322 |
-
"eval_loss": 4.96875,
|
323 |
-
"eval_runtime": 6.5945,
|
324 |
-
"eval_samples_per_second": 5.307,
|
325 |
-
"eval_steps_per_second": 0.303,
|
326 |
-
"step": 400
|
327 |
-
},
|
328 |
-
{
|
329 |
-
"epoch": 0.08,
|
330 |
-
"learning_rate": 0.0001931572109138155,
|
331 |
-
"loss": 4.1867,
|
332 |
-
"step": 410
|
333 |
-
},
|
334 |
-
{
|
335 |
-
"epoch": 0.09,
|
336 |
-
"learning_rate": 0.0001927241229969684,
|
337 |
-
"loss": 4.1687,
|
338 |
-
"step": 420
|
339 |
-
},
|
340 |
-
{
|
341 |
-
"epoch": 0.09,
|
342 |
-
"learning_rate": 0.00019229103508012127,
|
343 |
-
"loss": 4.1027,
|
344 |
-
"step": 430
|
345 |
-
},
|
346 |
-
{
|
347 |
-
"epoch": 0.09,
|
348 |
-
"learning_rate": 0.00019185794716327414,
|
349 |
-
"loss": 4.1233,
|
350 |
-
"step": 440
|
351 |
-
},
|
352 |
-
{
|
353 |
-
"epoch": 0.09,
|
354 |
-
"learning_rate": 0.00019142485924642704,
|
355 |
-
"loss": 4.0781,
|
356 |
-
"step": 450
|
357 |
-
},
|
358 |
-
{
|
359 |
-
"epoch": 0.09,
|
360 |
-
"eval_accuracy": 0.29638932496075354,
|
361 |
-
"eval_loss": 4.8515625,
|
362 |
-
"eval_runtime": 6.5996,
|
363 |
-
"eval_samples_per_second": 5.303,
|
364 |
-
"eval_steps_per_second": 0.303,
|
365 |
-
"step": 450
|
366 |
-
},
|
367 |
-
{
|
368 |
-
"epoch": 0.09,
|
369 |
-
"learning_rate": 0.0001909917713295799,
|
370 |
-
"loss": 4.0855,
|
371 |
-
"step": 460
|
372 |
-
},
|
373 |
-
{
|
374 |
-
"epoch": 0.1,
|
375 |
-
"learning_rate": 0.0001905586834127328,
|
376 |
-
"loss": 4.0859,
|
377 |
-
"step": 470
|
378 |
-
},
|
379 |
-
{
|
380 |
-
"epoch": 0.1,
|
381 |
-
"learning_rate": 0.00019012559549588568,
|
382 |
-
"loss": 4.0124,
|
383 |
-
"step": 480
|
384 |
-
},
|
385 |
-
{
|
386 |
-
"epoch": 0.1,
|
387 |
-
"learning_rate": 0.00018969250757903855,
|
388 |
-
"loss": 4.0151,
|
389 |
-
"step": 490
|
390 |
-
},
|
391 |
-
{
|
392 |
-
"epoch": 0.1,
|
393 |
-
"learning_rate": 0.00018925941966219142,
|
394 |
-
"loss": 4.0409,
|
395 |
-
"step": 500
|
396 |
-
},
|
397 |
-
{
|
398 |
-
"epoch": 0.1,
|
399 |
-
"eval_accuracy": 0.30181057038199893,
|
400 |
-
"eval_loss": 4.76953125,
|
401 |
-
"eval_runtime": 6.6117,
|
402 |
-
"eval_samples_per_second": 5.294,
|
403 |
-
"eval_steps_per_second": 0.302,
|
404 |
-
"step": 500
|
405 |
-
},
|
406 |
-
{
|
407 |
-
"epoch": 0.1,
|
408 |
-
"learning_rate": 0.00018882633174534431,
|
409 |
-
"loss": 3.9912,
|
410 |
-
"step": 510
|
411 |
-
},
|
412 |
-
{
|
413 |
-
"epoch": 0.11,
|
414 |
-
"learning_rate": 0.00018839324382849718,
|
415 |
-
"loss": 3.9383,
|
416 |
-
"step": 520
|
417 |
-
},
|
418 |
-
{
|
419 |
-
"epoch": 0.11,
|
420 |
-
"learning_rate": 0.00018796015591165008,
|
421 |
-
"loss": 3.9764,
|
422 |
-
"step": 530
|
423 |
-
},
|
424 |
-
{
|
425 |
-
"epoch": 0.11,
|
426 |
-
"learning_rate": 0.00018752706799480295,
|
427 |
-
"loss": 3.9809,
|
428 |
-
"step": 540
|
429 |
-
},
|
430 |
-
{
|
431 |
-
"epoch": 0.11,
|
432 |
-
"learning_rate": 0.00018709398007795585,
|
433 |
-
"loss": 3.9178,
|
434 |
-
"step": 550
|
435 |
-
},
|
436 |
-
{
|
437 |
-
"epoch": 0.11,
|
438 |
-
"eval_accuracy": 0.30575963718820864,
|
439 |
-
"eval_loss": 4.703125,
|
440 |
-
"eval_runtime": 6.6265,
|
441 |
-
"eval_samples_per_second": 5.282,
|
442 |
-
"eval_steps_per_second": 0.302,
|
443 |
-
"step": 550
|
444 |
-
},
|
445 |
-
{
|
446 |
-
"epoch": 0.12,
|
447 |
-
"learning_rate": 0.00018666089216110872,
|
448 |
-
"loss": 3.9073,
|
449 |
-
"step": 560
|
450 |
-
},
|
451 |
-
{
|
452 |
-
"epoch": 0.12,
|
453 |
-
"learning_rate": 0.0001862278042442616,
|
454 |
-
"loss": 3.9459,
|
455 |
-
"step": 570
|
456 |
-
},
|
457 |
-
{
|
458 |
-
"epoch": 0.12,
|
459 |
-
"learning_rate": 0.00018579471632741446,
|
460 |
-
"loss": 3.9535,
|
461 |
-
"step": 580
|
462 |
-
},
|
463 |
-
{
|
464 |
-
"epoch": 0.12,
|
465 |
-
"learning_rate": 0.00018536162841056733,
|
466 |
-
"loss": 3.8982,
|
467 |
-
"step": 590
|
468 |
-
},
|
469 |
-
{
|
470 |
-
"epoch": 0.12,
|
471 |
-
"learning_rate": 0.00018492854049372023,
|
472 |
-
"loss": 3.834,
|
473 |
-
"step": 600
|
474 |
-
},
|
475 |
-
{
|
476 |
-
"epoch": 0.12,
|
477 |
-
"eval_accuracy": 0.3082574568288854,
|
478 |
-
"eval_loss": 4.65625,
|
479 |
-
"eval_runtime": 6.6135,
|
480 |
-
"eval_samples_per_second": 5.292,
|
481 |
-
"eval_steps_per_second": 0.302,
|
482 |
-
"step": 600
|
483 |
-
},
|
484 |
-
{
|
485 |
-
"epoch": 0.13,
|
486 |
-
"learning_rate": 0.0001844954525768731,
|
487 |
-
"loss": 3.8759,
|
488 |
-
"step": 610
|
489 |
-
},
|
490 |
-
{
|
491 |
-
"epoch": 0.13,
|
492 |
-
"learning_rate": 0.000184062364660026,
|
493 |
-
"loss": 3.8835,
|
494 |
-
"step": 620
|
495 |
-
},
|
496 |
-
{
|
497 |
-
"epoch": 0.13,
|
498 |
-
"learning_rate": 0.00018362927674317887,
|
499 |
-
"loss": 3.9003,
|
500 |
-
"step": 630
|
501 |
-
},
|
502 |
-
{
|
503 |
-
"epoch": 0.13,
|
504 |
-
"learning_rate": 0.00018319618882633177,
|
505 |
-
"loss": 3.8538,
|
506 |
-
"step": 640
|
507 |
-
},
|
508 |
-
{
|
509 |
-
"epoch": 0.13,
|
510 |
-
"learning_rate": 0.00018276310090948464,
|
511 |
-
"loss": 3.8316,
|
512 |
-
"step": 650
|
513 |
-
},
|
514 |
-
{
|
515 |
-
"epoch": 0.13,
|
516 |
-
"eval_accuracy": 0.30981336124193265,
|
517 |
-
"eval_loss": 4.625,
|
518 |
-
"eval_runtime": 6.6039,
|
519 |
-
"eval_samples_per_second": 5.3,
|
520 |
-
"eval_steps_per_second": 0.303,
|
521 |
-
"step": 650
|
522 |
-
},
|
523 |
-
{
|
524 |
-
"epoch": 0.14,
|
525 |
-
"learning_rate": 0.00018233001299263754,
|
526 |
-
"loss": 3.9085,
|
527 |
-
"step": 660
|
528 |
-
},
|
529 |
-
{
|
530 |
-
"epoch": 0.14,
|
531 |
-
"learning_rate": 0.00018189692507579038,
|
532 |
-
"loss": 3.7825,
|
533 |
-
"step": 670
|
534 |
-
},
|
535 |
-
{
|
536 |
-
"epoch": 0.14,
|
537 |
-
"learning_rate": 0.00018146383715894328,
|
538 |
-
"loss": 3.824,
|
539 |
-
"step": 680
|
540 |
-
},
|
541 |
-
{
|
542 |
-
"epoch": 0.14,
|
543 |
-
"learning_rate": 0.00018103074924209615,
|
544 |
-
"loss": 3.8457,
|
545 |
-
"step": 690
|
546 |
-
},
|
547 |
-
{
|
548 |
-
"epoch": 0.14,
|
549 |
-
"learning_rate": 0.00018059766132524902,
|
550 |
-
"loss": 3.8197,
|
551 |
-
"step": 700
|
552 |
-
},
|
553 |
-
{
|
554 |
-
"epoch": 0.14,
|
555 |
-
"eval_accuracy": 0.3116902145473574,
|
556 |
-
"eval_loss": 4.59765625,
|
557 |
-
"eval_runtime": 6.6155,
|
558 |
-
"eval_samples_per_second": 5.291,
|
559 |
-
"eval_steps_per_second": 0.302,
|
560 |
-
"step": 700
|
561 |
-
},
|
562 |
-
{
|
563 |
-
"epoch": 0.15,
|
564 |
-
"learning_rate": 0.00018016457340840192,
|
565 |
-
"loss": 3.79,
|
566 |
-
"step": 710
|
567 |
-
},
|
568 |
-
{
|
569 |
-
"epoch": 0.15,
|
570 |
-
"learning_rate": 0.00017973148549155479,
|
571 |
-
"loss": 3.7907,
|
572 |
-
"step": 720
|
573 |
-
},
|
574 |
-
{
|
575 |
-
"epoch": 0.15,
|
576 |
-
"learning_rate": 0.00017929839757470768,
|
577 |
-
"loss": 3.7797,
|
578 |
-
"step": 730
|
579 |
-
},
|
580 |
-
{
|
581 |
-
"epoch": 0.15,
|
582 |
-
"learning_rate": 0.00017886530965786055,
|
583 |
-
"loss": 3.7533,
|
584 |
-
"step": 740
|
585 |
-
},
|
586 |
-
{
|
587 |
-
"epoch": 0.15,
|
588 |
-
"learning_rate": 0.00017843222174101345,
|
589 |
-
"loss": 3.7464,
|
590 |
-
"step": 750
|
591 |
-
},
|
592 |
-
{
|
593 |
-
"epoch": 0.15,
|
594 |
-
"eval_accuracy": 0.31469038897610324,
|
595 |
-
"eval_loss": 4.5625,
|
596 |
-
"eval_runtime": 6.5988,
|
597 |
-
"eval_samples_per_second": 5.304,
|
598 |
-
"eval_steps_per_second": 0.303,
|
599 |
-
"step": 750
|
600 |
-
},
|
601 |
-
{
|
602 |
-
"epoch": 0.16,
|
603 |
-
"learning_rate": 0.00017799913382416632,
|
604 |
-
"loss": 3.7347,
|
605 |
-
"step": 760
|
606 |
-
},
|
607 |
-
{
|
608 |
-
"epoch": 0.16,
|
609 |
-
"learning_rate": 0.0001775660459073192,
|
610 |
-
"loss": 3.7917,
|
611 |
-
"step": 770
|
612 |
-
},
|
613 |
-
{
|
614 |
-
"epoch": 0.16,
|
615 |
-
"learning_rate": 0.00017713295799047206,
|
616 |
-
"loss": 3.8106,
|
617 |
-
"step": 780
|
618 |
-
},
|
619 |
-
{
|
620 |
-
"epoch": 0.16,
|
621 |
-
"learning_rate": 0.00017669987007362496,
|
622 |
-
"loss": 3.7289,
|
623 |
-
"step": 790
|
624 |
-
},
|
625 |
-
{
|
626 |
-
"epoch": 0.16,
|
627 |
-
"learning_rate": 0.00017626678215677783,
|
628 |
-
"loss": 3.767,
|
629 |
-
"step": 800
|
630 |
-
},
|
631 |
-
{
|
632 |
-
"epoch": 0.16,
|
633 |
-
"eval_accuracy": 0.3163718820861678,
|
634 |
-
"eval_loss": 4.5390625,
|
635 |
-
"eval_runtime": 6.604,
|
636 |
-
"eval_samples_per_second": 5.3,
|
637 |
-
"eval_steps_per_second": 0.303,
|
638 |
-
"step": 800
|
639 |
-
},
|
640 |
-
{
|
641 |
-
"epoch": 0.17,
|
642 |
-
"learning_rate": 0.00017583369423993073,
|
643 |
-
"loss": 3.7362,
|
644 |
-
"step": 810
|
645 |
-
},
|
646 |
-
{
|
647 |
-
"epoch": 0.17,
|
648 |
-
"learning_rate": 0.0001754006063230836,
|
649 |
-
"loss": 3.7474,
|
650 |
-
"step": 820
|
651 |
-
},
|
652 |
-
{
|
653 |
-
"epoch": 0.17,
|
654 |
-
"learning_rate": 0.00017496751840623647,
|
655 |
-
"loss": 3.7485,
|
656 |
-
"step": 830
|
657 |
-
},
|
658 |
-
{
|
659 |
-
"epoch": 0.17,
|
660 |
-
"learning_rate": 0.00017453443048938937,
|
661 |
-
"loss": 3.7341,
|
662 |
-
"step": 840
|
663 |
-
},
|
664 |
-
{
|
665 |
-
"epoch": 0.17,
|
666 |
-
"learning_rate": 0.00017410134257254224,
|
667 |
-
"loss": 3.7511,
|
668 |
-
"step": 850
|
669 |
-
},
|
670 |
-
{
|
671 |
-
"epoch": 0.17,
|
672 |
-
"eval_accuracy": 0.3172649572649573,
|
673 |
-
"eval_loss": 4.515625,
|
674 |
-
"eval_runtime": 6.6093,
|
675 |
-
"eval_samples_per_second": 5.296,
|
676 |
-
"eval_steps_per_second": 0.303,
|
677 |
-
"step": 850
|
678 |
-
},
|
679 |
-
{
|
680 |
-
"epoch": 0.18,
|
681 |
-
"learning_rate": 0.0001736682546556951,
|
682 |
-
"loss": 3.6862,
|
683 |
-
"step": 860
|
684 |
-
},
|
685 |
-
{
|
686 |
-
"epoch": 0.18,
|
687 |
-
"learning_rate": 0.00017323516673884798,
|
688 |
-
"loss": 3.6411,
|
689 |
-
"step": 870
|
690 |
-
},
|
691 |
-
{
|
692 |
-
"epoch": 0.18,
|
693 |
-
"learning_rate": 0.00017280207882200088,
|
694 |
-
"loss": 3.7181,
|
695 |
-
"step": 880
|
696 |
-
},
|
697 |
-
{
|
698 |
-
"epoch": 0.18,
|
699 |
-
"learning_rate": 0.00017236899090515375,
|
700 |
-
"loss": 3.6471,
|
701 |
-
"step": 890
|
702 |
-
},
|
703 |
-
{
|
704 |
-
"epoch": 0.19,
|
705 |
-
"learning_rate": 0.00017193590298830665,
|
706 |
-
"loss": 3.7166,
|
707 |
-
"step": 900
|
708 |
-
},
|
709 |
-
{
|
710 |
-
"epoch": 0.19,
|
711 |
-
"eval_accuracy": 0.3188278388278388,
|
712 |
-
"eval_loss": 4.4921875,
|
713 |
-
"eval_runtime": 6.6113,
|
714 |
-
"eval_samples_per_second": 5.294,
|
715 |
-
"eval_steps_per_second": 0.303,
|
716 |
-
"step": 900
|
717 |
-
},
|
718 |
-
{
|
719 |
-
"epoch": 0.19,
|
720 |
-
"learning_rate": 0.00017150281507145952,
|
721 |
-
"loss": 3.6869,
|
722 |
-
"step": 910
|
723 |
-
},
|
724 |
-
{
|
725 |
-
"epoch": 0.19,
|
726 |
-
"learning_rate": 0.00017106972715461241,
|
727 |
-
"loss": 3.6728,
|
728 |
-
"step": 920
|
729 |
-
},
|
730 |
-
{
|
731 |
-
"epoch": 0.19,
|
732 |
-
"learning_rate": 0.00017063663923776528,
|
733 |
-
"loss": 3.7705,
|
734 |
-
"step": 930
|
735 |
-
},
|
736 |
-
{
|
737 |
-
"epoch": 0.19,
|
738 |
-
"learning_rate": 0.00017020355132091815,
|
739 |
-
"loss": 3.6728,
|
740 |
-
"step": 940
|
741 |
-
},
|
742 |
-
{
|
743 |
-
"epoch": 0.2,
|
744 |
-
"learning_rate": 0.00016977046340407103,
|
745 |
-
"loss": 3.6908,
|
746 |
-
"step": 950
|
747 |
-
},
|
748 |
-
{
|
749 |
-
"epoch": 0.2,
|
750 |
-
"eval_accuracy": 0.3190441304727019,
|
751 |
-
"eval_loss": 4.48046875,
|
752 |
-
"eval_runtime": 6.5961,
|
753 |
-
"eval_samples_per_second": 5.306,
|
754 |
-
"eval_steps_per_second": 0.303,
|
755 |
-
"step": 950
|
756 |
-
},
|
757 |
-
{
|
758 |
-
"epoch": 0.2,
|
759 |
-
"learning_rate": 0.0001693373754872239,
|
760 |
-
"loss": 3.5988,
|
761 |
-
"step": 960
|
762 |
-
},
|
763 |
-
{
|
764 |
-
"epoch": 0.2,
|
765 |
-
"learning_rate": 0.0001689042875703768,
|
766 |
-
"loss": 3.6713,
|
767 |
-
"step": 970
|
768 |
-
},
|
769 |
-
{
|
770 |
-
"epoch": 0.2,
|
771 |
-
"learning_rate": 0.00016847119965352966,
|
772 |
-
"loss": 3.7165,
|
773 |
-
"step": 980
|
774 |
-
},
|
775 |
-
{
|
776 |
-
"epoch": 0.2,
|
777 |
-
"learning_rate": 0.00016803811173668256,
|
778 |
-
"loss": 3.7098,
|
779 |
-
"step": 990
|
780 |
-
},
|
781 |
-
{
|
782 |
-
"epoch": 0.21,
|
783 |
-
"learning_rate": 0.00016760502381983543,
|
784 |
-
"loss": 3.617,
|
785 |
-
"step": 1000
|
786 |
-
},
|
787 |
-
{
|
788 |
-
"epoch": 0.21,
|
789 |
-
"eval_accuracy": 0.3208442351299494,
|
790 |
-
"eval_loss": 4.46484375,
|
791 |
-
"eval_runtime": 6.5949,
|
792 |
-
"eval_samples_per_second": 5.307,
|
793 |
-
"eval_steps_per_second": 0.303,
|
794 |
-
"step": 1000
|
795 |
-
},
|
796 |
-
{
|
797 |
-
"epoch": 0.21,
|
798 |
-
"learning_rate": 0.00016717193590298833,
|
799 |
-
"loss": 3.6629,
|
800 |
-
"step": 1010
|
801 |
-
},
|
802 |
-
{
|
803 |
-
"epoch": 0.21,
|
804 |
-
"learning_rate": 0.0001667388479861412,
|
805 |
-
"loss": 3.684,
|
806 |
-
"step": 1020
|
807 |
-
},
|
808 |
-
{
|
809 |
-
"epoch": 0.21,
|
810 |
-
"learning_rate": 0.00016630576006929407,
|
811 |
-
"loss": 3.6877,
|
812 |
-
"step": 1030
|
813 |
-
},
|
814 |
-
{
|
815 |
-
"epoch": 0.21,
|
816 |
-
"learning_rate": 0.00016587267215244694,
|
817 |
-
"loss": 3.6274,
|
818 |
-
"step": 1040
|
819 |
-
},
|
820 |
-
{
|
821 |
-
"epoch": 0.22,
|
822 |
-
"learning_rate": 0.00016543958423559984,
|
823 |
-
"loss": 3.6734,
|
824 |
-
"step": 1050
|
825 |
-
},
|
826 |
-
{
|
827 |
-
"epoch": 0.22,
|
828 |
-
"eval_accuracy": 0.3213535670678528,
|
829 |
-
"eval_loss": 4.453125,
|
830 |
-
"eval_runtime": 6.5974,
|
831 |
-
"eval_samples_per_second": 5.305,
|
832 |
-
"eval_steps_per_second": 0.303,
|
833 |
-
"step": 1050
|
834 |
-
},
|
835 |
-
{
|
836 |
-
"epoch": 0.22,
|
837 |
-
"learning_rate": 0.0001650064963187527,
|
838 |
-
"loss": 3.6895,
|
839 |
-
"step": 1060
|
840 |
-
},
|
841 |
-
{
|
842 |
-
"epoch": 0.22,
|
843 |
-
"learning_rate": 0.0001645734084019056,
|
844 |
-
"loss": 3.6662,
|
845 |
-
"step": 1070
|
846 |
-
},
|
847 |
-
{
|
848 |
-
"epoch": 0.22,
|
849 |
-
"learning_rate": 0.00016414032048505848,
|
850 |
-
"loss": 3.6152,
|
851 |
-
"step": 1080
|
852 |
-
},
|
853 |
-
{
|
854 |
-
"epoch": 0.22,
|
855 |
-
"learning_rate": 0.00016370723256821135,
|
856 |
-
"loss": 3.6531,
|
857 |
-
"step": 1090
|
858 |
-
},
|
859 |
-
{
|
860 |
-
"epoch": 0.23,
|
861 |
-
"learning_rate": 0.00016327414465136425,
|
862 |
-
"loss": 3.6916,
|
863 |
-
"step": 1100
|
864 |
-
},
|
865 |
-
{
|
866 |
-
"epoch": 0.23,
|
867 |
-
"eval_accuracy": 0.32398395255538115,
|
868 |
-
"eval_loss": 4.43359375,
|
869 |
-
"eval_runtime": 6.5897,
|
870 |
-
"eval_samples_per_second": 5.311,
|
871 |
-
"eval_steps_per_second": 0.304,
|
872 |
-
"step": 1100
|
873 |
-
},
|
874 |
-
{
|
875 |
-
"epoch": 0.23,
|
876 |
-
"learning_rate": 0.00016284105673451712,
|
877 |
-
"loss": 3.6961,
|
878 |
-
"step": 1110
|
879 |
-
},
|
880 |
-
{
|
881 |
-
"epoch": 0.23,
|
882 |
-
"learning_rate": 0.00016240796881767,
|
883 |
-
"loss": 3.5811,
|
884 |
-
"step": 1120
|
885 |
-
},
|
886 |
-
{
|
887 |
-
"epoch": 0.23,
|
888 |
-
"learning_rate": 0.00016197488090082286,
|
889 |
-
"loss": 3.6161,
|
890 |
-
"step": 1130
|
891 |
-
},
|
892 |
-
{
|
893 |
-
"epoch": 0.23,
|
894 |
-
"learning_rate": 0.00016154179298397576,
|
895 |
-
"loss": 3.6305,
|
896 |
-
"step": 1140
|
897 |
-
},
|
898 |
-
{
|
899 |
-
"epoch": 0.24,
|
900 |
-
"learning_rate": 0.00016110870506712863,
|
901 |
-
"loss": 3.629,
|
902 |
-
"step": 1150
|
903 |
-
},
|
904 |
-
{
|
905 |
-
"epoch": 0.24,
|
906 |
-
"eval_accuracy": 0.32400488400488403,
|
907 |
-
"eval_loss": 4.421875,
|
908 |
-
"eval_runtime": 6.5985,
|
909 |
-
"eval_samples_per_second": 5.304,
|
910 |
-
"eval_steps_per_second": 0.303,
|
911 |
-
"step": 1150
|
912 |
-
},
|
913 |
-
{
|
914 |
-
"epoch": 0.24,
|
915 |
-
"learning_rate": 0.00016067561715028152,
|
916 |
-
"loss": 3.6467,
|
917 |
-
"step": 1160
|
918 |
-
},
|
919 |
-
{
|
920 |
-
"epoch": 0.24,
|
921 |
-
"learning_rate": 0.0001602425292334344,
|
922 |
-
"loss": 3.6573,
|
923 |
-
"step": 1170
|
924 |
-
},
|
925 |
-
{
|
926 |
-
"epoch": 0.24,
|
927 |
-
"learning_rate": 0.0001598094413165873,
|
928 |
-
"loss": 3.6372,
|
929 |
-
"step": 1180
|
930 |
-
},
|
931 |
-
{
|
932 |
-
"epoch": 0.24,
|
933 |
-
"learning_rate": 0.00015937635339974016,
|
934 |
-
"loss": 3.6369,
|
935 |
-
"step": 1190
|
936 |
-
},
|
937 |
-
{
|
938 |
-
"epoch": 0.25,
|
939 |
-
"learning_rate": 0.00015894326548289303,
|
940 |
-
"loss": 3.6001,
|
941 |
-
"step": 1200
|
942 |
-
},
|
943 |
-
{
|
944 |
-
"epoch": 0.25,
|
945 |
-
"eval_accuracy": 0.3247165532879819,
|
946 |
-
"eval_loss": 4.4140625,
|
947 |
-
"eval_runtime": 6.602,
|
948 |
-
"eval_samples_per_second": 5.301,
|
949 |
-
"eval_steps_per_second": 0.303,
|
950 |
-
"step": 1200
|
951 |
-
},
|
952 |
-
{
|
953 |
-
"epoch": 0.25,
|
954 |
-
"learning_rate": 0.0001585101775660459,
|
955 |
-
"loss": 3.5843,
|
956 |
-
"step": 1210
|
957 |
-
},
|
958 |
-
{
|
959 |
-
"epoch": 0.25,
|
960 |
-
"learning_rate": 0.00015807708964919877,
|
961 |
-
"loss": 3.6407,
|
962 |
-
"step": 1220
|
963 |
-
},
|
964 |
-
{
|
965 |
-
"epoch": 0.25,
|
966 |
-
"learning_rate": 0.00015764400173235167,
|
967 |
-
"loss": 3.6413,
|
968 |
-
"step": 1230
|
969 |
-
},
|
970 |
-
{
|
971 |
-
"epoch": 0.26,
|
972 |
-
"learning_rate": 0.00015721091381550454,
|
973 |
-
"loss": 3.5963,
|
974 |
-
"step": 1240
|
975 |
-
},
|
976 |
-
{
|
977 |
-
"epoch": 0.26,
|
978 |
-
"learning_rate": 0.00015677782589865744,
|
979 |
-
"loss": 3.6053,
|
980 |
-
"step": 1250
|
981 |
-
},
|
982 |
-
{
|
983 |
-
"epoch": 0.26,
|
984 |
-
"eval_accuracy": 0.32607709750566893,
|
985 |
-
"eval_loss": 4.40234375,
|
986 |
-
"eval_runtime": 6.5982,
|
987 |
-
"eval_samples_per_second": 5.304,
|
988 |
-
"eval_steps_per_second": 0.303,
|
989 |
-
"step": 1250
|
990 |
-
},
|
991 |
-
{
|
992 |
-
"epoch": 0.26,
|
993 |
-
"learning_rate": 0.0001563447379818103,
|
994 |
-
"loss": 3.6049,
|
995 |
-
"step": 1260
|
996 |
-
},
|
997 |
-
{
|
998 |
-
"epoch": 0.26,
|
999 |
-
"learning_rate": 0.0001559116500649632,
|
1000 |
-
"loss": 3.6112,
|
1001 |
-
"step": 1270
|
1002 |
-
},
|
1003 |
-
{
|
1004 |
-
"epoch": 0.26,
|
1005 |
-
"learning_rate": 0.00015547856214811608,
|
1006 |
-
"loss": 3.5872,
|
1007 |
-
"step": 1280
|
1008 |
-
},
|
1009 |
-
{
|
1010 |
-
"epoch": 0.27,
|
1011 |
-
"learning_rate": 0.00015504547423126895,
|
1012 |
-
"loss": 3.6328,
|
1013 |
-
"step": 1290
|
1014 |
-
},
|
1015 |
-
{
|
1016 |
-
"epoch": 0.27,
|
1017 |
-
"learning_rate": 0.00015461238631442182,
|
1018 |
-
"loss": 3.5803,
|
1019 |
-
"step": 1300
|
1020 |
-
},
|
1021 |
-
{
|
1022 |
-
"epoch": 0.27,
|
1023 |
-
"eval_accuracy": 0.32540729112157685,
|
1024 |
-
"eval_loss": 4.390625,
|
1025 |
-
"eval_runtime": 6.5995,
|
1026 |
-
"eval_samples_per_second": 5.303,
|
1027 |
-
"eval_steps_per_second": 0.303,
|
1028 |
-
"step": 1300
|
1029 |
-
},
|
1030 |
-
{
|
1031 |
-
"epoch": 0.27,
|
1032 |
-
"learning_rate": 0.00015417929839757472,
|
1033 |
-
"loss": 3.6153,
|
1034 |
-
"step": 1310
|
1035 |
-
},
|
1036 |
-
{
|
1037 |
-
"epoch": 0.27,
|
1038 |
-
"learning_rate": 0.0001537462104807276,
|
1039 |
-
"loss": 3.5493,
|
1040 |
-
"step": 1320
|
1041 |
-
},
|
1042 |
-
{
|
1043 |
-
"epoch": 0.27,
|
1044 |
-
"learning_rate": 0.00015331312256388049,
|
1045 |
-
"loss": 3.5854,
|
1046 |
-
"step": 1330
|
1047 |
-
},
|
1048 |
-
{
|
1049 |
-
"epoch": 0.28,
|
1050 |
-
"learning_rate": 0.00015288003464703336,
|
1051 |
-
"loss": 3.6029,
|
1052 |
-
"step": 1340
|
1053 |
-
},
|
1054 |
-
{
|
1055 |
-
"epoch": 0.28,
|
1056 |
-
"learning_rate": 0.00015244694673018623,
|
1057 |
-
"loss": 3.5886,
|
1058 |
-
"step": 1350
|
1059 |
-
},
|
1060 |
-
{
|
1061 |
-
"epoch": 0.28,
|
1062 |
-
"eval_accuracy": 0.32738880167451595,
|
1063 |
-
"eval_loss": 4.37890625,
|
1064 |
-
"eval_runtime": 6.5931,
|
1065 |
-
"eval_samples_per_second": 5.309,
|
1066 |
-
"eval_steps_per_second": 0.303,
|
1067 |
-
"step": 1350
|
1068 |
-
},
|
1069 |
-
{
|
1070 |
-
"epoch": 0.28,
|
1071 |
-
"learning_rate": 0.00015201385881333913,
|
1072 |
-
"loss": 3.5496,
|
1073 |
-
"step": 1360
|
1074 |
-
},
|
1075 |
-
{
|
1076 |
-
"epoch": 0.28,
|
1077 |
-
"learning_rate": 0.000151580770896492,
|
1078 |
-
"loss": 3.557,
|
1079 |
-
"step": 1370
|
1080 |
-
},
|
1081 |
-
{
|
1082 |
-
"epoch": 0.28,
|
1083 |
-
"learning_rate": 0.00015114768297964487,
|
1084 |
-
"loss": 3.5647,
|
1085 |
-
"step": 1380
|
1086 |
-
},
|
1087 |
-
{
|
1088 |
-
"epoch": 0.29,
|
1089 |
-
"learning_rate": 0.00015071459506279774,
|
1090 |
-
"loss": 3.5912,
|
1091 |
-
"step": 1390
|
1092 |
-
},
|
1093 |
-
{
|
1094 |
-
"epoch": 0.29,
|
1095 |
-
"learning_rate": 0.00015028150714595063,
|
1096 |
-
"loss": 3.5033,
|
1097 |
-
"step": 1400
|
1098 |
-
},
|
1099 |
-
{
|
1100 |
-
"epoch": 0.29,
|
1101 |
-
"eval_accuracy": 0.3287981859410431,
|
1102 |
-
"eval_loss": 4.3671875,
|
1103 |
-
"eval_runtime": 6.593,
|
1104 |
-
"eval_samples_per_second": 5.309,
|
1105 |
-
"eval_steps_per_second": 0.303,
|
1106 |
-
"step": 1400
|
1107 |
-
},
|
1108 |
-
{
|
1109 |
-
"epoch": 0.29,
|
1110 |
-
"learning_rate": 0.0001498484192291035,
|
1111 |
-
"loss": 3.5814,
|
1112 |
-
"step": 1410
|
1113 |
-
},
|
1114 |
-
{
|
1115 |
-
"epoch": 0.29,
|
1116 |
-
"learning_rate": 0.0001494153313122564,
|
1117 |
-
"loss": 3.5834,
|
1118 |
-
"step": 1420
|
1119 |
-
},
|
1120 |
-
{
|
1121 |
-
"epoch": 0.29,
|
1122 |
-
"learning_rate": 0.00014898224339540927,
|
1123 |
-
"loss": 3.5661,
|
1124 |
-
"step": 1430
|
1125 |
-
},
|
1126 |
-
{
|
1127 |
-
"epoch": 0.3,
|
1128 |
-
"learning_rate": 0.00014854915547856217,
|
1129 |
-
"loss": 3.5844,
|
1130 |
-
"step": 1440
|
1131 |
-
},
|
1132 |
-
{
|
1133 |
-
"epoch": 0.3,
|
1134 |
-
"learning_rate": 0.00014811606756171504,
|
1135 |
-
"loss": 3.58,
|
1136 |
-
"step": 1450
|
1137 |
-
},
|
1138 |
-
{
|
1139 |
-
"epoch": 0.3,
|
1140 |
-
"eval_accuracy": 0.3283656026513169,
|
1141 |
-
"eval_loss": 4.36328125,
|
1142 |
-
"eval_runtime": 6.6004,
|
1143 |
-
"eval_samples_per_second": 5.303,
|
1144 |
-
"eval_steps_per_second": 0.303,
|
1145 |
-
"step": 1450
|
1146 |
-
},
|
1147 |
-
{
|
1148 |
-
"epoch": 0.3,
|
1149 |
-
"learning_rate": 0.0001476829796448679,
|
1150 |
-
"loss": 3.5874,
|
1151 |
-
"step": 1460
|
1152 |
-
},
|
1153 |
-
{
|
1154 |
-
"epoch": 0.3,
|
1155 |
-
"learning_rate": 0.00014724989172802078,
|
1156 |
-
"loss": 3.5736,
|
1157 |
-
"step": 1470
|
1158 |
-
},
|
1159 |
-
{
|
1160 |
-
"epoch": 0.3,
|
1161 |
-
"learning_rate": 0.00014681680381117365,
|
1162 |
-
"loss": 3.5659,
|
1163 |
-
"step": 1480
|
1164 |
-
},
|
1165 |
-
{
|
1166 |
-
"epoch": 0.31,
|
1167 |
-
"learning_rate": 0.00014638371589432655,
|
1168 |
-
"loss": 3.5632,
|
1169 |
-
"step": 1490
|
1170 |
-
},
|
1171 |
-
{
|
1172 |
-
"epoch": 0.31,
|
1173 |
-
"learning_rate": 0.00014595062797747942,
|
1174 |
-
"loss": 3.4966,
|
1175 |
-
"step": 1500
|
1176 |
-
},
|
1177 |
-
{
|
1178 |
-
"epoch": 0.31,
|
1179 |
-
"eval_accuracy": 0.32832373975231116,
|
1180 |
-
"eval_loss": 4.3515625,
|
1181 |
-
"eval_runtime": 6.592,
|
1182 |
-
"eval_samples_per_second": 5.309,
|
1183 |
-
"eval_steps_per_second": 0.303,
|
1184 |
-
"step": 1500
|
1185 |
-
},
|
1186 |
-
{
|
1187 |
-
"epoch": 0.31,
|
1188 |
-
"learning_rate": 0.00014551754006063232,
|
1189 |
-
"loss": 3.5236,
|
1190 |
-
"step": 1510
|
1191 |
-
},
|
1192 |
-
{
|
1193 |
-
"epoch": 0.31,
|
1194 |
-
"learning_rate": 0.0001450844521437852,
|
1195 |
-
"loss": 3.5277,
|
1196 |
-
"step": 1520
|
1197 |
-
},
|
1198 |
-
{
|
1199 |
-
"epoch": 0.31,
|
1200 |
-
"learning_rate": 0.0001446513642269381,
|
1201 |
-
"loss": 3.5237,
|
1202 |
-
"step": 1530
|
1203 |
-
},
|
1204 |
-
{
|
1205 |
-
"epoch": 0.32,
|
1206 |
-
"learning_rate": 0.00014421827631009096,
|
1207 |
-
"loss": 3.5719,
|
1208 |
-
"step": 1540
|
1209 |
-
},
|
1210 |
-
{
|
1211 |
-
"epoch": 0.32,
|
1212 |
-
"learning_rate": 0.00014378518839324383,
|
1213 |
-
"loss": 3.5411,
|
1214 |
-
"step": 1550
|
1215 |
-
},
|
1216 |
-
{
|
1217 |
-
"epoch": 0.32,
|
1218 |
-
"eval_accuracy": 0.32884004884004886,
|
1219 |
-
"eval_loss": 4.3515625,
|
1220 |
-
"eval_runtime": 6.6086,
|
1221 |
-
"eval_samples_per_second": 5.296,
|
1222 |
-
"eval_steps_per_second": 0.303,
|
1223 |
-
"step": 1550
|
1224 |
-
},
|
1225 |
-
{
|
1226 |
-
"epoch": 0.32,
|
1227 |
-
"learning_rate": 0.0001433521004763967,
|
1228 |
-
"loss": 3.5287,
|
1229 |
-
"step": 1560
|
1230 |
-
},
|
1231 |
-
{
|
1232 |
-
"epoch": 0.32,
|
1233 |
-
"learning_rate": 0.0001429190125595496,
|
1234 |
-
"loss": 3.5965,
|
1235 |
-
"step": 1570
|
1236 |
-
},
|
1237 |
-
{
|
1238 |
-
"epoch": 0.32,
|
1239 |
-
"learning_rate": 0.00014248592464270247,
|
1240 |
-
"loss": 3.5435,
|
1241 |
-
"step": 1580
|
1242 |
-
},
|
1243 |
-
{
|
1244 |
-
"epoch": 0.33,
|
1245 |
-
"learning_rate": 0.00014205283672585536,
|
1246 |
-
"loss": 3.5536,
|
1247 |
-
"step": 1590
|
1248 |
-
},
|
1249 |
-
{
|
1250 |
-
"epoch": 0.33,
|
1251 |
-
"learning_rate": 0.00014161974880900824,
|
1252 |
-
"loss": 3.527,
|
1253 |
-
"step": 1600
|
1254 |
-
},
|
1255 |
-
{
|
1256 |
-
"epoch": 0.33,
|
1257 |
-
"eval_accuracy": 0.33027036455607883,
|
1258 |
-
"eval_loss": 4.33984375,
|
1259 |
-
"eval_runtime": 6.5917,
|
1260 |
-
"eval_samples_per_second": 5.31,
|
1261 |
-
"eval_steps_per_second": 0.303,
|
1262 |
-
"step": 1600
|
1263 |
-
},
|
1264 |
-
{
|
1265 |
-
"epoch": 0.33,
|
1266 |
-
"learning_rate": 0.0001411866608921611,
|
1267 |
-
"loss": 3.5765,
|
1268 |
-
"step": 1610
|
1269 |
-
},
|
1270 |
-
{
|
1271 |
-
"epoch": 0.33,
|
1272 |
-
"learning_rate": 0.000140753572975314,
|
1273 |
-
"loss": 3.5882,
|
1274 |
-
"step": 1620
|
1275 |
-
},
|
1276 |
-
{
|
1277 |
-
"epoch": 0.34,
|
1278 |
-
"learning_rate": 0.00014032048505846687,
|
1279 |
-
"loss": 3.5135,
|
1280 |
-
"step": 1630
|
1281 |
-
},
|
1282 |
-
{
|
1283 |
-
"epoch": 0.34,
|
1284 |
-
"learning_rate": 0.00013988739714161974,
|
1285 |
-
"loss": 3.4924,
|
1286 |
-
"step": 1640
|
1287 |
-
},
|
1288 |
-
{
|
1289 |
-
"epoch": 0.34,
|
1290 |
-
"learning_rate": 0.00013945430922477262,
|
1291 |
-
"loss": 3.6018,
|
1292 |
-
"step": 1650
|
1293 |
-
},
|
1294 |
-
{
|
1295 |
-
"epoch": 0.34,
|
1296 |
-
"eval_accuracy": 0.3299284842141985,
|
1297 |
-
"eval_loss": 4.33203125,
|
1298 |
-
"eval_runtime": 6.5992,
|
1299 |
-
"eval_samples_per_second": 5.304,
|
1300 |
-
"eval_steps_per_second": 0.303,
|
1301 |
-
"step": 1650
|
1302 |
-
},
|
1303 |
-
{
|
1304 |
-
"epoch": 0.34,
|
1305 |
-
"learning_rate": 0.0001390212213079255,
|
1306 |
-
"loss": 3.5646,
|
1307 |
-
"step": 1660
|
1308 |
-
},
|
1309 |
-
{
|
1310 |
-
"epoch": 0.34,
|
1311 |
-
"learning_rate": 0.00013858813339107838,
|
1312 |
-
"loss": 3.5164,
|
1313 |
-
"step": 1670
|
1314 |
-
},
|
1315 |
-
{
|
1316 |
-
"epoch": 0.35,
|
1317 |
-
"learning_rate": 0.00013815504547423128,
|
1318 |
-
"loss": 3.5433,
|
1319 |
-
"step": 1680
|
1320 |
-
},
|
1321 |
-
{
|
1322 |
-
"epoch": 0.35,
|
1323 |
-
"learning_rate": 0.00013772195755738415,
|
1324 |
-
"loss": 3.4929,
|
1325 |
-
"step": 1690
|
1326 |
-
},
|
1327 |
-
{
|
1328 |
-
"epoch": 0.35,
|
1329 |
-
"learning_rate": 0.00013728886964053705,
|
1330 |
-
"loss": 3.4802,
|
1331 |
-
"step": 1700
|
1332 |
-
},
|
1333 |
-
{
|
1334 |
-
"epoch": 0.35,
|
1335 |
-
"eval_accuracy": 0.3301238444095587,
|
1336 |
-
"eval_loss": 4.32421875,
|
1337 |
-
"eval_runtime": 6.5899,
|
1338 |
-
"eval_samples_per_second": 5.311,
|
1339 |
-
"eval_steps_per_second": 0.303,
|
1340 |
-
"step": 1700
|
1341 |
-
},
|
1342 |
-
{
|
1343 |
-
"epoch": 0.35,
|
1344 |
-
"learning_rate": 0.00013685578172368992,
|
1345 |
-
"loss": 3.5211,
|
1346 |
-
"step": 1710
|
1347 |
-
},
|
1348 |
-
{
|
1349 |
-
"epoch": 0.35,
|
1350 |
-
"learning_rate": 0.00013642269380684282,
|
1351 |
-
"loss": 3.5424,
|
1352 |
-
"step": 1720
|
1353 |
-
},
|
1354 |
-
{
|
1355 |
-
"epoch": 0.36,
|
1356 |
-
"learning_rate": 0.00013598960588999566,
|
1357 |
-
"loss": 3.468,
|
1358 |
-
"step": 1730
|
1359 |
-
},
|
1360 |
-
{
|
1361 |
-
"epoch": 0.36,
|
1362 |
-
"learning_rate": 0.00013555651797314853,
|
1363 |
-
"loss": 3.5342,
|
1364 |
-
"step": 1740
|
1365 |
-
},
|
1366 |
-
{
|
1367 |
-
"epoch": 0.36,
|
1368 |
-
"learning_rate": 0.00013512343005630143,
|
1369 |
-
"loss": 3.4375,
|
1370 |
-
"step": 1750
|
1371 |
-
},
|
1372 |
-
{
|
1373 |
-
"epoch": 0.36,
|
1374 |
-
"eval_accuracy": 0.330584336298622,
|
1375 |
-
"eval_loss": 4.32421875,
|
1376 |
-
"eval_runtime": 6.5928,
|
1377 |
-
"eval_samples_per_second": 5.309,
|
1378 |
-
"eval_steps_per_second": 0.303,
|
1379 |
-
"step": 1750
|
1380 |
-
},
|
1381 |
-
{
|
1382 |
-
"epoch": 0.36,
|
1383 |
-
"learning_rate": 0.0001346903421394543,
|
1384 |
-
"loss": 3.5239,
|
1385 |
-
"step": 1760
|
1386 |
-
},
|
1387 |
-
{
|
1388 |
-
"epoch": 0.36,
|
1389 |
-
"learning_rate": 0.0001342572542226072,
|
1390 |
-
"loss": 3.5176,
|
1391 |
-
"step": 1770
|
1392 |
-
},
|
1393 |
-
{
|
1394 |
-
"epoch": 0.37,
|
1395 |
-
"learning_rate": 0.00013382416630576007,
|
1396 |
-
"loss": 3.4992,
|
1397 |
-
"step": 1780
|
1398 |
-
},
|
1399 |
-
{
|
1400 |
-
"epoch": 0.37,
|
1401 |
-
"learning_rate": 0.00013339107838891297,
|
1402 |
-
"loss": 3.457,
|
1403 |
-
"step": 1790
|
1404 |
-
},
|
1405 |
-
{
|
1406 |
-
"epoch": 0.37,
|
1407 |
-
"learning_rate": 0.00013295799047206584,
|
1408 |
-
"loss": 3.4873,
|
1409 |
-
"step": 1800
|
1410 |
-
},
|
1411 |
-
{
|
1412 |
-
"epoch": 0.37,
|
1413 |
-
"eval_accuracy": 0.33109366823652536,
|
1414 |
-
"eval_loss": 4.3203125,
|
1415 |
-
"eval_runtime": 6.6125,
|
1416 |
-
"eval_samples_per_second": 5.293,
|
1417 |
-
"eval_steps_per_second": 0.302,
|
1418 |
-
"step": 1800
|
1419 |
-
},
|
1420 |
-
{
|
1421 |
-
"epoch": 0.37,
|
1422 |
-
"learning_rate": 0.00013252490255521873,
|
1423 |
-
"loss": 3.5021,
|
1424 |
-
"step": 1810
|
1425 |
-
},
|
1426 |
-
{
|
1427 |
-
"epoch": 0.37,
|
1428 |
-
"learning_rate": 0.0001320918146383716,
|
1429 |
-
"loss": 3.4871,
|
1430 |
-
"step": 1820
|
1431 |
-
},
|
1432 |
-
{
|
1433 |
-
"epoch": 0.38,
|
1434 |
-
"learning_rate": 0.00013165872672152448,
|
1435 |
-
"loss": 3.47,
|
1436 |
-
"step": 1830
|
1437 |
-
},
|
1438 |
-
{
|
1439 |
-
"epoch": 0.38,
|
1440 |
-
"learning_rate": 0.00013122563880467735,
|
1441 |
-
"loss": 3.4462,
|
1442 |
-
"step": 1840
|
1443 |
-
},
|
1444 |
-
{
|
1445 |
-
"epoch": 0.38,
|
1446 |
-
"learning_rate": 0.00013079255088783024,
|
1447 |
-
"loss": 3.435,
|
1448 |
-
"step": 1850
|
1449 |
-
},
|
1450 |
-
{
|
1451 |
-
"epoch": 0.38,
|
1452 |
-
"eval_accuracy": 0.3309401709401709,
|
1453 |
-
"eval_loss": 4.3125,
|
1454 |
-
"eval_runtime": 6.5916,
|
1455 |
-
"eval_samples_per_second": 5.31,
|
1456 |
-
"eval_steps_per_second": 0.303,
|
1457 |
-
"step": 1850
|
1458 |
-
},
|
1459 |
-
{
|
1460 |
-
"epoch": 0.38,
|
1461 |
-
"learning_rate": 0.00013035946297098311,
|
1462 |
-
"loss": 3.4994,
|
1463 |
-
"step": 1860
|
1464 |
-
},
|
1465 |
-
{
|
1466 |
-
"epoch": 0.38,
|
1467 |
-
"learning_rate": 0.00012992637505413598,
|
1468 |
-
"loss": 3.5665,
|
1469 |
-
"step": 1870
|
1470 |
-
},
|
1471 |
-
{
|
1472 |
-
"epoch": 0.39,
|
1473 |
-
"learning_rate": 0.00012949328713728888,
|
1474 |
-
"loss": 3.5552,
|
1475 |
-
"step": 1880
|
1476 |
-
},
|
1477 |
-
{
|
1478 |
-
"epoch": 0.39,
|
1479 |
-
"learning_rate": 0.00012906019922044175,
|
1480 |
-
"loss": 3.5092,
|
1481 |
-
"step": 1890
|
1482 |
-
},
|
1483 |
-
{
|
1484 |
-
"epoch": 0.39,
|
1485 |
-
"learning_rate": 0.00012862711130359465,
|
1486 |
-
"loss": 3.4335,
|
1487 |
-
"step": 1900
|
1488 |
-
},
|
1489 |
-
{
|
1490 |
-
"epoch": 0.39,
|
1491 |
-
"eval_accuracy": 0.3317634746206175,
|
1492 |
-
"eval_loss": 4.3046875,
|
1493 |
-
"eval_runtime": 6.6183,
|
1494 |
-
"eval_samples_per_second": 5.288,
|
1495 |
-
"eval_steps_per_second": 0.302,
|
1496 |
-
"step": 1900
|
1497 |
-
},
|
1498 |
-
{
|
1499 |
-
"epoch": 0.39,
|
1500 |
-
"learning_rate": 0.00012819402338674752,
|
1501 |
-
"loss": 3.5154,
|
1502 |
-
"step": 1910
|
1503 |
-
},
|
1504 |
-
{
|
1505 |
-
"epoch": 0.39,
|
1506 |
-
"learning_rate": 0.0001277609354699004,
|
1507 |
-
"loss": 3.5207,
|
1508 |
-
"step": 1920
|
1509 |
-
},
|
1510 |
-
{
|
1511 |
-
"epoch": 0.4,
|
1512 |
-
"learning_rate": 0.00012732784755305326,
|
1513 |
-
"loss": 3.4869,
|
1514 |
-
"step": 1930
|
1515 |
-
},
|
1516 |
-
{
|
1517 |
-
"epoch": 0.4,
|
1518 |
-
"learning_rate": 0.00012689475963620616,
|
1519 |
-
"loss": 3.4773,
|
1520 |
-
"step": 1940
|
1521 |
-
},
|
1522 |
-
{
|
1523 |
-
"epoch": 0.4,
|
1524 |
-
"learning_rate": 0.00012646167171935903,
|
1525 |
-
"loss": 3.4595,
|
1526 |
-
"step": 1950
|
1527 |
-
},
|
1528 |
-
{
|
1529 |
-
"epoch": 0.4,
|
1530 |
-
"eval_accuracy": 0.3325239839525554,
|
1531 |
-
"eval_loss": 4.296875,
|
1532 |
-
"eval_runtime": 6.5968,
|
1533 |
-
"eval_samples_per_second": 5.306,
|
1534 |
-
"eval_steps_per_second": 0.303,
|
1535 |
-
"step": 1950
|
1536 |
-
},
|
1537 |
-
{
|
1538 |
-
"epoch": 0.4,
|
1539 |
-
"learning_rate": 0.00012602858380251193,
|
1540 |
-
"loss": 3.5035,
|
1541 |
-
"step": 1960
|
1542 |
-
},
|
1543 |
-
{
|
1544 |
-
"epoch": 0.41,
|
1545 |
-
"learning_rate": 0.0001255954958856648,
|
1546 |
-
"loss": 3.4692,
|
1547 |
-
"step": 1970
|
1548 |
-
},
|
1549 |
-
{
|
1550 |
-
"epoch": 0.41,
|
1551 |
-
"learning_rate": 0.0001251624079688177,
|
1552 |
-
"loss": 3.4712,
|
1553 |
-
"step": 1980
|
1554 |
-
},
|
1555 |
-
{
|
1556 |
-
"epoch": 0.41,
|
1557 |
-
"learning_rate": 0.00012472932005197057,
|
1558 |
-
"loss": 3.4558,
|
1559 |
-
"step": 1990
|
1560 |
-
},
|
1561 |
-
{
|
1562 |
-
"epoch": 0.41,
|
1563 |
-
"learning_rate": 0.00012429623213512344,
|
1564 |
-
"loss": 3.4937,
|
1565 |
-
"step": 2000
|
1566 |
-
},
|
1567 |
-
{
|
1568 |
-
"epoch": 0.41,
|
1569 |
-
"eval_accuracy": 0.3319448805163091,
|
1570 |
-
"eval_loss": 4.29296875,
|
1571 |
-
"eval_runtime": 6.5908,
|
1572 |
-
"eval_samples_per_second": 5.31,
|
1573 |
-
"eval_steps_per_second": 0.303,
|
1574 |
-
"step": 2000
|
1575 |
-
},
|
1576 |
-
{
|
1577 |
-
"epoch": 0.41,
|
1578 |
-
"learning_rate": 0.0001238631442182763,
|
1579 |
-
"loss": 3.5093,
|
1580 |
-
"step": 2010
|
1581 |
-
},
|
1582 |
-
{
|
1583 |
-
"epoch": 0.42,
|
1584 |
-
"learning_rate": 0.00012343005630142918,
|
1585 |
-
"loss": 3.4941,
|
1586 |
-
"step": 2020
|
1587 |
-
},
|
1588 |
-
{
|
1589 |
-
"epoch": 0.42,
|
1590 |
-
"learning_rate": 0.00012299696838458208,
|
1591 |
-
"loss": 3.4706,
|
1592 |
-
"step": 2030
|
1593 |
-
},
|
1594 |
-
{
|
1595 |
-
"epoch": 0.42,
|
1596 |
-
"learning_rate": 0.00012256388046773495,
|
1597 |
-
"loss": 3.5106,
|
1598 |
-
"step": 2040
|
1599 |
-
},
|
1600 |
-
{
|
1601 |
-
"epoch": 0.42,
|
1602 |
-
"learning_rate": 0.00012213079255088784,
|
1603 |
-
"loss": 3.4959,
|
1604 |
-
"step": 2050
|
1605 |
-
},
|
1606 |
-
{
|
1607 |
-
"epoch": 0.42,
|
1608 |
-
"eval_accuracy": 0.3324123495552067,
|
1609 |
-
"eval_loss": 4.28515625,
|
1610 |
-
"eval_runtime": 6.6003,
|
1611 |
-
"eval_samples_per_second": 5.303,
|
1612 |
-
"eval_steps_per_second": 0.303,
|
1613 |
-
"step": 2050
|
1614 |
-
},
|
1615 |
-
{
|
1616 |
-
"epoch": 0.42,
|
1617 |
-
"learning_rate": 0.00012169770463404072,
|
1618 |
-
"loss": 3.5218,
|
1619 |
-
"step": 2060
|
1620 |
-
},
|
1621 |
-
{
|
1622 |
-
"epoch": 0.43,
|
1623 |
-
"learning_rate": 0.0001212646167171936,
|
1624 |
-
"loss": 3.4813,
|
1625 |
-
"step": 2070
|
1626 |
-
},
|
1627 |
-
{
|
1628 |
-
"epoch": 0.43,
|
1629 |
-
"learning_rate": 0.00012083152880034647,
|
1630 |
-
"loss": 3.5107,
|
1631 |
-
"step": 2080
|
1632 |
-
},
|
1633 |
-
{
|
1634 |
-
"epoch": 0.43,
|
1635 |
-
"learning_rate": 0.00012039844088349937,
|
1636 |
-
"loss": 3.4568,
|
1637 |
-
"step": 2090
|
1638 |
-
},
|
1639 |
-
{
|
1640 |
-
"epoch": 0.43,
|
1641 |
-
"learning_rate": 0.00011996535296665224,
|
1642 |
-
"loss": 3.4987,
|
1643 |
-
"step": 2100
|
1644 |
-
},
|
1645 |
-
{
|
1646 |
-
"epoch": 0.43,
|
1647 |
-
"eval_accuracy": 0.3331658817373103,
|
1648 |
-
"eval_loss": 4.28515625,
|
1649 |
-
"eval_runtime": 6.6107,
|
1650 |
-
"eval_samples_per_second": 5.294,
|
1651 |
-
"eval_steps_per_second": 0.303,
|
1652 |
-
"step": 2100
|
1653 |
-
},
|
1654 |
-
{
|
1655 |
-
"epoch": 0.43,
|
1656 |
-
"learning_rate": 0.00011957557384148984,
|
1657 |
-
"loss": 3.458,
|
1658 |
-
"step": 2110
|
1659 |
-
},
|
1660 |
-
{
|
1661 |
-
"epoch": 0.44,
|
1662 |
-
"learning_rate": 0.0001191424859246427,
|
1663 |
-
"loss": 3.4656,
|
1664 |
-
"step": 2120
|
1665 |
-
},
|
1666 |
-
{
|
1667 |
-
"epoch": 0.44,
|
1668 |
-
"learning_rate": 0.00011870939800779559,
|
1669 |
-
"loss": 3.4505,
|
1670 |
-
"step": 2130
|
1671 |
-
},
|
1672 |
-
{
|
1673 |
-
"epoch": 0.44,
|
1674 |
-
"learning_rate": 0.00011827631009094846,
|
1675 |
-
"loss": 3.4182,
|
1676 |
-
"step": 2140
|
1677 |
-
},
|
1678 |
-
{
|
1679 |
-
"epoch": 0.44,
|
1680 |
-
"learning_rate": 0.00011784322217410136,
|
1681 |
-
"loss": 3.4001,
|
1682 |
-
"step": 2150
|
1683 |
-
},
|
1684 |
-
{
|
1685 |
-
"epoch": 0.44,
|
1686 |
-
"eval_accuracy": 0.3336403279260422,
|
1687 |
-
"eval_loss": 4.28515625,
|
1688 |
-
"eval_runtime": 6.5938,
|
1689 |
-
"eval_samples_per_second": 5.308,
|
1690 |
-
"eval_steps_per_second": 0.303,
|
1691 |
-
"step": 2150
|
1692 |
-
},
|
1693 |
-
{
|
1694 |
-
"epoch": 0.44,
|
1695 |
-
"learning_rate": 0.00011741013425725423,
|
1696 |
-
"loss": 3.4289,
|
1697 |
-
"step": 2160
|
1698 |
-
},
|
1699 |
-
{
|
1700 |
-
"epoch": 0.45,
|
1701 |
-
"learning_rate": 0.00011697704634040711,
|
1702 |
-
"loss": 3.4228,
|
1703 |
-
"step": 2170
|
1704 |
-
},
|
1705 |
-
{
|
1706 |
-
"epoch": 0.45,
|
1707 |
-
"learning_rate": 0.00011654395842355998,
|
1708 |
-
"loss": 3.4066,
|
1709 |
-
"step": 2180
|
1710 |
-
},
|
1711 |
-
{
|
1712 |
-
"epoch": 0.45,
|
1713 |
-
"learning_rate": 0.00011611087050671288,
|
1714 |
-
"loss": 3.4823,
|
1715 |
-
"step": 2190
|
1716 |
-
},
|
1717 |
-
{
|
1718 |
-
"epoch": 0.45,
|
1719 |
-
"learning_rate": 0.00011567778258986575,
|
1720 |
-
"loss": 3.4497,
|
1721 |
-
"step": 2200
|
1722 |
-
},
|
1723 |
-
{
|
1724 |
-
"epoch": 0.45,
|
1725 |
-
"eval_accuracy": 0.3340380254665969,
|
1726 |
-
"eval_loss": 4.28125,
|
1727 |
-
"eval_runtime": 6.606,
|
1728 |
-
"eval_samples_per_second": 5.298,
|
1729 |
-
"eval_steps_per_second": 0.303,
|
1730 |
-
"step": 2200
|
1731 |
-
},
|
1732 |
-
{
|
1733 |
-
"epoch": 0.45,
|
1734 |
-
"learning_rate": 0.00011524469467301864,
|
1735 |
-
"loss": 3.4737,
|
1736 |
-
"step": 2210
|
1737 |
-
},
|
1738 |
-
{
|
1739 |
-
"epoch": 0.46,
|
1740 |
-
"learning_rate": 0.0001148116067561715,
|
1741 |
-
"loss": 3.4564,
|
1742 |
-
"step": 2220
|
1743 |
-
},
|
1744 |
-
{
|
1745 |
-
"epoch": 0.46,
|
1746 |
-
"learning_rate": 0.00011437851883932438,
|
1747 |
-
"loss": 3.4865,
|
1748 |
-
"step": 2230
|
1749 |
-
},
|
1750 |
-
{
|
1751 |
-
"epoch": 0.46,
|
1752 |
-
"learning_rate": 0.00011394543092247727,
|
1753 |
-
"loss": 3.4496,
|
1754 |
-
"step": 2240
|
1755 |
-
},
|
1756 |
-
{
|
1757 |
-
"epoch": 0.46,
|
1758 |
-
"learning_rate": 0.00011351234300563015,
|
1759 |
-
"loss": 3.4068,
|
1760 |
-
"step": 2250
|
1761 |
-
},
|
1762 |
-
{
|
1763 |
-
"epoch": 0.46,
|
1764 |
-
"eval_accuracy": 0.33329147043432755,
|
1765 |
-
"eval_loss": 4.27734375,
|
1766 |
-
"eval_runtime": 6.6133,
|
1767 |
-
"eval_samples_per_second": 5.292,
|
1768 |
-
"eval_steps_per_second": 0.302,
|
1769 |
-
"step": 2250
|
1770 |
-
},
|
1771 |
-
{
|
1772 |
-
"epoch": 0.46,
|
1773 |
-
"learning_rate": 0.00011307925508878303,
|
1774 |
-
"loss": 3.4337,
|
1775 |
-
"step": 2260
|
1776 |
-
},
|
1777 |
-
{
|
1778 |
-
"epoch": 0.47,
|
1779 |
-
"learning_rate": 0.0001126461671719359,
|
1780 |
-
"loss": 3.4338,
|
1781 |
-
"step": 2270
|
1782 |
-
},
|
1783 |
-
{
|
1784 |
-
"epoch": 0.47,
|
1785 |
-
"learning_rate": 0.0001122130792550888,
|
1786 |
-
"loss": 3.4241,
|
1787 |
-
"step": 2280
|
1788 |
-
},
|
1789 |
-
{
|
1790 |
-
"epoch": 0.47,
|
1791 |
-
"learning_rate": 0.00011177999133824167,
|
1792 |
-
"loss": 3.497,
|
1793 |
-
"step": 2290
|
1794 |
-
},
|
1795 |
-
{
|
1796 |
-
"epoch": 0.47,
|
1797 |
-
"learning_rate": 0.00011134690342139455,
|
1798 |
-
"loss": 3.4634,
|
1799 |
-
"step": 2300
|
1800 |
-
},
|
1801 |
-
{
|
1802 |
-
"epoch": 0.47,
|
1803 |
-
"eval_accuracy": 0.3339822082679226,
|
1804 |
-
"eval_loss": 4.2734375,
|
1805 |
-
"eval_runtime": 6.5925,
|
1806 |
-
"eval_samples_per_second": 5.309,
|
1807 |
-
"eval_steps_per_second": 0.303,
|
1808 |
-
"step": 2300
|
1809 |
-
},
|
1810 |
-
{
|
1811 |
-
"epoch": 0.48,
|
1812 |
-
"learning_rate": 0.00011091381550454742,
|
1813 |
-
"loss": 3.4245,
|
1814 |
-
"step": 2310
|
1815 |
-
},
|
1816 |
-
{
|
1817 |
-
"epoch": 0.48,
|
1818 |
-
"learning_rate": 0.00011048072758770032,
|
1819 |
-
"loss": 3.4465,
|
1820 |
-
"step": 2320
|
1821 |
-
},
|
1822 |
-
{
|
1823 |
-
"epoch": 0.48,
|
1824 |
-
"learning_rate": 0.00011004763967085319,
|
1825 |
-
"loss": 3.498,
|
1826 |
-
"step": 2330
|
1827 |
-
},
|
1828 |
-
{
|
1829 |
-
"epoch": 0.48,
|
1830 |
-
"learning_rate": 0.00010961455175400607,
|
1831 |
-
"loss": 3.3637,
|
1832 |
-
"step": 2340
|
1833 |
-
},
|
1834 |
-
{
|
1835 |
-
"epoch": 0.48,
|
1836 |
-
"learning_rate": 0.00010918146383715895,
|
1837 |
-
"loss": 3.4324,
|
1838 |
-
"step": 2350
|
1839 |
-
},
|
1840 |
-
{
|
1841 |
-
"epoch": 0.48,
|
1842 |
-
"eval_accuracy": 0.333849642421071,
|
1843 |
-
"eval_loss": 4.27734375,
|
1844 |
-
"eval_runtime": 6.5996,
|
1845 |
-
"eval_samples_per_second": 5.303,
|
1846 |
-
"eval_steps_per_second": 0.303,
|
1847 |
-
"step": 2350
|
1848 |
-
},
|
1849 |
-
{
|
1850 |
-
"epoch": 0.49,
|
1851 |
-
"learning_rate": 0.00010874837592031182,
|
1852 |
-
"loss": 3.3816,
|
1853 |
-
"step": 2360
|
1854 |
-
},
|
1855 |
-
{
|
1856 |
-
"epoch": 0.49,
|
1857 |
-
"learning_rate": 0.00010831528800346471,
|
1858 |
-
"loss": 3.4245,
|
1859 |
-
"step": 2370
|
1860 |
-
},
|
1861 |
-
{
|
1862 |
-
"epoch": 0.49,
|
1863 |
-
"learning_rate": 0.00010788220008661758,
|
1864 |
-
"loss": 3.4567,
|
1865 |
-
"step": 2380
|
1866 |
-
},
|
1867 |
-
{
|
1868 |
-
"epoch": 0.49,
|
1869 |
-
"learning_rate": 0.00010744911216977047,
|
1870 |
-
"loss": 3.4122,
|
1871 |
-
"step": 2390
|
1872 |
-
},
|
1873 |
-
{
|
1874 |
-
"epoch": 0.49,
|
1875 |
-
"learning_rate": 0.00010701602425292334,
|
1876 |
-
"loss": 3.4039,
|
1877 |
-
"step": 2400
|
1878 |
-
},
|
1879 |
-
{
|
1880 |
-
"epoch": 0.49,
|
1881 |
-
"eval_accuracy": 0.3344147915576487,
|
1882 |
-
"eval_loss": 4.265625,
|
1883 |
-
"eval_runtime": 6.5992,
|
1884 |
-
"eval_samples_per_second": 5.304,
|
1885 |
-
"eval_steps_per_second": 0.303,
|
1886 |
-
"step": 2400
|
1887 |
-
},
|
1888 |
-
{
|
1889 |
-
"epoch": 0.5,
|
1890 |
-
"learning_rate": 0.00010658293633607624,
|
1891 |
-
"loss": 3.4124,
|
1892 |
-
"step": 2410
|
1893 |
-
},
|
1894 |
-
{
|
1895 |
-
"epoch": 0.5,
|
1896 |
-
"learning_rate": 0.00010614984841922911,
|
1897 |
-
"loss": 3.4309,
|
1898 |
-
"step": 2420
|
1899 |
-
},
|
1900 |
-
{
|
1901 |
-
"epoch": 0.5,
|
1902 |
-
"learning_rate": 0.00010571676050238199,
|
1903 |
-
"loss": 3.4464,
|
1904 |
-
"step": 2430
|
1905 |
-
},
|
1906 |
-
{
|
1907 |
-
"epoch": 0.5,
|
1908 |
-
"learning_rate": 0.00010528367258553486,
|
1909 |
-
"loss": 3.4136,
|
1910 |
-
"step": 2440
|
1911 |
-
},
|
1912 |
-
{
|
1913 |
-
"epoch": 0.5,
|
1914 |
-
"learning_rate": 0.00010485058466868776,
|
1915 |
-
"loss": 3.4502,
|
1916 |
-
"step": 2450
|
1917 |
-
},
|
1918 |
-
{
|
1919 |
-
"epoch": 0.5,
|
1920 |
-
"eval_accuracy": 0.33453340310483165,
|
1921 |
-
"eval_loss": 4.265625,
|
1922 |
-
"eval_runtime": 6.7512,
|
1923 |
-
"eval_samples_per_second": 5.184,
|
1924 |
-
"eval_steps_per_second": 0.296,
|
1925 |
-
"step": 2450
|
1926 |
-
},
|
1927 |
-
{
|
1928 |
-
"epoch": 0.51,
|
1929 |
-
"learning_rate": 0.00010441749675184063,
|
1930 |
-
"loss": 3.4062,
|
1931 |
-
"step": 2460
|
1932 |
-
},
|
1933 |
-
{
|
1934 |
-
"epoch": 0.51,
|
1935 |
-
"learning_rate": 0.00010398440883499351,
|
1936 |
-
"loss": 3.3951,
|
1937 |
-
"step": 2470
|
1938 |
-
},
|
1939 |
-
{
|
1940 |
-
"epoch": 0.51,
|
1941 |
-
"learning_rate": 0.00010355132091814638,
|
1942 |
-
"loss": 3.4322,
|
1943 |
-
"step": 2480
|
1944 |
-
},
|
1945 |
-
{
|
1946 |
-
"epoch": 0.51,
|
1947 |
-
"learning_rate": 0.00010311823300129926,
|
1948 |
-
"loss": 3.4107,
|
1949 |
-
"step": 2490
|
1950 |
-
},
|
1951 |
-
{
|
1952 |
-
"epoch": 0.51,
|
1953 |
-
"learning_rate": 0.00010268514508445215,
|
1954 |
-
"loss": 3.4104,
|
1955 |
-
"step": 2500
|
1956 |
-
},
|
1957 |
-
{
|
1958 |
-
"epoch": 0.51,
|
1959 |
-
"eval_accuracy": 0.33495900924472355,
|
1960 |
-
"eval_loss": 4.2578125,
|
1961 |
-
"eval_runtime": 6.5889,
|
1962 |
-
"eval_samples_per_second": 5.312,
|
1963 |
-
"eval_steps_per_second": 0.304,
|
1964 |
-
"step": 2500
|
1965 |
-
},
|
1966 |
-
{
|
1967 |
-
"epoch": 0.52,
|
1968 |
-
"learning_rate": 0.00010225205716760502,
|
1969 |
-
"loss": 3.4181,
|
1970 |
-
"step": 2510
|
1971 |
-
},
|
1972 |
-
{
|
1973 |
-
"epoch": 0.52,
|
1974 |
-
"learning_rate": 0.00010181896925075791,
|
1975 |
-
"loss": 3.3979,
|
1976 |
-
"step": 2520
|
1977 |
-
},
|
1978 |
-
{
|
1979 |
-
"epoch": 0.52,
|
1980 |
-
"learning_rate": 0.00010138588133391078,
|
1981 |
-
"loss": 3.3911,
|
1982 |
-
"step": 2530
|
1983 |
-
},
|
1984 |
-
{
|
1985 |
-
"epoch": 0.52,
|
1986 |
-
"learning_rate": 0.00010095279341706368,
|
1987 |
-
"loss": 3.4303,
|
1988 |
-
"step": 2540
|
1989 |
-
},
|
1990 |
-
{
|
1991 |
-
"epoch": 0.52,
|
1992 |
-
"learning_rate": 0.00010051970550021655,
|
1993 |
-
"loss": 3.5251,
|
1994 |
-
"step": 2550
|
1995 |
-
},
|
1996 |
-
{
|
1997 |
-
"epoch": 0.52,
|
1998 |
-
"eval_accuracy": 0.33599860457003317,
|
1999 |
-
"eval_loss": 4.2578125,
|
2000 |
-
"eval_runtime": 6.5946,
|
2001 |
-
"eval_samples_per_second": 5.307,
|
2002 |
-
"eval_steps_per_second": 0.303,
|
2003 |
-
"step": 2550
|
2004 |
-
},
|
2005 |
-
{
|
2006 |
-
"epoch": 0.53,
|
2007 |
-
"learning_rate": 0.00010008661758336943,
|
2008 |
-
"loss": 3.4337,
|
2009 |
-
"step": 2560
|
2010 |
-
},
|
2011 |
-
{
|
2012 |
-
"epoch": 0.53,
|
2013 |
-
"learning_rate": 9.96535296665223e-05,
|
2014 |
-
"loss": 3.3906,
|
2015 |
-
"step": 2570
|
2016 |
-
},
|
2017 |
-
{
|
2018 |
-
"epoch": 0.53,
|
2019 |
-
"learning_rate": 9.922044174967519e-05,
|
2020 |
-
"loss": 3.4714,
|
2021 |
-
"step": 2580
|
2022 |
-
},
|
2023 |
-
{
|
2024 |
-
"epoch": 0.53,
|
2025 |
-
"learning_rate": 9.878735383282807e-05,
|
2026 |
-
"loss": 3.4061,
|
2027 |
-
"step": 2590
|
2028 |
-
},
|
2029 |
-
{
|
2030 |
-
"epoch": 0.53,
|
2031 |
-
"learning_rate": 9.835426591598095e-05,
|
2032 |
-
"loss": 3.4176,
|
2033 |
-
"step": 2600
|
2034 |
-
},
|
2035 |
-
{
|
2036 |
-
"epoch": 0.53,
|
2037 |
-
"eval_accuracy": 0.3363404849119135,
|
2038 |
-
"eval_loss": 4.25,
|
2039 |
-
"eval_runtime": 6.594,
|
2040 |
-
"eval_samples_per_second": 5.308,
|
2041 |
-
"eval_steps_per_second": 0.303,
|
2042 |
-
"step": 2600
|
2043 |
-
},
|
2044 |
-
{
|
2045 |
-
"epoch": 0.54,
|
2046 |
-
"learning_rate": 9.792117799913382e-05,
|
2047 |
-
"loss": 3.37,
|
2048 |
-
"step": 2610
|
2049 |
-
},
|
2050 |
-
{
|
2051 |
-
"epoch": 0.54,
|
2052 |
-
"learning_rate": 9.748809008228671e-05,
|
2053 |
-
"loss": 3.3931,
|
2054 |
-
"step": 2620
|
2055 |
-
},
|
2056 |
-
{
|
2057 |
-
"epoch": 0.54,
|
2058 |
-
"learning_rate": 9.705500216543959e-05,
|
2059 |
-
"loss": 3.4052,
|
2060 |
-
"step": 2630
|
2061 |
-
},
|
2062 |
-
{
|
2063 |
-
"epoch": 0.54,
|
2064 |
-
"learning_rate": 9.662191424859248e-05,
|
2065 |
-
"loss": 3.4269,
|
2066 |
-
"step": 2640
|
2067 |
-
},
|
2068 |
-
{
|
2069 |
-
"epoch": 0.54,
|
2070 |
-
"learning_rate": 9.618882633174535e-05,
|
2071 |
-
"loss": 3.3795,
|
2072 |
-
"step": 2650
|
2073 |
-
},
|
2074 |
-
{
|
2075 |
-
"epoch": 0.54,
|
2076 |
-
"eval_accuracy": 0.3354404325832897,
|
2077 |
-
"eval_loss": 4.25,
|
2078 |
-
"eval_runtime": 6.6039,
|
2079 |
-
"eval_samples_per_second": 5.3,
|
2080 |
-
"eval_steps_per_second": 0.303,
|
2081 |
-
"step": 2650
|
2082 |
-
},
|
2083 |
-
{
|
2084 |
-
"epoch": 0.55,
|
2085 |
-
"learning_rate": 9.575573841489823e-05,
|
2086 |
-
"loss": 3.4064,
|
2087 |
-
"step": 2660
|
2088 |
-
},
|
2089 |
-
{
|
2090 |
-
"epoch": 0.55,
|
2091 |
-
"learning_rate": 9.532265049805112e-05,
|
2092 |
-
"loss": 3.4207,
|
2093 |
-
"step": 2670
|
2094 |
-
},
|
2095 |
-
{
|
2096 |
-
"epoch": 0.55,
|
2097 |
-
"learning_rate": 9.488956258120399e-05,
|
2098 |
-
"loss": 3.4353,
|
2099 |
-
"step": 2680
|
2100 |
-
},
|
2101 |
-
{
|
2102 |
-
"epoch": 0.55,
|
2103 |
-
"learning_rate": 9.445647466435687e-05,
|
2104 |
-
"loss": 3.4497,
|
2105 |
-
"step": 2690
|
2106 |
-
},
|
2107 |
-
{
|
2108 |
-
"epoch": 0.56,
|
2109 |
-
"learning_rate": 9.402338674750974e-05,
|
2110 |
-
"loss": 3.3656,
|
2111 |
-
"step": 2700
|
2112 |
-
},
|
2113 |
-
{
|
2114 |
-
"epoch": 0.56,
|
2115 |
-
"eval_accuracy": 0.33636839351125064,
|
2116 |
-
"eval_loss": 4.25,
|
2117 |
-
"eval_runtime": 6.5948,
|
2118 |
-
"eval_samples_per_second": 5.307,
|
2119 |
-
"eval_steps_per_second": 0.303,
|
2120 |
-
"step": 2700
|
2121 |
-
},
|
2122 |
-
{
|
2123 |
-
"epoch": 0.56,
|
2124 |
-
"learning_rate": 9.359029883066262e-05,
|
2125 |
-
"loss": 3.3736,
|
2126 |
-
"step": 2710
|
2127 |
-
},
|
2128 |
-
{
|
2129 |
-
"epoch": 0.56,
|
2130 |
-
"learning_rate": 9.315721091381551e-05,
|
2131 |
-
"loss": 3.4236,
|
2132 |
-
"step": 2720
|
2133 |
-
},
|
2134 |
-
{
|
2135 |
-
"epoch": 0.56,
|
2136 |
-
"learning_rate": 9.272412299696839e-05,
|
2137 |
-
"loss": 3.4234,
|
2138 |
-
"step": 2730
|
2139 |
-
},
|
2140 |
-
{
|
2141 |
-
"epoch": 0.56,
|
2142 |
-
"learning_rate": 9.229103508012126e-05,
|
2143 |
-
"loss": 3.3849,
|
2144 |
-
"step": 2740
|
2145 |
-
},
|
2146 |
-
{
|
2147 |
-
"epoch": 0.57,
|
2148 |
-
"learning_rate": 9.185794716327415e-05,
|
2149 |
-
"loss": 3.3938,
|
2150 |
-
"step": 2750
|
2151 |
-
},
|
2152 |
-
{
|
2153 |
-
"epoch": 0.57,
|
2154 |
-
"eval_accuracy": 0.33627769056340484,
|
2155 |
-
"eval_loss": 4.24609375,
|
2156 |
-
"eval_runtime": 6.5953,
|
2157 |
-
"eval_samples_per_second": 5.307,
|
2158 |
-
"eval_steps_per_second": 0.303,
|
2159 |
-
"step": 2750
|
2160 |
-
},
|
2161 |
-
{
|
2162 |
-
"epoch": 0.57,
|
2163 |
-
"learning_rate": 9.142485924642703e-05,
|
2164 |
-
"loss": 3.375,
|
2165 |
-
"step": 2760
|
2166 |
-
},
|
2167 |
-
{
|
2168 |
-
"epoch": 0.57,
|
2169 |
-
"learning_rate": 9.099177132957992e-05,
|
2170 |
-
"loss": 3.4365,
|
2171 |
-
"step": 2770
|
2172 |
-
},
|
2173 |
-
{
|
2174 |
-
"epoch": 0.57,
|
2175 |
-
"learning_rate": 9.055868341273279e-05,
|
2176 |
-
"loss": 3.4068,
|
2177 |
-
"step": 2780
|
2178 |
-
},
|
2179 |
-
{
|
2180 |
-
"epoch": 0.57,
|
2181 |
-
"learning_rate": 9.012559549588567e-05,
|
2182 |
-
"loss": 3.4333,
|
2183 |
-
"step": 2790
|
2184 |
-
},
|
2185 |
-
{
|
2186 |
-
"epoch": 0.58,
|
2187 |
-
"learning_rate": 8.969250757903855e-05,
|
2188 |
-
"loss": 3.3757,
|
2189 |
-
"step": 2800
|
2190 |
-
},
|
2191 |
-
{
|
2192 |
-
"epoch": 0.58,
|
2193 |
-
"eval_accuracy": 0.33639630211058785,
|
2194 |
-
"eval_loss": 4.24609375,
|
2195 |
-
"eval_runtime": 6.5987,
|
2196 |
-
"eval_samples_per_second": 5.304,
|
2197 |
-
"eval_steps_per_second": 0.303,
|
2198 |
-
"step": 2800
|
2199 |
-
},
|
2200 |
-
{
|
2201 |
-
"epoch": 0.58,
|
2202 |
-
"learning_rate": 8.925941966219143e-05,
|
2203 |
-
"loss": 3.4111,
|
2204 |
-
"step": 2810
|
2205 |
-
},
|
2206 |
-
{
|
2207 |
-
"epoch": 0.58,
|
2208 |
-
"learning_rate": 8.882633174534431e-05,
|
2209 |
-
"loss": 3.4143,
|
2210 |
-
"step": 2820
|
2211 |
-
},
|
2212 |
-
{
|
2213 |
-
"epoch": 0.58,
|
2214 |
-
"learning_rate": 8.839324382849718e-05,
|
2215 |
-
"loss": 3.4307,
|
2216 |
-
"step": 2830
|
2217 |
-
},
|
2218 |
-
{
|
2219 |
-
"epoch": 0.58,
|
2220 |
-
"learning_rate": 8.796015591165006e-05,
|
2221 |
-
"loss": 3.3366,
|
2222 |
-
"step": 2840
|
2223 |
-
},
|
2224 |
-
{
|
2225 |
-
"epoch": 0.59,
|
2226 |
-
"learning_rate": 8.752706799480295e-05,
|
2227 |
-
"loss": 3.407,
|
2228 |
-
"step": 2850
|
2229 |
-
},
|
2230 |
-
{
|
2231 |
-
"epoch": 0.59,
|
2232 |
-
"eval_accuracy": 0.33731728588871446,
|
2233 |
-
"eval_loss": 4.234375,
|
2234 |
-
"eval_runtime": 6.6038,
|
2235 |
-
"eval_samples_per_second": 5.3,
|
2236 |
-
"eval_steps_per_second": 0.303,
|
2237 |
-
"step": 2850
|
2238 |
-
},
|
2239 |
-
{
|
2240 |
-
"epoch": 0.59,
|
2241 |
-
"learning_rate": 8.709398007795583e-05,
|
2242 |
-
"loss": 3.3506,
|
2243 |
-
"step": 2860
|
2244 |
-
},
|
2245 |
-
{
|
2246 |
-
"epoch": 0.59,
|
2247 |
-
"learning_rate": 8.66608921611087e-05,
|
2248 |
-
"loss": 3.429,
|
2249 |
-
"step": 2870
|
2250 |
-
},
|
2251 |
-
{
|
2252 |
-
"epoch": 0.59,
|
2253 |
-
"learning_rate": 8.622780424426159e-05,
|
2254 |
-
"loss": 3.338,
|
2255 |
-
"step": 2880
|
2256 |
-
},
|
2257 |
-
{
|
2258 |
-
"epoch": 0.59,
|
2259 |
-
"learning_rate": 8.579471632741447e-05,
|
2260 |
-
"loss": 3.4252,
|
2261 |
-
"step": 2890
|
2262 |
-
},
|
2263 |
-
{
|
2264 |
-
"epoch": 0.6,
|
2265 |
-
"learning_rate": 8.536162841056736e-05,
|
2266 |
-
"loss": 3.3986,
|
2267 |
-
"step": 2900
|
2268 |
-
},
|
2269 |
-
{
|
2270 |
-
"epoch": 0.6,
|
2271 |
-
"eval_accuracy": 0.33657770800627945,
|
2272 |
-
"eval_loss": 4.23828125,
|
2273 |
-
"eval_runtime": 6.5874,
|
2274 |
-
"eval_samples_per_second": 5.313,
|
2275 |
-
"eval_steps_per_second": 0.304,
|
2276 |
-
"step": 2900
|
2277 |
-
},
|
2278 |
-
{
|
2279 |
-
"epoch": 0.6,
|
2280 |
-
"learning_rate": 8.492854049372023e-05,
|
2281 |
-
"loss": 3.4023,
|
2282 |
-
"step": 2910
|
2283 |
-
},
|
2284 |
-
{
|
2285 |
-
"epoch": 0.6,
|
2286 |
-
"learning_rate": 8.449545257687311e-05,
|
2287 |
-
"loss": 3.3919,
|
2288 |
-
"step": 2920
|
2289 |
-
},
|
2290 |
-
{
|
2291 |
-
"epoch": 0.6,
|
2292 |
-
"learning_rate": 8.4062364660026e-05,
|
2293 |
-
"loss": 3.3691,
|
2294 |
-
"step": 2930
|
2295 |
-
},
|
2296 |
-
{
|
2297 |
-
"epoch": 0.6,
|
2298 |
-
"learning_rate": 8.362927674317888e-05,
|
2299 |
-
"loss": 3.4017,
|
2300 |
-
"step": 2940
|
2301 |
-
},
|
2302 |
-
{
|
2303 |
-
"epoch": 0.61,
|
2304 |
-
"learning_rate": 8.319618882633175e-05,
|
2305 |
-
"loss": 3.4311,
|
2306 |
-
"step": 2950
|
2307 |
-
},
|
2308 |
-
{
|
2309 |
-
"epoch": 0.61,
|
2310 |
-
"eval_accuracy": 0.3370940170940171,
|
2311 |
-
"eval_loss": 4.234375,
|
2312 |
-
"eval_runtime": 6.5922,
|
2313 |
-
"eval_samples_per_second": 5.309,
|
2314 |
-
"eval_steps_per_second": 0.303,
|
2315 |
-
"step": 2950
|
2316 |
-
},
|
2317 |
-
{
|
2318 |
-
"epoch": 0.61,
|
2319 |
-
"learning_rate": 8.276310090948462e-05,
|
2320 |
-
"loss": 3.4224,
|
2321 |
-
"step": 2960
|
2322 |
-
},
|
2323 |
-
{
|
2324 |
-
"epoch": 0.61,
|
2325 |
-
"learning_rate": 8.23300129926375e-05,
|
2326 |
-
"loss": 3.3781,
|
2327 |
-
"step": 2970
|
2328 |
-
},
|
2329 |
-
{
|
2330 |
-
"epoch": 0.61,
|
2331 |
-
"learning_rate": 8.189692507579039e-05,
|
2332 |
-
"loss": 3.383,
|
2333 |
-
"step": 2980
|
2334 |
-
},
|
2335 |
-
{
|
2336 |
-
"epoch": 0.61,
|
2337 |
-
"learning_rate": 8.146383715894327e-05,
|
2338 |
-
"loss": 3.4332,
|
2339 |
-
"step": 2990
|
2340 |
-
},
|
2341 |
-
{
|
2342 |
-
"epoch": 0.62,
|
2343 |
-
"learning_rate": 8.103074924209616e-05,
|
2344 |
-
"loss": 3.3716,
|
2345 |
-
"step": 3000
|
2346 |
-
},
|
2347 |
-
{
|
2348 |
-
"epoch": 0.62,
|
2349 |
-
"eval_accuracy": 0.33713587999302286,
|
2350 |
-
"eval_loss": 4.234375,
|
2351 |
-
"eval_runtime": 6.5875,
|
2352 |
-
"eval_samples_per_second": 5.313,
|
2353 |
-
"eval_steps_per_second": 0.304,
|
2354 |
-
"step": 3000
|
2355 |
-
},
|
2356 |
-
{
|
2357 |
-
"epoch": 0.62,
|
2358 |
-
"learning_rate": 8.059766132524903e-05,
|
2359 |
-
"loss": 3.4123,
|
2360 |
-
"step": 3010
|
2361 |
-
},
|
2362 |
-
{
|
2363 |
-
"epoch": 0.62,
|
2364 |
-
"learning_rate": 8.016457340840191e-05,
|
2365 |
-
"loss": 3.4181,
|
2366 |
-
"step": 3020
|
2367 |
-
},
|
2368 |
-
{
|
2369 |
-
"epoch": 0.62,
|
2370 |
-
"learning_rate": 7.97314854915548e-05,
|
2371 |
-
"loss": 3.3851,
|
2372 |
-
"step": 3030
|
2373 |
-
},
|
2374 |
-
{
|
2375 |
-
"epoch": 0.63,
|
2376 |
-
"learning_rate": 7.929839757470768e-05,
|
2377 |
-
"loss": 3.4224,
|
2378 |
-
"step": 3040
|
2379 |
-
},
|
2380 |
-
{
|
2381 |
-
"epoch": 0.63,
|
2382 |
-
"learning_rate": 7.886530965786055e-05,
|
2383 |
-
"loss": 3.3831,
|
2384 |
-
"step": 3050
|
2385 |
-
},
|
2386 |
-
{
|
2387 |
-
"epoch": 0.63,
|
2388 |
-
"eval_accuracy": 0.33774986917844063,
|
2389 |
-
"eval_loss": 4.23046875,
|
2390 |
-
"eval_runtime": 6.5895,
|
2391 |
-
"eval_samples_per_second": 5.311,
|
2392 |
-
"eval_steps_per_second": 0.304,
|
2393 |
-
"step": 3050
|
2394 |
-
},
|
2395 |
-
{
|
2396 |
-
"epoch": 0.63,
|
2397 |
-
"learning_rate": 7.843222174101343e-05,
|
2398 |
-
"loss": 3.3965,
|
2399 |
-
"step": 3060
|
2400 |
-
},
|
2401 |
-
{
|
2402 |
-
"epoch": 0.63,
|
2403 |
-
"learning_rate": 7.799913382416632e-05,
|
2404 |
-
"loss": 3.4016,
|
2405 |
-
"step": 3070
|
2406 |
-
},
|
2407 |
-
{
|
2408 |
-
"epoch": 0.63,
|
2409 |
-
"learning_rate": 7.756604590731919e-05,
|
2410 |
-
"loss": 3.4047,
|
2411 |
-
"step": 3080
|
2412 |
-
},
|
2413 |
-
{
|
2414 |
-
"epoch": 0.64,
|
2415 |
-
"learning_rate": 7.713295799047207e-05,
|
2416 |
-
"loss": 3.3618,
|
2417 |
-
"step": 3090
|
2418 |
-
},
|
2419 |
-
{
|
2420 |
-
"epoch": 0.64,
|
2421 |
-
"learning_rate": 7.669987007362494e-05,
|
2422 |
-
"loss": 3.375,
|
2423 |
-
"step": 3100
|
2424 |
-
},
|
2425 |
-
{
|
2426 |
-
"epoch": 0.64,
|
2427 |
-
"eval_accuracy": 0.3377638234781092,
|
2428 |
-
"eval_loss": 4.23046875,
|
2429 |
-
"eval_runtime": 6.5887,
|
2430 |
-
"eval_samples_per_second": 5.312,
|
2431 |
-
"eval_steps_per_second": 0.304,
|
2432 |
-
"step": 3100
|
2433 |
-
},
|
2434 |
-
{
|
2435 |
-
"epoch": 0.64,
|
2436 |
-
"learning_rate": 7.626678215677783e-05,
|
2437 |
-
"loss": 3.3899,
|
2438 |
-
"step": 3110
|
2439 |
-
},
|
2440 |
-
{
|
2441 |
-
"epoch": 0.64,
|
2442 |
-
"learning_rate": 7.583369423993071e-05,
|
2443 |
-
"loss": 3.3723,
|
2444 |
-
"step": 3120
|
2445 |
-
},
|
2446 |
-
{
|
2447 |
-
"epoch": 0.64,
|
2448 |
-
"learning_rate": 7.54006063230836e-05,
|
2449 |
-
"loss": 3.381,
|
2450 |
-
"step": 3130
|
2451 |
-
},
|
2452 |
-
{
|
2453 |
-
"epoch": 0.65,
|
2454 |
-
"learning_rate": 7.496751840623647e-05,
|
2455 |
-
"loss": 3.3558,
|
2456 |
-
"step": 3140
|
2457 |
-
},
|
2458 |
-
{
|
2459 |
-
"epoch": 0.65,
|
2460 |
-
"learning_rate": 7.453443048938935e-05,
|
2461 |
-
"loss": 3.3677,
|
2462 |
-
"step": 3150
|
2463 |
-
},
|
2464 |
-
{
|
2465 |
-
"epoch": 0.65,
|
2466 |
-
"eval_accuracy": 0.33846851561137276,
|
2467 |
-
"eval_loss": 4.21875,
|
2468 |
-
"eval_runtime": 6.5962,
|
2469 |
-
"eval_samples_per_second": 5.306,
|
2470 |
-
"eval_steps_per_second": 0.303,
|
2471 |
-
"step": 3150
|
2472 |
-
},
|
2473 |
-
{
|
2474 |
-
"epoch": 0.65,
|
2475 |
-
"learning_rate": 7.410134257254223e-05,
|
2476 |
-
"loss": 3.3414,
|
2477 |
-
"step": 3160
|
2478 |
-
},
|
2479 |
-
{
|
2480 |
-
"epoch": 0.65,
|
2481 |
-
"learning_rate": 7.366825465569512e-05,
|
2482 |
-
"loss": 3.4144,
|
2483 |
-
"step": 3170
|
2484 |
-
},
|
2485 |
-
{
|
2486 |
-
"epoch": 0.65,
|
2487 |
-
"learning_rate": 7.323516673884799e-05,
|
2488 |
-
"loss": 3.3503,
|
2489 |
-
"step": 3180
|
2490 |
-
},
|
2491 |
-
{
|
2492 |
-
"epoch": 0.66,
|
2493 |
-
"learning_rate": 7.280207882200087e-05,
|
2494 |
-
"loss": 3.3716,
|
2495 |
-
"step": 3190
|
2496 |
-
},
|
2497 |
-
{
|
2498 |
-
"epoch": 0.66,
|
2499 |
-
"learning_rate": 7.236899090515376e-05,
|
2500 |
-
"loss": 3.3968,
|
2501 |
-
"step": 3200
|
2502 |
-
},
|
2503 |
-
{
|
2504 |
-
"epoch": 0.66,
|
2505 |
-
"eval_accuracy": 0.3386220129077272,
|
2506 |
-
"eval_loss": 4.22265625,
|
2507 |
-
"eval_runtime": 6.5929,
|
2508 |
-
"eval_samples_per_second": 5.309,
|
2509 |
-
"eval_steps_per_second": 0.303,
|
2510 |
-
"step": 3200
|
2511 |
-
},
|
2512 |
-
{
|
2513 |
-
"epoch": 0.66,
|
2514 |
-
"learning_rate": 7.193590298830663e-05,
|
2515 |
-
"loss": 3.3485,
|
2516 |
-
"step": 3210
|
2517 |
-
},
|
2518 |
-
{
|
2519 |
-
"epoch": 0.66,
|
2520 |
-
"learning_rate": 7.150281507145951e-05,
|
2521 |
-
"loss": 3.4172,
|
2522 |
-
"step": 3220
|
2523 |
-
},
|
2524 |
-
{
|
2525 |
-
"epoch": 0.66,
|
2526 |
-
"learning_rate": 7.106972715461238e-05,
|
2527 |
-
"loss": 3.3727,
|
2528 |
-
"step": 3230
|
2529 |
-
},
|
2530 |
-
{
|
2531 |
-
"epoch": 0.67,
|
2532 |
-
"learning_rate": 7.063663923776527e-05,
|
2533 |
-
"loss": 3.3616,
|
2534 |
-
"step": 3240
|
2535 |
-
},
|
2536 |
-
{
|
2537 |
-
"epoch": 0.67,
|
2538 |
-
"learning_rate": 7.020355132091815e-05,
|
2539 |
-
"loss": 3.4069,
|
2540 |
-
"step": 3250
|
2541 |
-
},
|
2542 |
-
{
|
2543 |
-
"epoch": 0.67,
|
2544 |
-
"eval_accuracy": 0.3380429094714809,
|
2545 |
-
"eval_loss": 4.21875,
|
2546 |
-
"eval_runtime": 6.5865,
|
2547 |
-
"eval_samples_per_second": 5.314,
|
2548 |
-
"eval_steps_per_second": 0.304,
|
2549 |
-
"step": 3250
|
2550 |
-
},
|
2551 |
-
{
|
2552 |
-
"epoch": 0.67,
|
2553 |
-
"learning_rate": 6.977046340407103e-05,
|
2554 |
-
"loss": 3.3583,
|
2555 |
-
"step": 3260
|
2556 |
-
},
|
2557 |
-
{
|
2558 |
-
"epoch": 0.67,
|
2559 |
-
"learning_rate": 6.93373754872239e-05,
|
2560 |
-
"loss": 3.3753,
|
2561 |
-
"step": 3270
|
2562 |
-
},
|
2563 |
-
{
|
2564 |
-
"epoch": 0.67,
|
2565 |
-
"learning_rate": 6.890428757037679e-05,
|
2566 |
-
"loss": 3.3443,
|
2567 |
-
"step": 3280
|
2568 |
-
},
|
2569 |
-
{
|
2570 |
-
"epoch": 0.68,
|
2571 |
-
"learning_rate": 6.847119965352967e-05,
|
2572 |
-
"loss": 3.3682,
|
2573 |
-
"step": 3290
|
2574 |
-
},
|
2575 |
-
{
|
2576 |
-
"epoch": 0.68,
|
2577 |
-
"learning_rate": 6.803811173668256e-05,
|
2578 |
-
"loss": 3.4192,
|
2579 |
-
"step": 3300
|
2580 |
-
},
|
2581 |
-
{
|
2582 |
-
"epoch": 0.68,
|
2583 |
-
"eval_accuracy": 0.33877551020408164,
|
2584 |
-
"eval_loss": 4.21484375,
|
2585 |
-
"eval_runtime": 6.6033,
|
2586 |
-
"eval_samples_per_second": 5.3,
|
2587 |
-
"eval_steps_per_second": 0.303,
|
2588 |
-
"step": 3300
|
2589 |
-
},
|
2590 |
-
{
|
2591 |
-
"epoch": 0.68,
|
2592 |
-
"learning_rate": 6.760502381983543e-05,
|
2593 |
-
"loss": 3.3657,
|
2594 |
-
"step": 3310
|
2595 |
-
},
|
2596 |
-
{
|
2597 |
-
"epoch": 0.68,
|
2598 |
-
"learning_rate": 6.717193590298831e-05,
|
2599 |
-
"loss": 3.3773,
|
2600 |
-
"step": 3320
|
2601 |
-
},
|
2602 |
-
{
|
2603 |
-
"epoch": 0.68,
|
2604 |
-
"learning_rate": 6.67388479861412e-05,
|
2605 |
-
"loss": 3.3604,
|
2606 |
-
"step": 3330
|
2607 |
-
},
|
2608 |
-
{
|
2609 |
-
"epoch": 0.69,
|
2610 |
-
"learning_rate": 6.630576006929407e-05,
|
2611 |
-
"loss": 3.404,
|
2612 |
-
"step": 3340
|
2613 |
-
},
|
2614 |
-
{
|
2615 |
-
"epoch": 0.69,
|
2616 |
-
"learning_rate": 6.587267215244695e-05,
|
2617 |
-
"loss": 3.3881,
|
2618 |
-
"step": 3350
|
2619 |
-
},
|
2620 |
-
{
|
2621 |
-
"epoch": 0.69,
|
2622 |
-
"eval_accuracy": 0.33830804116518404,
|
2623 |
-
"eval_loss": 4.21484375,
|
2624 |
-
"eval_runtime": 6.6138,
|
2625 |
-
"eval_samples_per_second": 5.292,
|
2626 |
-
"eval_steps_per_second": 0.302,
|
2627 |
-
"step": 3350
|
2628 |
-
},
|
2629 |
-
{
|
2630 |
-
"epoch": 0.69,
|
2631 |
-
"learning_rate": 6.543958423559982e-05,
|
2632 |
-
"loss": 3.3855,
|
2633 |
-
"step": 3360
|
2634 |
-
},
|
2635 |
-
{
|
2636 |
-
"epoch": 0.69,
|
2637 |
-
"learning_rate": 6.50064963187527e-05,
|
2638 |
-
"loss": 3.314,
|
2639 |
-
"step": 3370
|
2640 |
-
},
|
2641 |
-
{
|
2642 |
-
"epoch": 0.7,
|
2643 |
-
"learning_rate": 6.457340840190559e-05,
|
2644 |
-
"loss": 3.4034,
|
2645 |
-
"step": 3380
|
2646 |
-
},
|
2647 |
-
{
|
2648 |
-
"epoch": 0.7,
|
2649 |
-
"learning_rate": 6.414032048505847e-05,
|
2650 |
-
"loss": 3.3969,
|
2651 |
-
"step": 3390
|
2652 |
-
},
|
2653 |
-
{
|
2654 |
-
"epoch": 0.7,
|
2655 |
-
"learning_rate": 6.370723256821134e-05,
|
2656 |
-
"loss": 3.3858,
|
2657 |
-
"step": 3400
|
2658 |
-
},
|
2659 |
-
{
|
2660 |
-
"epoch": 0.7,
|
2661 |
-
"eval_accuracy": 0.33836385836385835,
|
2662 |
-
"eval_loss": 4.2109375,
|
2663 |
-
"eval_runtime": 6.6099,
|
2664 |
-
"eval_samples_per_second": 5.295,
|
2665 |
-
"eval_steps_per_second": 0.303,
|
2666 |
-
"step": 3400
|
2667 |
-
},
|
2668 |
-
{
|
2669 |
-
"epoch": 0.7,
|
2670 |
-
"learning_rate": 6.327414465136423e-05,
|
2671 |
-
"loss": 3.388,
|
2672 |
-
"step": 3410
|
2673 |
-
},
|
2674 |
-
{
|
2675 |
-
"epoch": 0.7,
|
2676 |
-
"learning_rate": 6.284105673451711e-05,
|
2677 |
-
"loss": 3.3832,
|
2678 |
-
"step": 3420
|
2679 |
-
},
|
2680 |
-
{
|
2681 |
-
"epoch": 0.71,
|
2682 |
-
"learning_rate": 6.240796881767e-05,
|
2683 |
-
"loss": 3.3098,
|
2684 |
-
"step": 3430
|
2685 |
-
},
|
2686 |
-
{
|
2687 |
-
"epoch": 0.71,
|
2688 |
-
"learning_rate": 6.197488090082287e-05,
|
2689 |
-
"loss": 3.3794,
|
2690 |
-
"step": 3440
|
2691 |
-
},
|
2692 |
-
{
|
2693 |
-
"epoch": 0.71,
|
2694 |
-
"learning_rate": 6.154179298397575e-05,
|
2695 |
-
"loss": 3.3999,
|
2696 |
-
"step": 3450
|
2697 |
-
},
|
2698 |
-
{
|
2699 |
-
"epoch": 0.71,
|
2700 |
-
"eval_accuracy": 0.33883132740275596,
|
2701 |
-
"eval_loss": 4.2109375,
|
2702 |
-
"eval_runtime": 6.6039,
|
2703 |
-
"eval_samples_per_second": 5.3,
|
2704 |
-
"eval_steps_per_second": 0.303,
|
2705 |
-
"step": 3450
|
2706 |
-
},
|
2707 |
-
{
|
2708 |
-
"epoch": 0.71,
|
2709 |
-
"learning_rate": 6.110870506712864e-05,
|
2710 |
-
"loss": 3.3278,
|
2711 |
-
"step": 3460
|
2712 |
-
},
|
2713 |
-
{
|
2714 |
-
"epoch": 0.71,
|
2715 |
-
"learning_rate": 6.0675617150281506e-05,
|
2716 |
-
"loss": 3.3645,
|
2717 |
-
"step": 3470
|
2718 |
-
},
|
2719 |
-
{
|
2720 |
-
"epoch": 0.72,
|
2721 |
-
"learning_rate": 6.024252923343438e-05,
|
2722 |
-
"loss": 3.3563,
|
2723 |
-
"step": 3480
|
2724 |
-
},
|
2725 |
-
{
|
2726 |
-
"epoch": 0.72,
|
2727 |
-
"learning_rate": 5.985275010827198e-05,
|
2728 |
-
"loss": 3.358,
|
2729 |
-
"step": 3490
|
2730 |
-
},
|
2731 |
-
{
|
2732 |
-
"epoch": 0.72,
|
2733 |
-
"learning_rate": 5.9419662191424864e-05,
|
2734 |
-
"loss": 3.3907,
|
2735 |
-
"step": 3500
|
2736 |
-
},
|
2737 |
-
{
|
2738 |
-
"epoch": 0.72,
|
2739 |
-
"eval_accuracy": 0.33894296180010464,
|
2740 |
-
"eval_loss": 4.2109375,
|
2741 |
-
"eval_runtime": 6.6054,
|
2742 |
-
"eval_samples_per_second": 5.299,
|
2743 |
-
"eval_steps_per_second": 0.303,
|
2744 |
-
"step": 3500
|
2745 |
-
},
|
2746 |
-
{
|
2747 |
-
"epoch": 0.72,
|
2748 |
-
"learning_rate": 5.898657427457774e-05,
|
2749 |
-
"loss": 3.4023,
|
2750 |
-
"step": 3510
|
2751 |
-
},
|
2752 |
-
{
|
2753 |
-
"epoch": 0.72,
|
2754 |
-
"learning_rate": 5.8553486357730626e-05,
|
2755 |
-
"loss": 3.3519,
|
2756 |
-
"step": 3520
|
2757 |
-
},
|
2758 |
-
{
|
2759 |
-
"epoch": 0.73,
|
2760 |
-
"learning_rate": 5.81203984408835e-05,
|
2761 |
-
"loss": 3.3514,
|
2762 |
-
"step": 3530
|
2763 |
-
},
|
2764 |
-
{
|
2765 |
-
"epoch": 0.73,
|
2766 |
-
"learning_rate": 5.768731052403639e-05,
|
2767 |
-
"loss": 3.3151,
|
2768 |
-
"step": 3540
|
2769 |
-
},
|
2770 |
-
{
|
2771 |
-
"epoch": 0.73,
|
2772 |
-
"learning_rate": 5.7254222607189265e-05,
|
2773 |
-
"loss": 3.3929,
|
2774 |
-
"step": 3550
|
2775 |
-
},
|
2776 |
-
{
|
2777 |
-
"epoch": 0.73,
|
2778 |
-
"eval_accuracy": 0.33893598465027036,
|
2779 |
-
"eval_loss": 4.2109375,
|
2780 |
-
"eval_runtime": 6.5935,
|
2781 |
-
"eval_samples_per_second": 5.308,
|
2782 |
-
"eval_steps_per_second": 0.303,
|
2783 |
-
"step": 3550
|
2784 |
-
},
|
2785 |
-
{
|
2786 |
-
"epoch": 0.73,
|
2787 |
-
"learning_rate": 5.6821134690342135e-05,
|
2788 |
-
"loss": 3.3983,
|
2789 |
-
"step": 3560
|
2790 |
-
},
|
2791 |
-
{
|
2792 |
-
"epoch": 0.73,
|
2793 |
-
"learning_rate": 5.638804677349502e-05,
|
2794 |
-
"loss": 3.3281,
|
2795 |
-
"step": 3570
|
2796 |
-
},
|
2797 |
-
{
|
2798 |
-
"epoch": 0.74,
|
2799 |
-
"learning_rate": 5.59549588566479e-05,
|
2800 |
-
"loss": 3.3654,
|
2801 |
-
"step": 3580
|
2802 |
-
},
|
2803 |
-
{
|
2804 |
-
"epoch": 0.74,
|
2805 |
-
"learning_rate": 5.552187093980078e-05,
|
2806 |
-
"loss": 3.379,
|
2807 |
-
"step": 3590
|
2808 |
-
},
|
2809 |
-
{
|
2810 |
-
"epoch": 0.74,
|
2811 |
-
"learning_rate": 5.508878302295366e-05,
|
2812 |
-
"loss": 3.3738,
|
2813 |
-
"step": 3600
|
2814 |
-
},
|
2815 |
-
{
|
2816 |
-
"epoch": 0.74,
|
2817 |
-
"eval_accuracy": 0.33964067678353393,
|
2818 |
-
"eval_loss": 4.20703125,
|
2819 |
-
"eval_runtime": 6.6063,
|
2820 |
-
"eval_samples_per_second": 5.298,
|
2821 |
-
"eval_steps_per_second": 0.303,
|
2822 |
-
"step": 3600
|
2823 |
-
},
|
2824 |
-
{
|
2825 |
-
"epoch": 0.74,
|
2826 |
-
"learning_rate": 5.465569510610654e-05,
|
2827 |
-
"loss": 3.3727,
|
2828 |
-
"step": 3610
|
2829 |
-
},
|
2830 |
-
{
|
2831 |
-
"epoch": 0.74,
|
2832 |
-
"learning_rate": 5.426591598094414e-05,
|
2833 |
-
"loss": 3.3401,
|
2834 |
-
"step": 3620
|
2835 |
-
},
|
2836 |
-
{
|
2837 |
-
"epoch": 0.75,
|
2838 |
-
"learning_rate": 5.3832828064097017e-05,
|
2839 |
-
"loss": 3.3583,
|
2840 |
-
"step": 3630
|
2841 |
-
},
|
2842 |
-
{
|
2843 |
-
"epoch": 0.75,
|
2844 |
-
"learning_rate": 5.33997401472499e-05,
|
2845 |
-
"loss": 3.3868,
|
2846 |
-
"step": 3640
|
2847 |
-
},
|
2848 |
-
{
|
2849 |
-
"epoch": 0.75,
|
2850 |
-
"learning_rate": 5.296665223040277e-05,
|
2851 |
-
"loss": 3.3839,
|
2852 |
-
"step": 3650
|
2853 |
-
},
|
2854 |
-
{
|
2855 |
-
"epoch": 0.75,
|
2856 |
-
"eval_accuracy": 0.3392918192918193,
|
2857 |
-
"eval_loss": 4.20703125,
|
2858 |
-
"eval_runtime": 6.5952,
|
2859 |
-
"eval_samples_per_second": 5.307,
|
2860 |
-
"eval_steps_per_second": 0.303,
|
2861 |
-
"step": 3650
|
2862 |
-
},
|
2863 |
-
{
|
2864 |
-
"epoch": 0.75,
|
2865 |
-
"learning_rate": 5.253356431355565e-05,
|
2866 |
-
"loss": 3.3713,
|
2867 |
-
"step": 3660
|
2868 |
-
},
|
2869 |
-
{
|
2870 |
-
"epoch": 0.75,
|
2871 |
-
"learning_rate": 5.210047639670853e-05,
|
2872 |
-
"loss": 3.3649,
|
2873 |
-
"step": 3670
|
2874 |
-
},
|
2875 |
-
{
|
2876 |
-
"epoch": 0.76,
|
2877 |
-
"learning_rate": 5.166738847986141e-05,
|
2878 |
-
"loss": 3.3836,
|
2879 |
-
"step": 3680
|
2880 |
-
},
|
2881 |
-
{
|
2882 |
-
"epoch": 0.76,
|
2883 |
-
"learning_rate": 5.1234300563014294e-05,
|
2884 |
-
"loss": 3.3227,
|
2885 |
-
"step": 3690
|
2886 |
-
},
|
2887 |
-
{
|
2888 |
-
"epoch": 0.76,
|
2889 |
-
"learning_rate": 5.080121264616717e-05,
|
2890 |
-
"loss": 3.3854,
|
2891 |
-
"step": 3700
|
2892 |
-
},
|
2893 |
-
{
|
2894 |
-
"epoch": 0.76,
|
2895 |
-
"eval_accuracy": 0.3399546485260771,
|
2896 |
-
"eval_loss": 4.20703125,
|
2897 |
-
"eval_runtime": 6.6101,
|
2898 |
-
"eval_samples_per_second": 5.295,
|
2899 |
-
"eval_steps_per_second": 0.303,
|
2900 |
-
"step": 3700
|
2901 |
-
},
|
2902 |
-
{
|
2903 |
-
"epoch": 0.76,
|
2904 |
-
"learning_rate": 5.0368124729320056e-05,
|
2905 |
-
"loss": 3.3536,
|
2906 |
-
"step": 3710
|
2907 |
-
},
|
2908 |
-
{
|
2909 |
-
"epoch": 0.77,
|
2910 |
-
"learning_rate": 4.993503681247293e-05,
|
2911 |
-
"loss": 3.3943,
|
2912 |
-
"step": 3720
|
2913 |
-
},
|
2914 |
-
{
|
2915 |
-
"epoch": 0.77,
|
2916 |
-
"learning_rate": 4.950194889562582e-05,
|
2917 |
-
"loss": 3.3705,
|
2918 |
-
"step": 3730
|
2919 |
-
},
|
2920 |
-
{
|
2921 |
-
"epoch": 0.77,
|
2922 |
-
"learning_rate": 4.9068860978778694e-05,
|
2923 |
-
"loss": 3.3525,
|
2924 |
-
"step": 3740
|
2925 |
-
},
|
2926 |
-
{
|
2927 |
-
"epoch": 0.77,
|
2928 |
-
"learning_rate": 4.863577306193157e-05,
|
2929 |
-
"loss": 3.297,
|
2930 |
-
"step": 3750
|
2931 |
-
},
|
2932 |
-
{
|
2933 |
-
"epoch": 0.77,
|
2934 |
-
"eval_accuracy": 0.339236002093145,
|
2935 |
-
"eval_loss": 4.20703125,
|
2936 |
-
"eval_runtime": 6.5942,
|
2937 |
-
"eval_samples_per_second": 5.308,
|
2938 |
-
"eval_steps_per_second": 0.303,
|
2939 |
-
"step": 3750
|
2940 |
-
},
|
2941 |
-
{
|
2942 |
-
"epoch": 0.77,
|
2943 |
-
"learning_rate": 4.8202685145084456e-05,
|
2944 |
-
"loss": 3.3305,
|
2945 |
-
"step": 3760
|
2946 |
-
},
|
2947 |
-
{
|
2948 |
-
"epoch": 0.78,
|
2949 |
-
"learning_rate": 4.776959722823733e-05,
|
2950 |
-
"loss": 3.3789,
|
2951 |
-
"step": 3770
|
2952 |
-
},
|
2953 |
-
{
|
2954 |
-
"epoch": 0.78,
|
2955 |
-
"learning_rate": 4.733650931139022e-05,
|
2956 |
-
"loss": 3.4127,
|
2957 |
-
"step": 3780
|
2958 |
-
},
|
2959 |
-
{
|
2960 |
-
"epoch": 0.78,
|
2961 |
-
"learning_rate": 4.6903421394543095e-05,
|
2962 |
-
"loss": 3.3234,
|
2963 |
-
"step": 3790
|
2964 |
-
},
|
2965 |
-
{
|
2966 |
-
"epoch": 0.78,
|
2967 |
-
"learning_rate": 4.647033347769598e-05,
|
2968 |
-
"loss": 3.2951,
|
2969 |
-
"step": 3800
|
2970 |
-
},
|
2971 |
-
{
|
2972 |
-
"epoch": 0.78,
|
2973 |
-
"eval_accuracy": 0.3394732251875109,
|
2974 |
-
"eval_loss": 4.203125,
|
2975 |
-
"eval_runtime": 6.5899,
|
2976 |
-
"eval_samples_per_second": 5.311,
|
2977 |
-
"eval_steps_per_second": 0.303,
|
2978 |
-
"step": 3800
|
2979 |
-
},
|
2980 |
-
{
|
2981 |
-
"epoch": 0.78,
|
2982 |
-
"learning_rate": 4.603724556084885e-05,
|
2983 |
-
"loss": 3.3538,
|
2984 |
-
"step": 3810
|
2985 |
-
},
|
2986 |
-
{
|
2987 |
-
"epoch": 0.79,
|
2988 |
-
"learning_rate": 4.5604157644001733e-05,
|
2989 |
-
"loss": 3.4101,
|
2990 |
-
"step": 3820
|
2991 |
-
},
|
2992 |
-
{
|
2993 |
-
"epoch": 0.79,
|
2994 |
-
"learning_rate": 4.517106972715461e-05,
|
2995 |
-
"loss": 3.3232,
|
2996 |
-
"step": 3830
|
2997 |
-
},
|
2998 |
-
{
|
2999 |
-
"epoch": 0.79,
|
3000 |
-
"learning_rate": 4.4737981810307495e-05,
|
3001 |
-
"loss": 3.3519,
|
3002 |
-
"step": 3840
|
3003 |
-
},
|
3004 |
-
{
|
3005 |
-
"epoch": 0.79,
|
3006 |
-
"learning_rate": 4.430489389346037e-05,
|
3007 |
-
"loss": 3.3587,
|
3008 |
-
"step": 3850
|
3009 |
-
},
|
3010 |
-
{
|
3011 |
-
"epoch": 0.79,
|
3012 |
-
"eval_accuracy": 0.3401709401709402,
|
3013 |
-
"eval_loss": 4.19921875,
|
3014 |
-
"eval_runtime": 6.5915,
|
3015 |
-
"eval_samples_per_second": 5.31,
|
3016 |
-
"eval_steps_per_second": 0.303,
|
3017 |
-
"step": 3850
|
3018 |
-
},
|
3019 |
-
{
|
3020 |
-
"epoch": 0.79,
|
3021 |
-
"learning_rate": 4.3871805976613256e-05,
|
3022 |
-
"loss": 3.3519,
|
3023 |
-
"step": 3860
|
3024 |
-
},
|
3025 |
-
{
|
3026 |
-
"epoch": 0.8,
|
3027 |
-
"learning_rate": 4.3438718059766134e-05,
|
3028 |
-
"loss": 3.3635,
|
3029 |
-
"step": 3870
|
3030 |
-
},
|
3031 |
-
{
|
3032 |
-
"epoch": 0.8,
|
3033 |
-
"learning_rate": 4.300563014291901e-05,
|
3034 |
-
"loss": 3.4013,
|
3035 |
-
"step": 3880
|
3036 |
-
},
|
3037 |
-
{
|
3038 |
-
"epoch": 0.8,
|
3039 |
-
"learning_rate": 4.2572542226071895e-05,
|
3040 |
-
"loss": 3.3224,
|
3041 |
-
"step": 3890
|
3042 |
-
},
|
3043 |
-
{
|
3044 |
-
"epoch": 0.8,
|
3045 |
-
"learning_rate": 4.213945430922477e-05,
|
3046 |
-
"loss": 3.3237,
|
3047 |
-
"step": 3900
|
3048 |
-
},
|
3049 |
-
{
|
3050 |
-
"epoch": 0.8,
|
3051 |
-
"eval_accuracy": 0.33935461364032793,
|
3052 |
-
"eval_loss": 4.203125,
|
3053 |
-
"eval_runtime": 6.5894,
|
3054 |
-
"eval_samples_per_second": 5.312,
|
3055 |
-
"eval_steps_per_second": 0.304,
|
3056 |
-
"step": 3900
|
3057 |
-
},
|
3058 |
-
{
|
3059 |
-
"epoch": 0.8,
|
3060 |
-
"learning_rate": 4.170636639237766e-05,
|
3061 |
-
"loss": 3.3706,
|
3062 |
-
"step": 3910
|
3063 |
-
},
|
3064 |
-
{
|
3065 |
-
"epoch": 0.81,
|
3066 |
-
"learning_rate": 4.1273278475530534e-05,
|
3067 |
-
"loss": 3.3024,
|
3068 |
-
"step": 3920
|
3069 |
-
},
|
3070 |
-
{
|
3071 |
-
"epoch": 0.81,
|
3072 |
-
"learning_rate": 4.084019055868342e-05,
|
3073 |
-
"loss": 3.3717,
|
3074 |
-
"step": 3930
|
3075 |
-
},
|
3076 |
-
{
|
3077 |
-
"epoch": 0.81,
|
3078 |
-
"learning_rate": 4.0407102641836295e-05,
|
3079 |
-
"loss": 3.35,
|
3080 |
-
"step": 3940
|
3081 |
-
},
|
3082 |
-
{
|
3083 |
-
"epoch": 0.81,
|
3084 |
-
"learning_rate": 3.997401472498917e-05,
|
3085 |
-
"loss": 3.3136,
|
3086 |
-
"step": 3950
|
3087 |
-
},
|
3088 |
-
{
|
3089 |
-
"epoch": 0.81,
|
3090 |
-
"eval_accuracy": 0.3393267050409908,
|
3091 |
-
"eval_loss": 4.203125,
|
3092 |
-
"eval_runtime": 6.6033,
|
3093 |
-
"eval_samples_per_second": 5.3,
|
3094 |
-
"eval_steps_per_second": 0.303,
|
3095 |
-
"step": 3950
|
3096 |
-
},
|
3097 |
-
{
|
3098 |
-
"epoch": 0.81,
|
3099 |
-
"learning_rate": 3.954092680814206e-05,
|
3100 |
-
"loss": 3.3281,
|
3101 |
-
"step": 3960
|
3102 |
-
},
|
3103 |
-
{
|
3104 |
-
"epoch": 0.82,
|
3105 |
-
"learning_rate": 3.9107838891294934e-05,
|
3106 |
-
"loss": 3.2916,
|
3107 |
-
"step": 3970
|
3108 |
-
},
|
3109 |
-
{
|
3110 |
-
"epoch": 0.82,
|
3111 |
-
"learning_rate": 3.867475097444782e-05,
|
3112 |
-
"loss": 3.3578,
|
3113 |
-
"step": 3980
|
3114 |
-
},
|
3115 |
-
{
|
3116 |
-
"epoch": 0.82,
|
3117 |
-
"learning_rate": 3.8241663057600696e-05,
|
3118 |
-
"loss": 3.3789,
|
3119 |
-
"step": 3990
|
3120 |
-
},
|
3121 |
-
{
|
3122 |
-
"epoch": 0.82,
|
3123 |
-
"learning_rate": 3.780857514075358e-05,
|
3124 |
-
"loss": 3.3367,
|
3125 |
-
"step": 4000
|
3126 |
-
},
|
3127 |
-
{
|
3128 |
-
"epoch": 0.82,
|
3129 |
-
"eval_accuracy": 0.33942438513867085,
|
3130 |
-
"eval_loss": 4.203125,
|
3131 |
-
"eval_runtime": 6.5852,
|
3132 |
-
"eval_samples_per_second": 5.315,
|
3133 |
-
"eval_steps_per_second": 0.304,
|
3134 |
-
"step": 4000
|
3135 |
-
}
|
3136 |
-
],
|
3137 |
-
"max_steps": 4862,
|
3138 |
-
"num_train_epochs": 1,
|
3139 |
-
"total_flos": 4.250832212066304e+19,
|
3140 |
-
"trial_name": null,
|
3141 |
-
"trial_params": null
|
3142 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-4000/training_args.bin
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:8882a918bcb9a315dfcc77f1a3a879de69b8b68f463ac4b04ba7f0c085598f2d
|
3 |
-
size 5115
|
|
|
|
|
|
|
|
checkpoint-4000/zero_to_fp32.py
DELETED
@@ -1,578 +0,0 @@
|
|
1 |
-
#!/usr/bin/env python
|
2 |
-
|
3 |
-
# Copyright (c) Microsoft Corporation.
|
4 |
-
# SPDX-License-Identifier: Apache-2.0
|
5 |
-
|
6 |
-
# DeepSpeed Team
|
7 |
-
|
8 |
-
# This script extracts fp32 consolidated weights from a zero 2 and 3 DeepSpeed checkpoints. It gets
|
9 |
-
# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
|
10 |
-
# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
|
11 |
-
# application.
|
12 |
-
#
|
13 |
-
# example: python zero_to_fp32.py . pytorch_model.bin
|
14 |
-
|
15 |
-
import argparse
|
16 |
-
import torch
|
17 |
-
import glob
|
18 |
-
import math
|
19 |
-
import os
|
20 |
-
import re
|
21 |
-
from collections import OrderedDict
|
22 |
-
from dataclasses import dataclass
|
23 |
-
|
24 |
-
# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
|
25 |
-
# DeepSpeed data structures it has to be available in the current python environment.
|
26 |
-
from deepspeed.utils import logger
|
27 |
-
from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
|
28 |
-
FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
|
29 |
-
FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
|
30 |
-
|
31 |
-
|
32 |
-
@dataclass
|
33 |
-
class zero_model_state:
|
34 |
-
buffers: dict()
|
35 |
-
param_shapes: dict()
|
36 |
-
shared_params: list
|
37 |
-
ds_version: int
|
38 |
-
frozen_param_shapes: dict()
|
39 |
-
frozen_param_fragments: dict()
|
40 |
-
|
41 |
-
|
42 |
-
debug = 0
|
43 |
-
|
44 |
-
# load to cpu
|
45 |
-
device = torch.device('cpu')
|
46 |
-
|
47 |
-
|
48 |
-
def atoi(text):
|
49 |
-
return int(text) if text.isdigit() else text
|
50 |
-
|
51 |
-
|
52 |
-
def natural_keys(text):
|
53 |
-
'''
|
54 |
-
alist.sort(key=natural_keys) sorts in human order
|
55 |
-
http://nedbatchelder.com/blog/200712/human_sorting.html
|
56 |
-
(See Toothy's implementation in the comments)
|
57 |
-
'''
|
58 |
-
return [atoi(c) for c in re.split(r'(\d+)', text)]
|
59 |
-
|
60 |
-
|
61 |
-
def get_model_state_file(checkpoint_dir, zero_stage):
|
62 |
-
if not os.path.isdir(checkpoint_dir):
|
63 |
-
raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
|
64 |
-
|
65 |
-
# there should be only one file
|
66 |
-
if zero_stage == 2:
|
67 |
-
file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
|
68 |
-
elif zero_stage == 3:
|
69 |
-
file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
|
70 |
-
|
71 |
-
if not os.path.exists(file):
|
72 |
-
raise FileNotFoundError(f"can't find model states file at '{file}'")
|
73 |
-
|
74 |
-
return file
|
75 |
-
|
76 |
-
|
77 |
-
def get_checkpoint_files(checkpoint_dir, glob_pattern):
|
78 |
-
# XXX: need to test that this simple glob rule works for multi-node setup too
|
79 |
-
ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
|
80 |
-
|
81 |
-
if len(ckpt_files) == 0:
|
82 |
-
raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
|
83 |
-
|
84 |
-
return ckpt_files
|
85 |
-
|
86 |
-
|
87 |
-
def get_optim_files(checkpoint_dir):
|
88 |
-
return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
|
89 |
-
|
90 |
-
|
91 |
-
def get_model_state_files(checkpoint_dir):
|
92 |
-
return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
|
93 |
-
|
94 |
-
|
95 |
-
def parse_model_states(files):
|
96 |
-
zero_model_states = []
|
97 |
-
for file in files:
|
98 |
-
state_dict = torch.load(file, map_location=device)
|
99 |
-
|
100 |
-
if BUFFER_NAMES not in state_dict:
|
101 |
-
raise ValueError(f"{file} is not a model state checkpoint")
|
102 |
-
buffer_names = state_dict[BUFFER_NAMES]
|
103 |
-
if debug:
|
104 |
-
print("Found buffers:", buffer_names)
|
105 |
-
|
106 |
-
# recover just the buffers while restoring them to fp32 if they were saved in fp16
|
107 |
-
buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
|
108 |
-
param_shapes = state_dict[PARAM_SHAPES]
|
109 |
-
|
110 |
-
# collect parameters that are included in param_shapes
|
111 |
-
param_names = []
|
112 |
-
for s in param_shapes:
|
113 |
-
for name in s.keys():
|
114 |
-
param_names.append(name)
|
115 |
-
|
116 |
-
# update with frozen parameters
|
117 |
-
frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
|
118 |
-
if frozen_param_shapes is not None:
|
119 |
-
if debug:
|
120 |
-
print(f"Found frozen_param_shapes: {frozen_param_shapes}")
|
121 |
-
param_names += list(frozen_param_shapes.keys())
|
122 |
-
|
123 |
-
# handle shared params
|
124 |
-
shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
|
125 |
-
|
126 |
-
ds_version = state_dict.get(DS_VERSION, None)
|
127 |
-
|
128 |
-
frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
|
129 |
-
|
130 |
-
z_model_state = zero_model_state(buffers=buffers,
|
131 |
-
param_shapes=param_shapes,
|
132 |
-
shared_params=shared_params,
|
133 |
-
ds_version=ds_version,
|
134 |
-
frozen_param_shapes=frozen_param_shapes,
|
135 |
-
frozen_param_fragments=frozen_param_fragments)
|
136 |
-
zero_model_states.append(z_model_state)
|
137 |
-
|
138 |
-
return zero_model_states
|
139 |
-
|
140 |
-
|
141 |
-
def parse_optim_states(files, ds_checkpoint_dir):
|
142 |
-
|
143 |
-
total_files = len(files)
|
144 |
-
state_dicts = []
|
145 |
-
for f in files:
|
146 |
-
state_dicts.append(torch.load(f, map_location=device))
|
147 |
-
|
148 |
-
if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
|
149 |
-
raise ValueError(f"{files[0]} is not a zero checkpoint")
|
150 |
-
zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
|
151 |
-
world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
|
152 |
-
|
153 |
-
# For ZeRO-2 each param group can have different partition_count as data parallelism for expert
|
154 |
-
# parameters can be different from data parallelism for non-expert parameters. So we can just
|
155 |
-
# use the max of the partition_count to get the dp world_size.
|
156 |
-
|
157 |
-
if type(world_size) is list:
|
158 |
-
world_size = max(world_size)
|
159 |
-
|
160 |
-
if world_size != total_files:
|
161 |
-
raise ValueError(
|
162 |
-
f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
|
163 |
-
"Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
|
164 |
-
)
|
165 |
-
|
166 |
-
# the groups are named differently in each stage
|
167 |
-
if zero_stage == 2:
|
168 |
-
fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
|
169 |
-
elif zero_stage == 3:
|
170 |
-
fp32_groups_key = FP32_FLAT_GROUPS
|
171 |
-
else:
|
172 |
-
raise ValueError(f"unknown zero stage {zero_stage}")
|
173 |
-
|
174 |
-
if zero_stage == 2:
|
175 |
-
fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
|
176 |
-
elif zero_stage == 3:
|
177 |
-
# if there is more than one param group, there will be multiple flattened tensors - one
|
178 |
-
# flattened tensor per group - for simplicity merge them into a single tensor
|
179 |
-
#
|
180 |
-
# XXX: could make the script more memory efficient for when there are multiple groups - it
|
181 |
-
# will require matching the sub-lists of param_shapes for each param group flattened tensor
|
182 |
-
|
183 |
-
fp32_flat_groups = [
|
184 |
-
torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
|
185 |
-
]
|
186 |
-
|
187 |
-
return zero_stage, world_size, fp32_flat_groups
|
188 |
-
|
189 |
-
|
190 |
-
def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir):
|
191 |
-
"""
|
192 |
-
Returns fp32 state_dict reconstructed from ds checkpoint
|
193 |
-
|
194 |
-
Args:
|
195 |
-
- ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
|
196 |
-
|
197 |
-
"""
|
198 |
-
print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
|
199 |
-
|
200 |
-
optim_files = get_optim_files(ds_checkpoint_dir)
|
201 |
-
zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
|
202 |
-
print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
|
203 |
-
|
204 |
-
model_files = get_model_state_files(ds_checkpoint_dir)
|
205 |
-
|
206 |
-
zero_model_states = parse_model_states(model_files)
|
207 |
-
print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
|
208 |
-
|
209 |
-
if zero_stage == 2:
|
210 |
-
return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states)
|
211 |
-
elif zero_stage == 3:
|
212 |
-
return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states)
|
213 |
-
|
214 |
-
|
215 |
-
def _zero2_merge_frozen_params(state_dict, zero_model_states):
|
216 |
-
if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
|
217 |
-
return
|
218 |
-
|
219 |
-
frozen_param_shapes = zero_model_states[0].frozen_param_shapes
|
220 |
-
frozen_param_fragments = zero_model_states[0].frozen_param_fragments
|
221 |
-
|
222 |
-
if debug:
|
223 |
-
num_elem = sum(s.numel() for s in frozen_param_shapes.values())
|
224 |
-
print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
|
225 |
-
|
226 |
-
wanted_params = len(frozen_param_shapes)
|
227 |
-
wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
|
228 |
-
avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
|
229 |
-
print(f'Frozen params: Have {avail_numel} numels to process.')
|
230 |
-
print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
|
231 |
-
|
232 |
-
total_params = 0
|
233 |
-
total_numel = 0
|
234 |
-
for name, shape in frozen_param_shapes.items():
|
235 |
-
total_params += 1
|
236 |
-
unpartitioned_numel = shape.numel()
|
237 |
-
total_numel += unpartitioned_numel
|
238 |
-
|
239 |
-
state_dict[name] = frozen_param_fragments[name]
|
240 |
-
|
241 |
-
if debug:
|
242 |
-
print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
|
243 |
-
|
244 |
-
print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
|
245 |
-
|
246 |
-
|
247 |
-
def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
|
248 |
-
param_shapes = zero_model_states[0].param_shapes
|
249 |
-
|
250 |
-
# Reconstruction protocol:
|
251 |
-
#
|
252 |
-
# XXX: document this
|
253 |
-
|
254 |
-
if debug:
|
255 |
-
for i in range(world_size):
|
256 |
-
for j in range(len(fp32_flat_groups[0])):
|
257 |
-
print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
|
258 |
-
|
259 |
-
# XXX: memory usage doubles here (zero2)
|
260 |
-
num_param_groups = len(fp32_flat_groups[0])
|
261 |
-
merged_single_partition_of_fp32_groups = []
|
262 |
-
for i in range(num_param_groups):
|
263 |
-
merged_partitions = [sd[i] for sd in fp32_flat_groups]
|
264 |
-
full_single_fp32_vector = torch.cat(merged_partitions, 0)
|
265 |
-
merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
|
266 |
-
avail_numel = sum(
|
267 |
-
[full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
|
268 |
-
|
269 |
-
if debug:
|
270 |
-
wanted_params = sum([len(shapes) for shapes in param_shapes])
|
271 |
-
wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
|
272 |
-
# not asserting if there is a mismatch due to possible padding
|
273 |
-
print(f"Have {avail_numel} numels to process.")
|
274 |
-
print(f"Need {wanted_numel} numels in {wanted_params} params.")
|
275 |
-
|
276 |
-
# params
|
277 |
-
# XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
|
278 |
-
# out-of-core computing solution
|
279 |
-
total_numel = 0
|
280 |
-
total_params = 0
|
281 |
-
for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
|
282 |
-
offset = 0
|
283 |
-
avail_numel = full_single_fp32_vector.numel()
|
284 |
-
for name, shape in shapes.items():
|
285 |
-
|
286 |
-
unpartitioned_numel = shape.numel()
|
287 |
-
total_numel += unpartitioned_numel
|
288 |
-
total_params += 1
|
289 |
-
|
290 |
-
if debug:
|
291 |
-
print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
|
292 |
-
state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
|
293 |
-
offset += unpartitioned_numel
|
294 |
-
|
295 |
-
# Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
|
296 |
-
# avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
|
297 |
-
# paddings performed in the code it's almost impossible to predict the exact numbers w/o the
|
298 |
-
# live optimizer object, so we are checking that the numbers are within the right range
|
299 |
-
align_to = 2 * world_size
|
300 |
-
|
301 |
-
def zero2_align(x):
|
302 |
-
return align_to * math.ceil(x / align_to)
|
303 |
-
|
304 |
-
if debug:
|
305 |
-
print(f"original offset={offset}, avail_numel={avail_numel}")
|
306 |
-
|
307 |
-
offset = zero2_align(offset)
|
308 |
-
avail_numel = zero2_align(avail_numel)
|
309 |
-
|
310 |
-
if debug:
|
311 |
-
print(f"aligned offset={offset}, avail_numel={avail_numel}")
|
312 |
-
|
313 |
-
# Sanity check
|
314 |
-
if offset != avail_numel:
|
315 |
-
raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
|
316 |
-
|
317 |
-
print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
|
318 |
-
|
319 |
-
|
320 |
-
def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states):
|
321 |
-
state_dict = OrderedDict()
|
322 |
-
|
323 |
-
# buffers
|
324 |
-
buffers = zero_model_states[0].buffers
|
325 |
-
state_dict.update(buffers)
|
326 |
-
if debug:
|
327 |
-
print(f"added {len(buffers)} buffers")
|
328 |
-
|
329 |
-
_zero2_merge_frozen_params(state_dict, zero_model_states)
|
330 |
-
|
331 |
-
_zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
|
332 |
-
|
333 |
-
# recover shared parameters
|
334 |
-
for pair in zero_model_states[0].shared_params:
|
335 |
-
if pair[1] in state_dict:
|
336 |
-
state_dict[pair[0]] = state_dict[pair[1]]
|
337 |
-
|
338 |
-
return state_dict
|
339 |
-
|
340 |
-
|
341 |
-
def zero3_partitioned_param_info(unpartitioned_numel, world_size):
|
342 |
-
remainder = unpartitioned_numel % world_size
|
343 |
-
padding_numel = (world_size - remainder) if remainder else 0
|
344 |
-
partitioned_numel = math.ceil(unpartitioned_numel / world_size)
|
345 |
-
return partitioned_numel, padding_numel
|
346 |
-
|
347 |
-
|
348 |
-
def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
|
349 |
-
if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
|
350 |
-
return
|
351 |
-
|
352 |
-
if debug:
|
353 |
-
for i in range(world_size):
|
354 |
-
num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
|
355 |
-
print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
|
356 |
-
|
357 |
-
frozen_param_shapes = zero_model_states[0].frozen_param_shapes
|
358 |
-
wanted_params = len(frozen_param_shapes)
|
359 |
-
wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
|
360 |
-
avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
|
361 |
-
print(f'Frozen params: Have {avail_numel} numels to process.')
|
362 |
-
print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
|
363 |
-
|
364 |
-
total_params = 0
|
365 |
-
total_numel = 0
|
366 |
-
for name, shape in zero_model_states[0].frozen_param_shapes.items():
|
367 |
-
total_params += 1
|
368 |
-
unpartitioned_numel = shape.numel()
|
369 |
-
total_numel += unpartitioned_numel
|
370 |
-
|
371 |
-
param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
|
372 |
-
state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
|
373 |
-
|
374 |
-
partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
|
375 |
-
|
376 |
-
if debug:
|
377 |
-
print(
|
378 |
-
f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
|
379 |
-
)
|
380 |
-
|
381 |
-
print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
|
382 |
-
|
383 |
-
|
384 |
-
def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
|
385 |
-
param_shapes = zero_model_states[0].param_shapes
|
386 |
-
avail_numel = fp32_flat_groups[0].numel() * world_size
|
387 |
-
# Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
|
388 |
-
# param, re-consolidating each param, while dealing with padding if any
|
389 |
-
|
390 |
-
# merge list of dicts, preserving order
|
391 |
-
param_shapes = {k: v for d in param_shapes for k, v in d.items()}
|
392 |
-
|
393 |
-
if debug:
|
394 |
-
for i in range(world_size):
|
395 |
-
print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
|
396 |
-
|
397 |
-
wanted_params = len(param_shapes)
|
398 |
-
wanted_numel = sum(shape.numel() for shape in param_shapes.values())
|
399 |
-
# not asserting if there is a mismatch due to possible padding
|
400 |
-
avail_numel = fp32_flat_groups[0].numel() * world_size
|
401 |
-
print(f"Trainable params: Have {avail_numel} numels to process.")
|
402 |
-
print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
|
403 |
-
|
404 |
-
# params
|
405 |
-
# XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
|
406 |
-
# out-of-core computing solution
|
407 |
-
offset = 0
|
408 |
-
total_numel = 0
|
409 |
-
total_params = 0
|
410 |
-
for name, shape in param_shapes.items():
|
411 |
-
|
412 |
-
unpartitioned_numel = shape.numel()
|
413 |
-
total_numel += unpartitioned_numel
|
414 |
-
total_params += 1
|
415 |
-
|
416 |
-
partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
|
417 |
-
|
418 |
-
if debug:
|
419 |
-
print(
|
420 |
-
f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
|
421 |
-
)
|
422 |
-
|
423 |
-
# XXX: memory usage doubles here
|
424 |
-
state_dict[name] = torch.cat(
|
425 |
-
tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
|
426 |
-
0).narrow(0, 0, unpartitioned_numel).view(shape)
|
427 |
-
offset += partitioned_numel
|
428 |
-
|
429 |
-
offset *= world_size
|
430 |
-
|
431 |
-
# Sanity check
|
432 |
-
if offset != avail_numel:
|
433 |
-
raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
|
434 |
-
|
435 |
-
print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
|
436 |
-
|
437 |
-
|
438 |
-
def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states):
|
439 |
-
state_dict = OrderedDict()
|
440 |
-
|
441 |
-
# buffers
|
442 |
-
buffers = zero_model_states[0].buffers
|
443 |
-
state_dict.update(buffers)
|
444 |
-
if debug:
|
445 |
-
print(f"added {len(buffers)} buffers")
|
446 |
-
|
447 |
-
_zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
|
448 |
-
|
449 |
-
_zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
|
450 |
-
|
451 |
-
# recover shared parameters
|
452 |
-
for pair in zero_model_states[0].shared_params:
|
453 |
-
if pair[1] in state_dict:
|
454 |
-
state_dict[pair[0]] = state_dict[pair[1]]
|
455 |
-
|
456 |
-
return state_dict
|
457 |
-
|
458 |
-
|
459 |
-
def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None):
|
460 |
-
"""
|
461 |
-
Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
|
462 |
-
``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
|
463 |
-
via a model hub.
|
464 |
-
|
465 |
-
Args:
|
466 |
-
- ``checkpoint_dir``: path to the desired checkpoint folder
|
467 |
-
- ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
|
468 |
-
|
469 |
-
Returns:
|
470 |
-
- pytorch ``state_dict``
|
471 |
-
|
472 |
-
Note: this approach may not work if your application doesn't have sufficient free CPU memory and
|
473 |
-
you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
|
474 |
-
the checkpoint.
|
475 |
-
|
476 |
-
A typical usage might be ::
|
477 |
-
|
478 |
-
from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
|
479 |
-
# do the training and checkpoint saving
|
480 |
-
state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
|
481 |
-
model = model.cpu() # move to cpu
|
482 |
-
model.load_state_dict(state_dict)
|
483 |
-
# submit to model hub or save the model to share with others
|
484 |
-
|
485 |
-
In this example the ``model`` will no longer be usable in the deepspeed context of the same
|
486 |
-
application. i.e. you will need to re-initialize the deepspeed engine, since
|
487 |
-
``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
|
488 |
-
|
489 |
-
If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
|
490 |
-
|
491 |
-
"""
|
492 |
-
if tag is None:
|
493 |
-
latest_path = os.path.join(checkpoint_dir, 'latest')
|
494 |
-
if os.path.isfile(latest_path):
|
495 |
-
with open(latest_path, 'r') as fd:
|
496 |
-
tag = fd.read().strip()
|
497 |
-
else:
|
498 |
-
raise ValueError(f"Unable to find 'latest' file at {latest_path}")
|
499 |
-
|
500 |
-
ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
|
501 |
-
|
502 |
-
if not os.path.isdir(ds_checkpoint_dir):
|
503 |
-
raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
|
504 |
-
|
505 |
-
return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir)
|
506 |
-
|
507 |
-
|
508 |
-
def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None):
|
509 |
-
"""
|
510 |
-
Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
|
511 |
-
loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
|
512 |
-
|
513 |
-
Args:
|
514 |
-
- ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
|
515 |
-
- ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
|
516 |
-
- ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
|
517 |
-
"""
|
518 |
-
|
519 |
-
state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
|
520 |
-
print(f"Saving fp32 state dict to {output_file}")
|
521 |
-
torch.save(state_dict, output_file)
|
522 |
-
|
523 |
-
|
524 |
-
def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
|
525 |
-
"""
|
526 |
-
1. Put the provided model to cpu
|
527 |
-
2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
|
528 |
-
3. Load it into the provided model
|
529 |
-
|
530 |
-
Args:
|
531 |
-
- ``model``: the model object to update
|
532 |
-
- ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
|
533 |
-
- ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
|
534 |
-
|
535 |
-
Returns:
|
536 |
-
- ``model`: modified model
|
537 |
-
|
538 |
-
Make sure you have plenty of CPU memory available before you call this function. If you don't
|
539 |
-
have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
|
540 |
-
conveniently placed for you in the checkpoint folder.
|
541 |
-
|
542 |
-
A typical usage might be ::
|
543 |
-
|
544 |
-
from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
|
545 |
-
model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
|
546 |
-
# submit to model hub or save the model to share with others
|
547 |
-
|
548 |
-
Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
|
549 |
-
of the same application. i.e. you will need to re-initialize the deepspeed engine, since
|
550 |
-
``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
|
551 |
-
|
552 |
-
"""
|
553 |
-
logger.info(f"Extracting fp32 weights")
|
554 |
-
state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
|
555 |
-
|
556 |
-
logger.info(f"Overwriting model with fp32 weights")
|
557 |
-
model = model.cpu()
|
558 |
-
model.load_state_dict(state_dict, strict=False)
|
559 |
-
|
560 |
-
return model
|
561 |
-
|
562 |
-
|
563 |
-
if __name__ == "__main__":
|
564 |
-
|
565 |
-
parser = argparse.ArgumentParser()
|
566 |
-
parser.add_argument("checkpoint_dir",
|
567 |
-
type=str,
|
568 |
-
help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
|
569 |
-
parser.add_argument(
|
570 |
-
"output_file",
|
571 |
-
type=str,
|
572 |
-
help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
|
573 |
-
parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
|
574 |
-
args = parser.parse_args()
|
575 |
-
|
576 |
-
debug = args.debug
|
577 |
-
|
578 |
-
convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, args.output_file)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|