add baseline model checkpoints
Browse filesThis view is limited to 50 files because it contains too many changes. Β
See raw diff
- .gitattributes +10 -0
- {archived β function-base-flatten}/checkpoint-10000/config.json +2 -1
- generation_config.json β function-base-flatten/checkpoint-10000/generation_config.json +1 -1
- optimizer.pt β function-base-flatten/checkpoint-10000/optimizer.pt +2 -2
- pytorch_model.bin β function-base-flatten/checkpoint-10000/pytorch_model.bin +2 -2
- {archived β function-base-flatten}/checkpoint-10000/rng_state.pth +1 -1
- scheduler.pt β function-base-flatten/checkpoint-10000/scheduler.pt +1 -1
- {archived β function-base-flatten}/checkpoint-10000/trainer_state.json +133 -130
- function-base-flatten/checkpoint-10000/training_args.bin +3 -0
- config.json β function-base-flatten/checkpoint-15000/config.json +2 -1
- {archived/checkpoint-10000 β function-base-flatten/checkpoint-15000}/generation_config.json +1 -1
- {archived/checkpoint-10000 β function-base-flatten/checkpoint-15000}/optimizer.pt +2 -2
- {archived/checkpoint-10000 β function-base-flatten/checkpoint-15000}/pytorch_model.bin +2 -2
- rng_state.pth β function-base-flatten/checkpoint-15000/rng_state.pth +1 -1
- {archived/checkpoint-10000 β function-base-flatten/checkpoint-15000}/scheduler.pt +1 -1
- function-base-flatten/checkpoint-15000/trainer_state.json +439 -0
- function-base-flatten/checkpoint-15000/training_args.bin +3 -0
- function-base-flatten/checkpoint-20000/config.json +62 -0
- function-base-flatten/checkpoint-20000/generation_config.json +7 -0
- function-base-flatten/checkpoint-20000/optimizer.pt +3 -0
- function-base-flatten/checkpoint-20000/pytorch_model.bin +3 -0
- function-base-flatten/checkpoint-20000/rng_state.pth +3 -0
- function-base-flatten/checkpoint-20000/scheduler.pt +3 -0
- function-base-flatten/checkpoint-20000/trainer_state.json +579 -0
- function-base-flatten/checkpoint-20000/training_args.bin +3 -0
- function-base-flatten/checkpoint-5000/config.json +62 -0
- function-base-flatten/checkpoint-5000/generation_config.json +7 -0
- function-base-flatten/checkpoint-5000/optimizer.pt +3 -0
- function-base-flatten/checkpoint-5000/pytorch_model.bin +3 -0
- function-base-flatten/checkpoint-5000/rng_state.pth +3 -0
- function-base-flatten/checkpoint-5000/scheduler.pt +3 -0
- function-base-flatten/checkpoint-5000/trainer_state.json +159 -0
- function-base-flatten/checkpoint-5000/training_args.bin +3 -0
- function-base/checkpoint-10000/config.json +62 -0
- function-base/checkpoint-10000/generation_config.json +7 -0
- function-base/checkpoint-10000/optimizer.pt +3 -0
- function-base/checkpoint-10000/pytorch_model.bin +3 -0
- function-base/checkpoint-10000/rng_state.pth +3 -0
- function-base/checkpoint-10000/scheduler.pt +3 -0
- function-base/checkpoint-10000/trainer_state.json +299 -0
- training_args.bin β function-base/checkpoint-10000/training_args.bin +1 -1
- function-base/checkpoint-15000/config.json +62 -0
- function-base/checkpoint-15000/generation_config.json +7 -0
- function-base/checkpoint-15000/optimizer.pt +3 -0
- function-base/checkpoint-15000/pytorch_model.bin +3 -0
- function-base/checkpoint-15000/rng_state.pth +3 -0
- function-base/checkpoint-15000/scheduler.pt +3 -0
- function-base/checkpoint-15000/trainer_state.json +439 -0
- {archived/checkpoint-10000 β function-base/checkpoint-15000}/training_args.bin +1 -1
- function-base/checkpoint-20000/config.json +62 -0
.gitattributes
CHANGED
@@ -37,3 +37,13 @@ checkpoint-10000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
|
|
37 |
checkpoint-10000/training_args.bin filter=lfs diff=lfs merge=lfs -text
|
38 |
pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
|
39 |
**/**/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
37 |
checkpoint-10000/training_args.bin filter=lfs diff=lfs merge=lfs -text
|
38 |
pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
|
39 |
**/**/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
|
40 |
+
*/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
|
41 |
+
function-base/checkpoint-10000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
|
42 |
+
function-base/checkpoint-15000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
|
43 |
+
function-base/checkpoint-20000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
|
44 |
+
function-base/checkpoint-5000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
|
45 |
+
**/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
|
46 |
+
function-base-flatten/checkpoint-10000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
|
47 |
+
function-base-flatten/checkpoint-15000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
|
48 |
+
function-base-flatten/checkpoint-20000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
|
49 |
+
function-base-flatten/checkpoint-5000/pytorch_model.bin filter=lfs diff=lfs merge=lfs -text
|
{archived β function-base-flatten}/checkpoint-10000/config.json
RENAMED
@@ -3,6 +3,7 @@
|
|
3 |
"architectures": [
|
4 |
"FiDT5"
|
5 |
],
|
|
|
6 |
"d_ff": 2048,
|
7 |
"d_kv": 64,
|
8 |
"d_model": 768,
|
@@ -55,7 +56,7 @@
|
|
55 |
},
|
56 |
"tie_word_embeddings": false,
|
57 |
"torch_dtype": "float32",
|
58 |
-
"transformers_version": "4.
|
59 |
"use_cache": true,
|
60 |
"vocab_size": 32128
|
61 |
}
|
|
|
3 |
"architectures": [
|
4 |
"FiDT5"
|
5 |
],
|
6 |
+
"classifier_dropout": 0.0,
|
7 |
"d_ff": 2048,
|
8 |
"d_kv": 64,
|
9 |
"d_model": 768,
|
|
|
56 |
},
|
57 |
"tie_word_embeddings": false,
|
58 |
"torch_dtype": "float32",
|
59 |
+
"transformers_version": "4.33.1",
|
60 |
"use_cache": true,
|
61 |
"vocab_size": 32128
|
62 |
}
|
generation_config.json β function-base-flatten/checkpoint-10000/generation_config.json
RENAMED
@@ -3,5 +3,5 @@
|
|
3 |
"decoder_start_token_id": 0,
|
4 |
"eos_token_id": 1,
|
5 |
"pad_token_id": 0,
|
6 |
-
"transformers_version": "4.
|
7 |
}
|
|
|
3 |
"decoder_start_token_id": 0,
|
4 |
"eos_token_id": 1,
|
5 |
"pad_token_id": 0,
|
6 |
+
"transformers_version": "4.33.1"
|
7 |
}
|
optimizer.pt β function-base-flatten/checkpoint-10000/optimizer.pt
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:02862daee9f5279b5eb8cd89e34c2669c673cdf1d59a98c134566a9b2e31be64
|
3 |
+
size 2372293
|
pytorch_model.bin β function-base-flatten/checkpoint-10000/pytorch_model.bin
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:298486418d9e1499d63285dd7ef1ea2dc836d8fcb2372dba48dac0d9bd8acf4e
|
3 |
+
size 990410745
|
{archived β function-base-flatten}/checkpoint-10000/rng_state.pth
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14575
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7ffe83d5f6e5972b42f5e84134778b78cd2b53ac4084db95835b17046131594e
|
3 |
size 14575
|
scheduler.pt β function-base-flatten/checkpoint-10000/scheduler.pt
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 627
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5ca84735770f2a357642bb07a692869b77f0409ed01b19bf3c890697e28d947e
|
3 |
size 627
|
{archived β function-base-flatten}/checkpoint-10000/trainer_state.json
RENAMED
@@ -1,7 +1,8 @@
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
-
"epoch": 1.
|
|
|
5 |
"global_step": 10000,
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
@@ -9,288 +10,290 @@
|
|
9 |
"log_history": [
|
10 |
{
|
11 |
"epoch": 0.06,
|
12 |
-
"learning_rate":
|
13 |
-
"loss":
|
14 |
"step": 500
|
15 |
},
|
16 |
{
|
17 |
"epoch": 0.06,
|
18 |
-
"eval_loss": 0.
|
19 |
-
"eval_runtime":
|
20 |
-
"eval_samples_per_second":
|
21 |
-
"eval_steps_per_second":
|
22 |
"step": 500
|
23 |
},
|
24 |
{
|
25 |
"epoch": 0.13,
|
26 |
-
"learning_rate":
|
27 |
-
"loss": 0.
|
28 |
"step": 1000
|
29 |
},
|
30 |
{
|
31 |
"epoch": 0.13,
|
32 |
-
"eval_loss": 0.
|
33 |
-
"eval_runtime":
|
34 |
-
"eval_samples_per_second":
|
35 |
-
"eval_steps_per_second":
|
36 |
"step": 1000
|
37 |
},
|
38 |
{
|
39 |
"epoch": 0.19,
|
40 |
-
"learning_rate":
|
41 |
-
"loss": 0.
|
42 |
"step": 1500
|
43 |
},
|
44 |
{
|
45 |
"epoch": 0.19,
|
46 |
-
"eval_loss": 0.
|
47 |
-
"eval_runtime":
|
48 |
-
"eval_samples_per_second":
|
49 |
-
"eval_steps_per_second":
|
50 |
"step": 1500
|
51 |
},
|
52 |
{
|
53 |
"epoch": 0.25,
|
54 |
-
"learning_rate":
|
55 |
-
"loss": 0.
|
56 |
"step": 2000
|
57 |
},
|
58 |
{
|
59 |
"epoch": 0.25,
|
60 |
-
"eval_loss": 0.
|
61 |
-
"eval_runtime":
|
62 |
-
"eval_samples_per_second":
|
63 |
-
"eval_steps_per_second":
|
64 |
"step": 2000
|
65 |
},
|
66 |
{
|
67 |
-
"epoch": 0.
|
68 |
-
"learning_rate":
|
69 |
-
"loss": 0.
|
70 |
"step": 2500
|
71 |
},
|
72 |
{
|
73 |
-
"epoch": 0.
|
74 |
-
"eval_loss": 0.
|
75 |
-
"eval_runtime":
|
76 |
-
"eval_samples_per_second":
|
77 |
-
"eval_steps_per_second":
|
78 |
"step": 2500
|
79 |
},
|
80 |
{
|
81 |
"epoch": 0.38,
|
82 |
-
"learning_rate":
|
83 |
-
"loss": 0.
|
84 |
"step": 3000
|
85 |
},
|
86 |
{
|
87 |
"epoch": 0.38,
|
88 |
-
"eval_loss": 0.
|
89 |
-
"eval_runtime":
|
90 |
-
"eval_samples_per_second":
|
91 |
-
"eval_steps_per_second":
|
92 |
"step": 3000
|
93 |
},
|
94 |
{
|
95 |
"epoch": 0.44,
|
96 |
-
"learning_rate":
|
97 |
-
"loss": 0.
|
98 |
"step": 3500
|
99 |
},
|
100 |
{
|
101 |
"epoch": 0.44,
|
102 |
-
"eval_loss": 0.
|
103 |
-
"eval_runtime":
|
104 |
-
"eval_samples_per_second":
|
105 |
-
"eval_steps_per_second":
|
106 |
"step": 3500
|
107 |
},
|
108 |
{
|
109 |
"epoch": 0.5,
|
110 |
-
"learning_rate":
|
111 |
-
"loss": 0.
|
112 |
"step": 4000
|
113 |
},
|
114 |
{
|
115 |
"epoch": 0.5,
|
116 |
-
"eval_loss": 0.
|
117 |
-
"eval_runtime":
|
118 |
-
"eval_samples_per_second":
|
119 |
-
"eval_steps_per_second":
|
120 |
"step": 4000
|
121 |
},
|
122 |
{
|
123 |
"epoch": 0.57,
|
124 |
-
"learning_rate":
|
125 |
-
"loss": 0.
|
126 |
"step": 4500
|
127 |
},
|
128 |
{
|
129 |
"epoch": 0.57,
|
130 |
-
"eval_loss": 0.
|
131 |
-
"eval_runtime":
|
132 |
-
"eval_samples_per_second":
|
133 |
-
"eval_steps_per_second":
|
134 |
"step": 4500
|
135 |
},
|
136 |
{
|
137 |
"epoch": 0.63,
|
138 |
-
"learning_rate":
|
139 |
-
"loss": 0.
|
140 |
"step": 5000
|
141 |
},
|
142 |
{
|
143 |
"epoch": 0.63,
|
144 |
-
"eval_loss": 0.
|
145 |
-
"eval_runtime":
|
146 |
-
"eval_samples_per_second":
|
147 |
-
"eval_steps_per_second":
|
148 |
"step": 5000
|
149 |
},
|
150 |
{
|
151 |
"epoch": 0.69,
|
152 |
-
"learning_rate":
|
153 |
-
"loss": 0.
|
154 |
"step": 5500
|
155 |
},
|
156 |
{
|
157 |
"epoch": 0.69,
|
158 |
-
"eval_loss": 0.
|
159 |
-
"eval_runtime":
|
160 |
-
"eval_samples_per_second":
|
161 |
-
"eval_steps_per_second":
|
162 |
"step": 5500
|
163 |
},
|
164 |
{
|
165 |
"epoch": 0.76,
|
166 |
-
"learning_rate":
|
167 |
-
"loss": 0.
|
168 |
"step": 6000
|
169 |
},
|
170 |
{
|
171 |
"epoch": 0.76,
|
172 |
-
"eval_loss": 0.
|
173 |
-
"eval_runtime":
|
174 |
-
"eval_samples_per_second":
|
175 |
-
"eval_steps_per_second":
|
176 |
"step": 6000
|
177 |
},
|
178 |
{
|
179 |
"epoch": 0.82,
|
180 |
-
"learning_rate":
|
181 |
-
"loss": 0.
|
182 |
"step": 6500
|
183 |
},
|
184 |
{
|
185 |
"epoch": 0.82,
|
186 |
-
"eval_loss": 0.
|
187 |
-
"eval_runtime":
|
188 |
-
"eval_samples_per_second":
|
189 |
-
"eval_steps_per_second":
|
190 |
"step": 6500
|
191 |
},
|
192 |
{
|
193 |
"epoch": 0.88,
|
194 |
-
"learning_rate":
|
195 |
-
"loss": 0.
|
196 |
"step": 7000
|
197 |
},
|
198 |
{
|
199 |
"epoch": 0.88,
|
200 |
-
"eval_loss": 0.
|
201 |
-
"eval_runtime":
|
202 |
-
"eval_samples_per_second":
|
203 |
-
"eval_steps_per_second":
|
204 |
"step": 7000
|
205 |
},
|
206 |
{
|
207 |
-
"epoch": 0.
|
208 |
-
"learning_rate":
|
209 |
-
"loss": 0.
|
210 |
"step": 7500
|
211 |
},
|
212 |
{
|
213 |
-
"epoch": 0.
|
214 |
-
"eval_loss": 0.
|
215 |
-
"eval_runtime":
|
216 |
-
"eval_samples_per_second":
|
217 |
-
"eval_steps_per_second":
|
218 |
"step": 7500
|
219 |
},
|
220 |
{
|
221 |
"epoch": 1.01,
|
222 |
-
"learning_rate":
|
223 |
-
"loss": 0.
|
224 |
"step": 8000
|
225 |
},
|
226 |
{
|
227 |
"epoch": 1.01,
|
228 |
-
"eval_loss": 0.
|
229 |
-
"eval_runtime":
|
230 |
-
"eval_samples_per_second":
|
231 |
-
"eval_steps_per_second":
|
232 |
"step": 8000
|
233 |
},
|
234 |
{
|
235 |
"epoch": 1.07,
|
236 |
-
"learning_rate":
|
237 |
-
"loss": 0.
|
238 |
"step": 8500
|
239 |
},
|
240 |
{
|
241 |
"epoch": 1.07,
|
242 |
-
"eval_loss": 0.
|
243 |
-
"eval_runtime":
|
244 |
-
"eval_samples_per_second":
|
245 |
-
"eval_steps_per_second":
|
246 |
"step": 8500
|
247 |
},
|
248 |
{
|
249 |
-
"epoch": 1.
|
250 |
-
"learning_rate":
|
251 |
-
"loss": 0.
|
252 |
"step": 9000
|
253 |
},
|
254 |
{
|
255 |
-
"epoch": 1.
|
256 |
-
"eval_loss": 0.
|
257 |
-
"eval_runtime":
|
258 |
-
"eval_samples_per_second":
|
259 |
-
"eval_steps_per_second":
|
260 |
"step": 9000
|
261 |
},
|
262 |
{
|
263 |
"epoch": 1.2,
|
264 |
-
"learning_rate":
|
265 |
-
"loss": 0.
|
266 |
"step": 9500
|
267 |
},
|
268 |
{
|
269 |
"epoch": 1.2,
|
270 |
-
"eval_loss": 0.
|
271 |
-
"eval_runtime":
|
272 |
-
"eval_samples_per_second":
|
273 |
-
"eval_steps_per_second":
|
274 |
"step": 9500
|
275 |
},
|
276 |
{
|
277 |
"epoch": 1.26,
|
278 |
-
"learning_rate": 0.
|
279 |
-
"loss": 0.
|
280 |
"step": 10000
|
281 |
},
|
282 |
{
|
283 |
"epoch": 1.26,
|
284 |
-
"eval_loss": 0.
|
285 |
-
"eval_runtime":
|
286 |
-
"eval_samples_per_second":
|
287 |
-
"eval_steps_per_second":
|
288 |
"step": 10000
|
289 |
}
|
290 |
],
|
291 |
-
"
|
292 |
-
"
|
293 |
-
"
|
|
|
|
|
294 |
"trial_name": null,
|
295 |
"trial_params": null
|
296 |
}
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 1.2597631645250693,
|
5 |
+
"eval_steps": 500,
|
6 |
"global_step": 10000,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
|
|
10 |
"log_history": [
|
11 |
{
|
12 |
"epoch": 0.06,
|
13 |
+
"learning_rate": 0.000975,
|
14 |
+
"loss": 1.0322,
|
15 |
"step": 500
|
16 |
},
|
17 |
{
|
18 |
"epoch": 0.06,
|
19 |
+
"eval_loss": 0.8672707676887512,
|
20 |
+
"eval_runtime": 15.3736,
|
21 |
+
"eval_samples_per_second": 65.047,
|
22 |
+
"eval_steps_per_second": 32.523,
|
23 |
"step": 500
|
24 |
},
|
25 |
{
|
26 |
"epoch": 0.13,
|
27 |
+
"learning_rate": 0.00095,
|
28 |
+
"loss": 0.9884,
|
29 |
"step": 1000
|
30 |
},
|
31 |
{
|
32 |
"epoch": 0.13,
|
33 |
+
"eval_loss": 0.8287830948829651,
|
34 |
+
"eval_runtime": 15.4828,
|
35 |
+
"eval_samples_per_second": 64.588,
|
36 |
+
"eval_steps_per_second": 32.294,
|
37 |
"step": 1000
|
38 |
},
|
39 |
{
|
40 |
"epoch": 0.19,
|
41 |
+
"learning_rate": 0.000925,
|
42 |
+
"loss": 0.9418,
|
43 |
"step": 1500
|
44 |
},
|
45 |
{
|
46 |
"epoch": 0.19,
|
47 |
+
"eval_loss": 0.8468040823936462,
|
48 |
+
"eval_runtime": 15.4288,
|
49 |
+
"eval_samples_per_second": 64.814,
|
50 |
+
"eval_steps_per_second": 32.407,
|
51 |
"step": 1500
|
52 |
},
|
53 |
{
|
54 |
"epoch": 0.25,
|
55 |
+
"learning_rate": 0.0009000000000000001,
|
56 |
+
"loss": 0.9232,
|
57 |
"step": 2000
|
58 |
},
|
59 |
{
|
60 |
"epoch": 0.25,
|
61 |
+
"eval_loss": 0.8540903329849243,
|
62 |
+
"eval_runtime": 15.6504,
|
63 |
+
"eval_samples_per_second": 63.896,
|
64 |
+
"eval_steps_per_second": 31.948,
|
65 |
"step": 2000
|
66 |
},
|
67 |
{
|
68 |
+
"epoch": 0.31,
|
69 |
+
"learning_rate": 0.000875,
|
70 |
+
"loss": 0.9107,
|
71 |
"step": 2500
|
72 |
},
|
73 |
{
|
74 |
+
"epoch": 0.31,
|
75 |
+
"eval_loss": 0.809073269367218,
|
76 |
+
"eval_runtime": 15.3867,
|
77 |
+
"eval_samples_per_second": 64.991,
|
78 |
+
"eval_steps_per_second": 32.496,
|
79 |
"step": 2500
|
80 |
},
|
81 |
{
|
82 |
"epoch": 0.38,
|
83 |
+
"learning_rate": 0.00085,
|
84 |
+
"loss": 0.9233,
|
85 |
"step": 3000
|
86 |
},
|
87 |
{
|
88 |
"epoch": 0.38,
|
89 |
+
"eval_loss": 0.8151862621307373,
|
90 |
+
"eval_runtime": 15.0851,
|
91 |
+
"eval_samples_per_second": 66.291,
|
92 |
+
"eval_steps_per_second": 33.145,
|
93 |
"step": 3000
|
94 |
},
|
95 |
{
|
96 |
"epoch": 0.44,
|
97 |
+
"learning_rate": 0.000825,
|
98 |
+
"loss": 0.9038,
|
99 |
"step": 3500
|
100 |
},
|
101 |
{
|
102 |
"epoch": 0.44,
|
103 |
+
"eval_loss": 0.8232718706130981,
|
104 |
+
"eval_runtime": 15.5073,
|
105 |
+
"eval_samples_per_second": 64.486,
|
106 |
+
"eval_steps_per_second": 32.243,
|
107 |
"step": 3500
|
108 |
},
|
109 |
{
|
110 |
"epoch": 0.5,
|
111 |
+
"learning_rate": 0.0008,
|
112 |
+
"loss": 0.8747,
|
113 |
"step": 4000
|
114 |
},
|
115 |
{
|
116 |
"epoch": 0.5,
|
117 |
+
"eval_loss": 0.7864852547645569,
|
118 |
+
"eval_runtime": 15.3637,
|
119 |
+
"eval_samples_per_second": 65.089,
|
120 |
+
"eval_steps_per_second": 32.544,
|
121 |
"step": 4000
|
122 |
},
|
123 |
{
|
124 |
"epoch": 0.57,
|
125 |
+
"learning_rate": 0.0007750000000000001,
|
126 |
+
"loss": 0.8981,
|
127 |
"step": 4500
|
128 |
},
|
129 |
{
|
130 |
"epoch": 0.57,
|
131 |
+
"eval_loss": 0.770944356918335,
|
132 |
+
"eval_runtime": 15.3602,
|
133 |
+
"eval_samples_per_second": 65.103,
|
134 |
+
"eval_steps_per_second": 32.552,
|
135 |
"step": 4500
|
136 |
},
|
137 |
{
|
138 |
"epoch": 0.63,
|
139 |
+
"learning_rate": 0.00075,
|
140 |
+
"loss": 0.8538,
|
141 |
"step": 5000
|
142 |
},
|
143 |
{
|
144 |
"epoch": 0.63,
|
145 |
+
"eval_loss": 0.7905715107917786,
|
146 |
+
"eval_runtime": 16.0451,
|
147 |
+
"eval_samples_per_second": 62.324,
|
148 |
+
"eval_steps_per_second": 31.162,
|
149 |
"step": 5000
|
150 |
},
|
151 |
{
|
152 |
"epoch": 0.69,
|
153 |
+
"learning_rate": 0.000725,
|
154 |
+
"loss": 0.8295,
|
155 |
"step": 5500
|
156 |
},
|
157 |
{
|
158 |
"epoch": 0.69,
|
159 |
+
"eval_loss": 0.7717331051826477,
|
160 |
+
"eval_runtime": 16.2335,
|
161 |
+
"eval_samples_per_second": 61.601,
|
162 |
+
"eval_steps_per_second": 30.8,
|
163 |
"step": 5500
|
164 |
},
|
165 |
{
|
166 |
"epoch": 0.76,
|
167 |
+
"learning_rate": 0.0007,
|
168 |
+
"loss": 0.8346,
|
169 |
"step": 6000
|
170 |
},
|
171 |
{
|
172 |
"epoch": 0.76,
|
173 |
+
"eval_loss": 0.7587910890579224,
|
174 |
+
"eval_runtime": 15.3749,
|
175 |
+
"eval_samples_per_second": 65.041,
|
176 |
+
"eval_steps_per_second": 32.521,
|
177 |
"step": 6000
|
178 |
},
|
179 |
{
|
180 |
"epoch": 0.82,
|
181 |
+
"learning_rate": 0.000675,
|
182 |
+
"loss": 0.8366,
|
183 |
"step": 6500
|
184 |
},
|
185 |
{
|
186 |
"epoch": 0.82,
|
187 |
+
"eval_loss": 0.7654258012771606,
|
188 |
+
"eval_runtime": 15.1755,
|
189 |
+
"eval_samples_per_second": 65.895,
|
190 |
+
"eval_steps_per_second": 32.948,
|
191 |
"step": 6500
|
192 |
},
|
193 |
{
|
194 |
"epoch": 0.88,
|
195 |
+
"learning_rate": 0.0006500000000000001,
|
196 |
+
"loss": 0.8017,
|
197 |
"step": 7000
|
198 |
},
|
199 |
{
|
200 |
"epoch": 0.88,
|
201 |
+
"eval_loss": 0.759145975112915,
|
202 |
+
"eval_runtime": 15.3597,
|
203 |
+
"eval_samples_per_second": 65.106,
|
204 |
+
"eval_steps_per_second": 32.553,
|
205 |
"step": 7000
|
206 |
},
|
207 |
{
|
208 |
+
"epoch": 0.94,
|
209 |
+
"learning_rate": 0.000625,
|
210 |
+
"loss": 0.7788,
|
211 |
"step": 7500
|
212 |
},
|
213 |
{
|
214 |
+
"epoch": 0.94,
|
215 |
+
"eval_loss": 0.7703807950019836,
|
216 |
+
"eval_runtime": 15.9867,
|
217 |
+
"eval_samples_per_second": 62.552,
|
218 |
+
"eval_steps_per_second": 31.276,
|
219 |
"step": 7500
|
220 |
},
|
221 |
{
|
222 |
"epoch": 1.01,
|
223 |
+
"learning_rate": 0.0006,
|
224 |
+
"loss": 0.7764,
|
225 |
"step": 8000
|
226 |
},
|
227 |
{
|
228 |
"epoch": 1.01,
|
229 |
+
"eval_loss": 0.7488923668861389,
|
230 |
+
"eval_runtime": 16.1496,
|
231 |
+
"eval_samples_per_second": 61.921,
|
232 |
+
"eval_steps_per_second": 30.961,
|
233 |
"step": 8000
|
234 |
},
|
235 |
{
|
236 |
"epoch": 1.07,
|
237 |
+
"learning_rate": 0.000575,
|
238 |
+
"loss": 0.5766,
|
239 |
"step": 8500
|
240 |
},
|
241 |
{
|
242 |
"epoch": 1.07,
|
243 |
+
"eval_loss": 0.7688985466957092,
|
244 |
+
"eval_runtime": 15.5729,
|
245 |
+
"eval_samples_per_second": 64.214,
|
246 |
+
"eval_steps_per_second": 32.107,
|
247 |
"step": 8500
|
248 |
},
|
249 |
{
|
250 |
+
"epoch": 1.13,
|
251 |
+
"learning_rate": 0.00055,
|
252 |
+
"loss": 0.6024,
|
253 |
"step": 9000
|
254 |
},
|
255 |
{
|
256 |
+
"epoch": 1.13,
|
257 |
+
"eval_loss": 0.7472162842750549,
|
258 |
+
"eval_runtime": 15.4638,
|
259 |
+
"eval_samples_per_second": 64.667,
|
260 |
+
"eval_steps_per_second": 32.334,
|
261 |
"step": 9000
|
262 |
},
|
263 |
{
|
264 |
"epoch": 1.2,
|
265 |
+
"learning_rate": 0.0005250000000000001,
|
266 |
+
"loss": 0.5822,
|
267 |
"step": 9500
|
268 |
},
|
269 |
{
|
270 |
"epoch": 1.2,
|
271 |
+
"eval_loss": 0.7330933809280396,
|
272 |
+
"eval_runtime": 15.5588,
|
273 |
+
"eval_samples_per_second": 64.272,
|
274 |
+
"eval_steps_per_second": 32.136,
|
275 |
"step": 9500
|
276 |
},
|
277 |
{
|
278 |
"epoch": 1.26,
|
279 |
+
"learning_rate": 0.0005,
|
280 |
+
"loss": 0.5812,
|
281 |
"step": 10000
|
282 |
},
|
283 |
{
|
284 |
"epoch": 1.26,
|
285 |
+
"eval_loss": 0.7249069809913635,
|
286 |
+
"eval_runtime": 15.1664,
|
287 |
+
"eval_samples_per_second": 65.935,
|
288 |
+
"eval_steps_per_second": 32.968,
|
289 |
"step": 10000
|
290 |
}
|
291 |
],
|
292 |
+
"logging_steps": 500,
|
293 |
+
"max_steps": 20000,
|
294 |
+
"num_train_epochs": 3,
|
295 |
+
"save_steps": 5000,
|
296 |
+
"total_flos": 5.445321533503488e+16,
|
297 |
"trial_name": null,
|
298 |
"trial_params": null
|
299 |
}
|
function-base-flatten/checkpoint-10000/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:857168942562a5574ce0eadda06d15062fa7c087954ae9df9c3df06ec62f42c4
|
3 |
+
size 4283
|
config.json β function-base-flatten/checkpoint-15000/config.json
RENAMED
@@ -3,6 +3,7 @@
|
|
3 |
"architectures": [
|
4 |
"FiDT5"
|
5 |
],
|
|
|
6 |
"d_ff": 2048,
|
7 |
"d_kv": 64,
|
8 |
"d_model": 768,
|
@@ -55,7 +56,7 @@
|
|
55 |
},
|
56 |
"tie_word_embeddings": false,
|
57 |
"torch_dtype": "float32",
|
58 |
-
"transformers_version": "4.
|
59 |
"use_cache": true,
|
60 |
"vocab_size": 32128
|
61 |
}
|
|
|
3 |
"architectures": [
|
4 |
"FiDT5"
|
5 |
],
|
6 |
+
"classifier_dropout": 0.0,
|
7 |
"d_ff": 2048,
|
8 |
"d_kv": 64,
|
9 |
"d_model": 768,
|
|
|
56 |
},
|
57 |
"tie_word_embeddings": false,
|
58 |
"torch_dtype": "float32",
|
59 |
+
"transformers_version": "4.33.1",
|
60 |
"use_cache": true,
|
61 |
"vocab_size": 32128
|
62 |
}
|
{archived/checkpoint-10000 β function-base-flatten/checkpoint-15000}/generation_config.json
RENAMED
@@ -3,5 +3,5 @@
|
|
3 |
"decoder_start_token_id": 0,
|
4 |
"eos_token_id": 1,
|
5 |
"pad_token_id": 0,
|
6 |
-
"transformers_version": "4.
|
7 |
}
|
|
|
3 |
"decoder_start_token_id": 0,
|
4 |
"eos_token_id": 1,
|
5 |
"pad_token_id": 0,
|
6 |
+
"transformers_version": "4.33.1"
|
7 |
}
|
{archived/checkpoint-10000 β function-base-flatten/checkpoint-15000}/optimizer.pt
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e3fc6096a4bea0a15a0ef50591ecf5d74f46d983ae8dcc8255a582e94c0d786e
|
3 |
+
size 2372293
|
{archived/checkpoint-10000 β function-base-flatten/checkpoint-15000}/pytorch_model.bin
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bfc4d7035318a7d9356a8c80980d5e8e98a09127dace85985989b9a5114f139b
|
3 |
+
size 990410745
|
rng_state.pth β function-base-flatten/checkpoint-15000/rng_state.pth
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14575
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:84148658af0ffc7e402e61e276c1e47ca4c371c6495ebac7d6c4bc7bb1680e2c
|
3 |
size 14575
|
{archived/checkpoint-10000 β function-base-flatten/checkpoint-15000}/scheduler.pt
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 627
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:62b793f9ea6eba39185d82063b1e7434411e2aeca1bab5a010024f955d1696b3
|
3 |
size 627
|
function-base-flatten/checkpoint-15000/trainer_state.json
ADDED
@@ -0,0 +1,439 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 1.8896447467876039,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 15000,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.06,
|
13 |
+
"learning_rate": 0.000975,
|
14 |
+
"loss": 1.0322,
|
15 |
+
"step": 500
|
16 |
+
},
|
17 |
+
{
|
18 |
+
"epoch": 0.06,
|
19 |
+
"eval_loss": 0.8672707676887512,
|
20 |
+
"eval_runtime": 15.3736,
|
21 |
+
"eval_samples_per_second": 65.047,
|
22 |
+
"eval_steps_per_second": 32.523,
|
23 |
+
"step": 500
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.13,
|
27 |
+
"learning_rate": 0.00095,
|
28 |
+
"loss": 0.9884,
|
29 |
+
"step": 1000
|
30 |
+
},
|
31 |
+
{
|
32 |
+
"epoch": 0.13,
|
33 |
+
"eval_loss": 0.8287830948829651,
|
34 |
+
"eval_runtime": 15.4828,
|
35 |
+
"eval_samples_per_second": 64.588,
|
36 |
+
"eval_steps_per_second": 32.294,
|
37 |
+
"step": 1000
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.19,
|
41 |
+
"learning_rate": 0.000925,
|
42 |
+
"loss": 0.9418,
|
43 |
+
"step": 1500
|
44 |
+
},
|
45 |
+
{
|
46 |
+
"epoch": 0.19,
|
47 |
+
"eval_loss": 0.8468040823936462,
|
48 |
+
"eval_runtime": 15.4288,
|
49 |
+
"eval_samples_per_second": 64.814,
|
50 |
+
"eval_steps_per_second": 32.407,
|
51 |
+
"step": 1500
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.25,
|
55 |
+
"learning_rate": 0.0009000000000000001,
|
56 |
+
"loss": 0.9232,
|
57 |
+
"step": 2000
|
58 |
+
},
|
59 |
+
{
|
60 |
+
"epoch": 0.25,
|
61 |
+
"eval_loss": 0.8540903329849243,
|
62 |
+
"eval_runtime": 15.6504,
|
63 |
+
"eval_samples_per_second": 63.896,
|
64 |
+
"eval_steps_per_second": 31.948,
|
65 |
+
"step": 2000
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.31,
|
69 |
+
"learning_rate": 0.000875,
|
70 |
+
"loss": 0.9107,
|
71 |
+
"step": 2500
|
72 |
+
},
|
73 |
+
{
|
74 |
+
"epoch": 0.31,
|
75 |
+
"eval_loss": 0.809073269367218,
|
76 |
+
"eval_runtime": 15.3867,
|
77 |
+
"eval_samples_per_second": 64.991,
|
78 |
+
"eval_steps_per_second": 32.496,
|
79 |
+
"step": 2500
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 0.38,
|
83 |
+
"learning_rate": 0.00085,
|
84 |
+
"loss": 0.9233,
|
85 |
+
"step": 3000
|
86 |
+
},
|
87 |
+
{
|
88 |
+
"epoch": 0.38,
|
89 |
+
"eval_loss": 0.8151862621307373,
|
90 |
+
"eval_runtime": 15.0851,
|
91 |
+
"eval_samples_per_second": 66.291,
|
92 |
+
"eval_steps_per_second": 33.145,
|
93 |
+
"step": 3000
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 0.44,
|
97 |
+
"learning_rate": 0.000825,
|
98 |
+
"loss": 0.9038,
|
99 |
+
"step": 3500
|
100 |
+
},
|
101 |
+
{
|
102 |
+
"epoch": 0.44,
|
103 |
+
"eval_loss": 0.8232718706130981,
|
104 |
+
"eval_runtime": 15.5073,
|
105 |
+
"eval_samples_per_second": 64.486,
|
106 |
+
"eval_steps_per_second": 32.243,
|
107 |
+
"step": 3500
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 0.5,
|
111 |
+
"learning_rate": 0.0008,
|
112 |
+
"loss": 0.8747,
|
113 |
+
"step": 4000
|
114 |
+
},
|
115 |
+
{
|
116 |
+
"epoch": 0.5,
|
117 |
+
"eval_loss": 0.7864852547645569,
|
118 |
+
"eval_runtime": 15.3637,
|
119 |
+
"eval_samples_per_second": 65.089,
|
120 |
+
"eval_steps_per_second": 32.544,
|
121 |
+
"step": 4000
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"epoch": 0.57,
|
125 |
+
"learning_rate": 0.0007750000000000001,
|
126 |
+
"loss": 0.8981,
|
127 |
+
"step": 4500
|
128 |
+
},
|
129 |
+
{
|
130 |
+
"epoch": 0.57,
|
131 |
+
"eval_loss": 0.770944356918335,
|
132 |
+
"eval_runtime": 15.3602,
|
133 |
+
"eval_samples_per_second": 65.103,
|
134 |
+
"eval_steps_per_second": 32.552,
|
135 |
+
"step": 4500
|
136 |
+
},
|
137 |
+
{
|
138 |
+
"epoch": 0.63,
|
139 |
+
"learning_rate": 0.00075,
|
140 |
+
"loss": 0.8538,
|
141 |
+
"step": 5000
|
142 |
+
},
|
143 |
+
{
|
144 |
+
"epoch": 0.63,
|
145 |
+
"eval_loss": 0.7905715107917786,
|
146 |
+
"eval_runtime": 16.0451,
|
147 |
+
"eval_samples_per_second": 62.324,
|
148 |
+
"eval_steps_per_second": 31.162,
|
149 |
+
"step": 5000
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 0.69,
|
153 |
+
"learning_rate": 0.000725,
|
154 |
+
"loss": 0.8295,
|
155 |
+
"step": 5500
|
156 |
+
},
|
157 |
+
{
|
158 |
+
"epoch": 0.69,
|
159 |
+
"eval_loss": 0.7717331051826477,
|
160 |
+
"eval_runtime": 16.2335,
|
161 |
+
"eval_samples_per_second": 61.601,
|
162 |
+
"eval_steps_per_second": 30.8,
|
163 |
+
"step": 5500
|
164 |
+
},
|
165 |
+
{
|
166 |
+
"epoch": 0.76,
|
167 |
+
"learning_rate": 0.0007,
|
168 |
+
"loss": 0.8346,
|
169 |
+
"step": 6000
|
170 |
+
},
|
171 |
+
{
|
172 |
+
"epoch": 0.76,
|
173 |
+
"eval_loss": 0.7587910890579224,
|
174 |
+
"eval_runtime": 15.3749,
|
175 |
+
"eval_samples_per_second": 65.041,
|
176 |
+
"eval_steps_per_second": 32.521,
|
177 |
+
"step": 6000
|
178 |
+
},
|
179 |
+
{
|
180 |
+
"epoch": 0.82,
|
181 |
+
"learning_rate": 0.000675,
|
182 |
+
"loss": 0.8366,
|
183 |
+
"step": 6500
|
184 |
+
},
|
185 |
+
{
|
186 |
+
"epoch": 0.82,
|
187 |
+
"eval_loss": 0.7654258012771606,
|
188 |
+
"eval_runtime": 15.1755,
|
189 |
+
"eval_samples_per_second": 65.895,
|
190 |
+
"eval_steps_per_second": 32.948,
|
191 |
+
"step": 6500
|
192 |
+
},
|
193 |
+
{
|
194 |
+
"epoch": 0.88,
|
195 |
+
"learning_rate": 0.0006500000000000001,
|
196 |
+
"loss": 0.8017,
|
197 |
+
"step": 7000
|
198 |
+
},
|
199 |
+
{
|
200 |
+
"epoch": 0.88,
|
201 |
+
"eval_loss": 0.759145975112915,
|
202 |
+
"eval_runtime": 15.3597,
|
203 |
+
"eval_samples_per_second": 65.106,
|
204 |
+
"eval_steps_per_second": 32.553,
|
205 |
+
"step": 7000
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 0.94,
|
209 |
+
"learning_rate": 0.000625,
|
210 |
+
"loss": 0.7788,
|
211 |
+
"step": 7500
|
212 |
+
},
|
213 |
+
{
|
214 |
+
"epoch": 0.94,
|
215 |
+
"eval_loss": 0.7703807950019836,
|
216 |
+
"eval_runtime": 15.9867,
|
217 |
+
"eval_samples_per_second": 62.552,
|
218 |
+
"eval_steps_per_second": 31.276,
|
219 |
+
"step": 7500
|
220 |
+
},
|
221 |
+
{
|
222 |
+
"epoch": 1.01,
|
223 |
+
"learning_rate": 0.0006,
|
224 |
+
"loss": 0.7764,
|
225 |
+
"step": 8000
|
226 |
+
},
|
227 |
+
{
|
228 |
+
"epoch": 1.01,
|
229 |
+
"eval_loss": 0.7488923668861389,
|
230 |
+
"eval_runtime": 16.1496,
|
231 |
+
"eval_samples_per_second": 61.921,
|
232 |
+
"eval_steps_per_second": 30.961,
|
233 |
+
"step": 8000
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 1.07,
|
237 |
+
"learning_rate": 0.000575,
|
238 |
+
"loss": 0.5766,
|
239 |
+
"step": 8500
|
240 |
+
},
|
241 |
+
{
|
242 |
+
"epoch": 1.07,
|
243 |
+
"eval_loss": 0.7688985466957092,
|
244 |
+
"eval_runtime": 15.5729,
|
245 |
+
"eval_samples_per_second": 64.214,
|
246 |
+
"eval_steps_per_second": 32.107,
|
247 |
+
"step": 8500
|
248 |
+
},
|
249 |
+
{
|
250 |
+
"epoch": 1.13,
|
251 |
+
"learning_rate": 0.00055,
|
252 |
+
"loss": 0.6024,
|
253 |
+
"step": 9000
|
254 |
+
},
|
255 |
+
{
|
256 |
+
"epoch": 1.13,
|
257 |
+
"eval_loss": 0.7472162842750549,
|
258 |
+
"eval_runtime": 15.4638,
|
259 |
+
"eval_samples_per_second": 64.667,
|
260 |
+
"eval_steps_per_second": 32.334,
|
261 |
+
"step": 9000
|
262 |
+
},
|
263 |
+
{
|
264 |
+
"epoch": 1.2,
|
265 |
+
"learning_rate": 0.0005250000000000001,
|
266 |
+
"loss": 0.5822,
|
267 |
+
"step": 9500
|
268 |
+
},
|
269 |
+
{
|
270 |
+
"epoch": 1.2,
|
271 |
+
"eval_loss": 0.7330933809280396,
|
272 |
+
"eval_runtime": 15.5588,
|
273 |
+
"eval_samples_per_second": 64.272,
|
274 |
+
"eval_steps_per_second": 32.136,
|
275 |
+
"step": 9500
|
276 |
+
},
|
277 |
+
{
|
278 |
+
"epoch": 1.26,
|
279 |
+
"learning_rate": 0.0005,
|
280 |
+
"loss": 0.5812,
|
281 |
+
"step": 10000
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 1.26,
|
285 |
+
"eval_loss": 0.7249069809913635,
|
286 |
+
"eval_runtime": 15.1664,
|
287 |
+
"eval_samples_per_second": 65.935,
|
288 |
+
"eval_steps_per_second": 32.968,
|
289 |
+
"step": 10000
|
290 |
+
},
|
291 |
+
{
|
292 |
+
"epoch": 1.32,
|
293 |
+
"learning_rate": 0.000475,
|
294 |
+
"loss": 0.5758,
|
295 |
+
"step": 10500
|
296 |
+
},
|
297 |
+
{
|
298 |
+
"epoch": 1.32,
|
299 |
+
"eval_loss": 0.7344868779182434,
|
300 |
+
"eval_runtime": 15.267,
|
301 |
+
"eval_samples_per_second": 65.501,
|
302 |
+
"eval_steps_per_second": 32.75,
|
303 |
+
"step": 10500
|
304 |
+
},
|
305 |
+
{
|
306 |
+
"epoch": 1.39,
|
307 |
+
"learning_rate": 0.00045000000000000004,
|
308 |
+
"loss": 0.5714,
|
309 |
+
"step": 11000
|
310 |
+
},
|
311 |
+
{
|
312 |
+
"epoch": 1.39,
|
313 |
+
"eval_loss": 0.7404966950416565,
|
314 |
+
"eval_runtime": 15.2995,
|
315 |
+
"eval_samples_per_second": 65.361,
|
316 |
+
"eval_steps_per_second": 32.681,
|
317 |
+
"step": 11000
|
318 |
+
},
|
319 |
+
{
|
320 |
+
"epoch": 1.45,
|
321 |
+
"learning_rate": 0.000425,
|
322 |
+
"loss": 0.5787,
|
323 |
+
"step": 11500
|
324 |
+
},
|
325 |
+
{
|
326 |
+
"epoch": 1.45,
|
327 |
+
"eval_loss": 0.7168460488319397,
|
328 |
+
"eval_runtime": 15.277,
|
329 |
+
"eval_samples_per_second": 65.458,
|
330 |
+
"eval_steps_per_second": 32.729,
|
331 |
+
"step": 11500
|
332 |
+
},
|
333 |
+
{
|
334 |
+
"epoch": 1.51,
|
335 |
+
"learning_rate": 0.0004,
|
336 |
+
"loss": 0.5694,
|
337 |
+
"step": 12000
|
338 |
+
},
|
339 |
+
{
|
340 |
+
"epoch": 1.51,
|
341 |
+
"eval_loss": 0.7175167202949524,
|
342 |
+
"eval_runtime": 15.6988,
|
343 |
+
"eval_samples_per_second": 63.699,
|
344 |
+
"eval_steps_per_second": 31.85,
|
345 |
+
"step": 12000
|
346 |
+
},
|
347 |
+
{
|
348 |
+
"epoch": 1.57,
|
349 |
+
"learning_rate": 0.000375,
|
350 |
+
"loss": 0.5835,
|
351 |
+
"step": 12500
|
352 |
+
},
|
353 |
+
{
|
354 |
+
"epoch": 1.57,
|
355 |
+
"eval_loss": 0.7070069313049316,
|
356 |
+
"eval_runtime": 17.1549,
|
357 |
+
"eval_samples_per_second": 58.293,
|
358 |
+
"eval_steps_per_second": 29.146,
|
359 |
+
"step": 12500
|
360 |
+
},
|
361 |
+
{
|
362 |
+
"epoch": 1.64,
|
363 |
+
"learning_rate": 0.00035,
|
364 |
+
"loss": 0.5664,
|
365 |
+
"step": 13000
|
366 |
+
},
|
367 |
+
{
|
368 |
+
"epoch": 1.64,
|
369 |
+
"eval_loss": 0.7037546038627625,
|
370 |
+
"eval_runtime": 15.2843,
|
371 |
+
"eval_samples_per_second": 65.427,
|
372 |
+
"eval_steps_per_second": 32.713,
|
373 |
+
"step": 13000
|
374 |
+
},
|
375 |
+
{
|
376 |
+
"epoch": 1.7,
|
377 |
+
"learning_rate": 0.00032500000000000004,
|
378 |
+
"loss": 0.5534,
|
379 |
+
"step": 13500
|
380 |
+
},
|
381 |
+
{
|
382 |
+
"epoch": 1.7,
|
383 |
+
"eval_loss": 0.706645667552948,
|
384 |
+
"eval_runtime": 15.983,
|
385 |
+
"eval_samples_per_second": 62.567,
|
386 |
+
"eval_steps_per_second": 31.283,
|
387 |
+
"step": 13500
|
388 |
+
},
|
389 |
+
{
|
390 |
+
"epoch": 1.76,
|
391 |
+
"learning_rate": 0.0003,
|
392 |
+
"loss": 0.5588,
|
393 |
+
"step": 14000
|
394 |
+
},
|
395 |
+
{
|
396 |
+
"epoch": 1.76,
|
397 |
+
"eval_loss": 0.7000927329063416,
|
398 |
+
"eval_runtime": 15.4052,
|
399 |
+
"eval_samples_per_second": 64.913,
|
400 |
+
"eval_steps_per_second": 32.457,
|
401 |
+
"step": 14000
|
402 |
+
},
|
403 |
+
{
|
404 |
+
"epoch": 1.83,
|
405 |
+
"learning_rate": 0.000275,
|
406 |
+
"loss": 0.5568,
|
407 |
+
"step": 14500
|
408 |
+
},
|
409 |
+
{
|
410 |
+
"epoch": 1.83,
|
411 |
+
"eval_loss": 0.6883980631828308,
|
412 |
+
"eval_runtime": 15.226,
|
413 |
+
"eval_samples_per_second": 65.677,
|
414 |
+
"eval_steps_per_second": 32.839,
|
415 |
+
"step": 14500
|
416 |
+
},
|
417 |
+
{
|
418 |
+
"epoch": 1.89,
|
419 |
+
"learning_rate": 0.00025,
|
420 |
+
"loss": 0.5661,
|
421 |
+
"step": 15000
|
422 |
+
},
|
423 |
+
{
|
424 |
+
"epoch": 1.89,
|
425 |
+
"eval_loss": 0.6906747817993164,
|
426 |
+
"eval_runtime": 15.225,
|
427 |
+
"eval_samples_per_second": 65.681,
|
428 |
+
"eval_steps_per_second": 32.841,
|
429 |
+
"step": 15000
|
430 |
+
}
|
431 |
+
],
|
432 |
+
"logging_steps": 500,
|
433 |
+
"max_steps": 20000,
|
434 |
+
"num_train_epochs": 3,
|
435 |
+
"save_steps": 5000,
|
436 |
+
"total_flos": 8.170201074567168e+16,
|
437 |
+
"trial_name": null,
|
438 |
+
"trial_params": null
|
439 |
+
}
|
function-base-flatten/checkpoint-15000/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:857168942562a5574ce0eadda06d15062fa7c087954ae9df9c3df06ec62f42c4
|
3 |
+
size 4283
|
function-base-flatten/checkpoint-20000/config.json
ADDED
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "google/flan-t5-base",
|
3 |
+
"architectures": [
|
4 |
+
"FiDT5"
|
5 |
+
],
|
6 |
+
"classifier_dropout": 0.0,
|
7 |
+
"d_ff": 2048,
|
8 |
+
"d_kv": 64,
|
9 |
+
"d_model": 768,
|
10 |
+
"decoder_start_token_id": 0,
|
11 |
+
"dense_act_fn": "gelu_new",
|
12 |
+
"dropout_rate": 0.1,
|
13 |
+
"eos_token_id": 1,
|
14 |
+
"feed_forward_proj": "gated-gelu",
|
15 |
+
"initializer_factor": 1.0,
|
16 |
+
"is_encoder_decoder": true,
|
17 |
+
"is_gated_act": true,
|
18 |
+
"layer_norm_epsilon": 1e-06,
|
19 |
+
"model_type": "t5",
|
20 |
+
"n_positions": 512,
|
21 |
+
"num_decoder_layers": 12,
|
22 |
+
"num_heads": 12,
|
23 |
+
"num_layers": 12,
|
24 |
+
"output_past": true,
|
25 |
+
"pad_token_id": 0,
|
26 |
+
"relative_attention_max_distance": 128,
|
27 |
+
"relative_attention_num_buckets": 32,
|
28 |
+
"task_specific_params": {
|
29 |
+
"summarization": {
|
30 |
+
"early_stopping": true,
|
31 |
+
"length_penalty": 2.0,
|
32 |
+
"max_length": 200,
|
33 |
+
"min_length": 30,
|
34 |
+
"no_repeat_ngram_size": 3,
|
35 |
+
"num_beams": 4,
|
36 |
+
"prefix": "summarize: "
|
37 |
+
},
|
38 |
+
"translation_en_to_de": {
|
39 |
+
"early_stopping": true,
|
40 |
+
"max_length": 300,
|
41 |
+
"num_beams": 4,
|
42 |
+
"prefix": "translate English to German: "
|
43 |
+
},
|
44 |
+
"translation_en_to_fr": {
|
45 |
+
"early_stopping": true,
|
46 |
+
"max_length": 300,
|
47 |
+
"num_beams": 4,
|
48 |
+
"prefix": "translate English to French: "
|
49 |
+
},
|
50 |
+
"translation_en_to_ro": {
|
51 |
+
"early_stopping": true,
|
52 |
+
"max_length": 300,
|
53 |
+
"num_beams": 4,
|
54 |
+
"prefix": "translate English to Romanian: "
|
55 |
+
}
|
56 |
+
},
|
57 |
+
"tie_word_embeddings": false,
|
58 |
+
"torch_dtype": "float32",
|
59 |
+
"transformers_version": "4.33.1",
|
60 |
+
"use_cache": true,
|
61 |
+
"vocab_size": 32128
|
62 |
+
}
|
function-base-flatten/checkpoint-20000/generation_config.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"decoder_start_token_id": 0,
|
4 |
+
"eos_token_id": 1,
|
5 |
+
"pad_token_id": 0,
|
6 |
+
"transformers_version": "4.33.1"
|
7 |
+
}
|
function-base-flatten/checkpoint-20000/optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f48f34683d2ac110d59dd00b1585aaef74de33fcadd1e19a361eb0b516be913c
|
3 |
+
size 2372293
|
function-base-flatten/checkpoint-20000/pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:aca407d24447508342f9dfc23627a02773a09e95bd307d9602b44bd70f333450
|
3 |
+
size 990410745
|
function-base-flatten/checkpoint-20000/rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d45fe887815d6bbd1e0ca49c6430d2920293486016d8671f425007a097024ef6
|
3 |
+
size 14575
|
function-base-flatten/checkpoint-20000/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f58a0653df4cbbcf9d6cc03d846193b654e5e1cc8a7d6462c99377d7fbe445ea
|
3 |
+
size 627
|
function-base-flatten/checkpoint-20000/trainer_state.json
ADDED
@@ -0,0 +1,579 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 2.5195263290501386,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 20000,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.06,
|
13 |
+
"learning_rate": 0.000975,
|
14 |
+
"loss": 1.0322,
|
15 |
+
"step": 500
|
16 |
+
},
|
17 |
+
{
|
18 |
+
"epoch": 0.06,
|
19 |
+
"eval_loss": 0.8672707676887512,
|
20 |
+
"eval_runtime": 15.3736,
|
21 |
+
"eval_samples_per_second": 65.047,
|
22 |
+
"eval_steps_per_second": 32.523,
|
23 |
+
"step": 500
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.13,
|
27 |
+
"learning_rate": 0.00095,
|
28 |
+
"loss": 0.9884,
|
29 |
+
"step": 1000
|
30 |
+
},
|
31 |
+
{
|
32 |
+
"epoch": 0.13,
|
33 |
+
"eval_loss": 0.8287830948829651,
|
34 |
+
"eval_runtime": 15.4828,
|
35 |
+
"eval_samples_per_second": 64.588,
|
36 |
+
"eval_steps_per_second": 32.294,
|
37 |
+
"step": 1000
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.19,
|
41 |
+
"learning_rate": 0.000925,
|
42 |
+
"loss": 0.9418,
|
43 |
+
"step": 1500
|
44 |
+
},
|
45 |
+
{
|
46 |
+
"epoch": 0.19,
|
47 |
+
"eval_loss": 0.8468040823936462,
|
48 |
+
"eval_runtime": 15.4288,
|
49 |
+
"eval_samples_per_second": 64.814,
|
50 |
+
"eval_steps_per_second": 32.407,
|
51 |
+
"step": 1500
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.25,
|
55 |
+
"learning_rate": 0.0009000000000000001,
|
56 |
+
"loss": 0.9232,
|
57 |
+
"step": 2000
|
58 |
+
},
|
59 |
+
{
|
60 |
+
"epoch": 0.25,
|
61 |
+
"eval_loss": 0.8540903329849243,
|
62 |
+
"eval_runtime": 15.6504,
|
63 |
+
"eval_samples_per_second": 63.896,
|
64 |
+
"eval_steps_per_second": 31.948,
|
65 |
+
"step": 2000
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.31,
|
69 |
+
"learning_rate": 0.000875,
|
70 |
+
"loss": 0.9107,
|
71 |
+
"step": 2500
|
72 |
+
},
|
73 |
+
{
|
74 |
+
"epoch": 0.31,
|
75 |
+
"eval_loss": 0.809073269367218,
|
76 |
+
"eval_runtime": 15.3867,
|
77 |
+
"eval_samples_per_second": 64.991,
|
78 |
+
"eval_steps_per_second": 32.496,
|
79 |
+
"step": 2500
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 0.38,
|
83 |
+
"learning_rate": 0.00085,
|
84 |
+
"loss": 0.9233,
|
85 |
+
"step": 3000
|
86 |
+
},
|
87 |
+
{
|
88 |
+
"epoch": 0.38,
|
89 |
+
"eval_loss": 0.8151862621307373,
|
90 |
+
"eval_runtime": 15.0851,
|
91 |
+
"eval_samples_per_second": 66.291,
|
92 |
+
"eval_steps_per_second": 33.145,
|
93 |
+
"step": 3000
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 0.44,
|
97 |
+
"learning_rate": 0.000825,
|
98 |
+
"loss": 0.9038,
|
99 |
+
"step": 3500
|
100 |
+
},
|
101 |
+
{
|
102 |
+
"epoch": 0.44,
|
103 |
+
"eval_loss": 0.8232718706130981,
|
104 |
+
"eval_runtime": 15.5073,
|
105 |
+
"eval_samples_per_second": 64.486,
|
106 |
+
"eval_steps_per_second": 32.243,
|
107 |
+
"step": 3500
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 0.5,
|
111 |
+
"learning_rate": 0.0008,
|
112 |
+
"loss": 0.8747,
|
113 |
+
"step": 4000
|
114 |
+
},
|
115 |
+
{
|
116 |
+
"epoch": 0.5,
|
117 |
+
"eval_loss": 0.7864852547645569,
|
118 |
+
"eval_runtime": 15.3637,
|
119 |
+
"eval_samples_per_second": 65.089,
|
120 |
+
"eval_steps_per_second": 32.544,
|
121 |
+
"step": 4000
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"epoch": 0.57,
|
125 |
+
"learning_rate": 0.0007750000000000001,
|
126 |
+
"loss": 0.8981,
|
127 |
+
"step": 4500
|
128 |
+
},
|
129 |
+
{
|
130 |
+
"epoch": 0.57,
|
131 |
+
"eval_loss": 0.770944356918335,
|
132 |
+
"eval_runtime": 15.3602,
|
133 |
+
"eval_samples_per_second": 65.103,
|
134 |
+
"eval_steps_per_second": 32.552,
|
135 |
+
"step": 4500
|
136 |
+
},
|
137 |
+
{
|
138 |
+
"epoch": 0.63,
|
139 |
+
"learning_rate": 0.00075,
|
140 |
+
"loss": 0.8538,
|
141 |
+
"step": 5000
|
142 |
+
},
|
143 |
+
{
|
144 |
+
"epoch": 0.63,
|
145 |
+
"eval_loss": 0.7905715107917786,
|
146 |
+
"eval_runtime": 16.0451,
|
147 |
+
"eval_samples_per_second": 62.324,
|
148 |
+
"eval_steps_per_second": 31.162,
|
149 |
+
"step": 5000
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 0.69,
|
153 |
+
"learning_rate": 0.000725,
|
154 |
+
"loss": 0.8295,
|
155 |
+
"step": 5500
|
156 |
+
},
|
157 |
+
{
|
158 |
+
"epoch": 0.69,
|
159 |
+
"eval_loss": 0.7717331051826477,
|
160 |
+
"eval_runtime": 16.2335,
|
161 |
+
"eval_samples_per_second": 61.601,
|
162 |
+
"eval_steps_per_second": 30.8,
|
163 |
+
"step": 5500
|
164 |
+
},
|
165 |
+
{
|
166 |
+
"epoch": 0.76,
|
167 |
+
"learning_rate": 0.0007,
|
168 |
+
"loss": 0.8346,
|
169 |
+
"step": 6000
|
170 |
+
},
|
171 |
+
{
|
172 |
+
"epoch": 0.76,
|
173 |
+
"eval_loss": 0.7587910890579224,
|
174 |
+
"eval_runtime": 15.3749,
|
175 |
+
"eval_samples_per_second": 65.041,
|
176 |
+
"eval_steps_per_second": 32.521,
|
177 |
+
"step": 6000
|
178 |
+
},
|
179 |
+
{
|
180 |
+
"epoch": 0.82,
|
181 |
+
"learning_rate": 0.000675,
|
182 |
+
"loss": 0.8366,
|
183 |
+
"step": 6500
|
184 |
+
},
|
185 |
+
{
|
186 |
+
"epoch": 0.82,
|
187 |
+
"eval_loss": 0.7654258012771606,
|
188 |
+
"eval_runtime": 15.1755,
|
189 |
+
"eval_samples_per_second": 65.895,
|
190 |
+
"eval_steps_per_second": 32.948,
|
191 |
+
"step": 6500
|
192 |
+
},
|
193 |
+
{
|
194 |
+
"epoch": 0.88,
|
195 |
+
"learning_rate": 0.0006500000000000001,
|
196 |
+
"loss": 0.8017,
|
197 |
+
"step": 7000
|
198 |
+
},
|
199 |
+
{
|
200 |
+
"epoch": 0.88,
|
201 |
+
"eval_loss": 0.759145975112915,
|
202 |
+
"eval_runtime": 15.3597,
|
203 |
+
"eval_samples_per_second": 65.106,
|
204 |
+
"eval_steps_per_second": 32.553,
|
205 |
+
"step": 7000
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 0.94,
|
209 |
+
"learning_rate": 0.000625,
|
210 |
+
"loss": 0.7788,
|
211 |
+
"step": 7500
|
212 |
+
},
|
213 |
+
{
|
214 |
+
"epoch": 0.94,
|
215 |
+
"eval_loss": 0.7703807950019836,
|
216 |
+
"eval_runtime": 15.9867,
|
217 |
+
"eval_samples_per_second": 62.552,
|
218 |
+
"eval_steps_per_second": 31.276,
|
219 |
+
"step": 7500
|
220 |
+
},
|
221 |
+
{
|
222 |
+
"epoch": 1.01,
|
223 |
+
"learning_rate": 0.0006,
|
224 |
+
"loss": 0.7764,
|
225 |
+
"step": 8000
|
226 |
+
},
|
227 |
+
{
|
228 |
+
"epoch": 1.01,
|
229 |
+
"eval_loss": 0.7488923668861389,
|
230 |
+
"eval_runtime": 16.1496,
|
231 |
+
"eval_samples_per_second": 61.921,
|
232 |
+
"eval_steps_per_second": 30.961,
|
233 |
+
"step": 8000
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 1.07,
|
237 |
+
"learning_rate": 0.000575,
|
238 |
+
"loss": 0.5766,
|
239 |
+
"step": 8500
|
240 |
+
},
|
241 |
+
{
|
242 |
+
"epoch": 1.07,
|
243 |
+
"eval_loss": 0.7688985466957092,
|
244 |
+
"eval_runtime": 15.5729,
|
245 |
+
"eval_samples_per_second": 64.214,
|
246 |
+
"eval_steps_per_second": 32.107,
|
247 |
+
"step": 8500
|
248 |
+
},
|
249 |
+
{
|
250 |
+
"epoch": 1.13,
|
251 |
+
"learning_rate": 0.00055,
|
252 |
+
"loss": 0.6024,
|
253 |
+
"step": 9000
|
254 |
+
},
|
255 |
+
{
|
256 |
+
"epoch": 1.13,
|
257 |
+
"eval_loss": 0.7472162842750549,
|
258 |
+
"eval_runtime": 15.4638,
|
259 |
+
"eval_samples_per_second": 64.667,
|
260 |
+
"eval_steps_per_second": 32.334,
|
261 |
+
"step": 9000
|
262 |
+
},
|
263 |
+
{
|
264 |
+
"epoch": 1.2,
|
265 |
+
"learning_rate": 0.0005250000000000001,
|
266 |
+
"loss": 0.5822,
|
267 |
+
"step": 9500
|
268 |
+
},
|
269 |
+
{
|
270 |
+
"epoch": 1.2,
|
271 |
+
"eval_loss": 0.7330933809280396,
|
272 |
+
"eval_runtime": 15.5588,
|
273 |
+
"eval_samples_per_second": 64.272,
|
274 |
+
"eval_steps_per_second": 32.136,
|
275 |
+
"step": 9500
|
276 |
+
},
|
277 |
+
{
|
278 |
+
"epoch": 1.26,
|
279 |
+
"learning_rate": 0.0005,
|
280 |
+
"loss": 0.5812,
|
281 |
+
"step": 10000
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 1.26,
|
285 |
+
"eval_loss": 0.7249069809913635,
|
286 |
+
"eval_runtime": 15.1664,
|
287 |
+
"eval_samples_per_second": 65.935,
|
288 |
+
"eval_steps_per_second": 32.968,
|
289 |
+
"step": 10000
|
290 |
+
},
|
291 |
+
{
|
292 |
+
"epoch": 1.32,
|
293 |
+
"learning_rate": 0.000475,
|
294 |
+
"loss": 0.5758,
|
295 |
+
"step": 10500
|
296 |
+
},
|
297 |
+
{
|
298 |
+
"epoch": 1.32,
|
299 |
+
"eval_loss": 0.7344868779182434,
|
300 |
+
"eval_runtime": 15.267,
|
301 |
+
"eval_samples_per_second": 65.501,
|
302 |
+
"eval_steps_per_second": 32.75,
|
303 |
+
"step": 10500
|
304 |
+
},
|
305 |
+
{
|
306 |
+
"epoch": 1.39,
|
307 |
+
"learning_rate": 0.00045000000000000004,
|
308 |
+
"loss": 0.5714,
|
309 |
+
"step": 11000
|
310 |
+
},
|
311 |
+
{
|
312 |
+
"epoch": 1.39,
|
313 |
+
"eval_loss": 0.7404966950416565,
|
314 |
+
"eval_runtime": 15.2995,
|
315 |
+
"eval_samples_per_second": 65.361,
|
316 |
+
"eval_steps_per_second": 32.681,
|
317 |
+
"step": 11000
|
318 |
+
},
|
319 |
+
{
|
320 |
+
"epoch": 1.45,
|
321 |
+
"learning_rate": 0.000425,
|
322 |
+
"loss": 0.5787,
|
323 |
+
"step": 11500
|
324 |
+
},
|
325 |
+
{
|
326 |
+
"epoch": 1.45,
|
327 |
+
"eval_loss": 0.7168460488319397,
|
328 |
+
"eval_runtime": 15.277,
|
329 |
+
"eval_samples_per_second": 65.458,
|
330 |
+
"eval_steps_per_second": 32.729,
|
331 |
+
"step": 11500
|
332 |
+
},
|
333 |
+
{
|
334 |
+
"epoch": 1.51,
|
335 |
+
"learning_rate": 0.0004,
|
336 |
+
"loss": 0.5694,
|
337 |
+
"step": 12000
|
338 |
+
},
|
339 |
+
{
|
340 |
+
"epoch": 1.51,
|
341 |
+
"eval_loss": 0.7175167202949524,
|
342 |
+
"eval_runtime": 15.6988,
|
343 |
+
"eval_samples_per_second": 63.699,
|
344 |
+
"eval_steps_per_second": 31.85,
|
345 |
+
"step": 12000
|
346 |
+
},
|
347 |
+
{
|
348 |
+
"epoch": 1.57,
|
349 |
+
"learning_rate": 0.000375,
|
350 |
+
"loss": 0.5835,
|
351 |
+
"step": 12500
|
352 |
+
},
|
353 |
+
{
|
354 |
+
"epoch": 1.57,
|
355 |
+
"eval_loss": 0.7070069313049316,
|
356 |
+
"eval_runtime": 17.1549,
|
357 |
+
"eval_samples_per_second": 58.293,
|
358 |
+
"eval_steps_per_second": 29.146,
|
359 |
+
"step": 12500
|
360 |
+
},
|
361 |
+
{
|
362 |
+
"epoch": 1.64,
|
363 |
+
"learning_rate": 0.00035,
|
364 |
+
"loss": 0.5664,
|
365 |
+
"step": 13000
|
366 |
+
},
|
367 |
+
{
|
368 |
+
"epoch": 1.64,
|
369 |
+
"eval_loss": 0.7037546038627625,
|
370 |
+
"eval_runtime": 15.2843,
|
371 |
+
"eval_samples_per_second": 65.427,
|
372 |
+
"eval_steps_per_second": 32.713,
|
373 |
+
"step": 13000
|
374 |
+
},
|
375 |
+
{
|
376 |
+
"epoch": 1.7,
|
377 |
+
"learning_rate": 0.00032500000000000004,
|
378 |
+
"loss": 0.5534,
|
379 |
+
"step": 13500
|
380 |
+
},
|
381 |
+
{
|
382 |
+
"epoch": 1.7,
|
383 |
+
"eval_loss": 0.706645667552948,
|
384 |
+
"eval_runtime": 15.983,
|
385 |
+
"eval_samples_per_second": 62.567,
|
386 |
+
"eval_steps_per_second": 31.283,
|
387 |
+
"step": 13500
|
388 |
+
},
|
389 |
+
{
|
390 |
+
"epoch": 1.76,
|
391 |
+
"learning_rate": 0.0003,
|
392 |
+
"loss": 0.5588,
|
393 |
+
"step": 14000
|
394 |
+
},
|
395 |
+
{
|
396 |
+
"epoch": 1.76,
|
397 |
+
"eval_loss": 0.7000927329063416,
|
398 |
+
"eval_runtime": 15.4052,
|
399 |
+
"eval_samples_per_second": 64.913,
|
400 |
+
"eval_steps_per_second": 32.457,
|
401 |
+
"step": 14000
|
402 |
+
},
|
403 |
+
{
|
404 |
+
"epoch": 1.83,
|
405 |
+
"learning_rate": 0.000275,
|
406 |
+
"loss": 0.5568,
|
407 |
+
"step": 14500
|
408 |
+
},
|
409 |
+
{
|
410 |
+
"epoch": 1.83,
|
411 |
+
"eval_loss": 0.6883980631828308,
|
412 |
+
"eval_runtime": 15.226,
|
413 |
+
"eval_samples_per_second": 65.677,
|
414 |
+
"eval_steps_per_second": 32.839,
|
415 |
+
"step": 14500
|
416 |
+
},
|
417 |
+
{
|
418 |
+
"epoch": 1.89,
|
419 |
+
"learning_rate": 0.00025,
|
420 |
+
"loss": 0.5661,
|
421 |
+
"step": 15000
|
422 |
+
},
|
423 |
+
{
|
424 |
+
"epoch": 1.89,
|
425 |
+
"eval_loss": 0.6906747817993164,
|
426 |
+
"eval_runtime": 15.225,
|
427 |
+
"eval_samples_per_second": 65.681,
|
428 |
+
"eval_steps_per_second": 32.841,
|
429 |
+
"step": 15000
|
430 |
+
},
|
431 |
+
{
|
432 |
+
"epoch": 1.95,
|
433 |
+
"learning_rate": 0.00022500000000000002,
|
434 |
+
"loss": 0.536,
|
435 |
+
"step": 15500
|
436 |
+
},
|
437 |
+
{
|
438 |
+
"epoch": 1.95,
|
439 |
+
"eval_loss": 0.6931287050247192,
|
440 |
+
"eval_runtime": 15.5471,
|
441 |
+
"eval_samples_per_second": 64.321,
|
442 |
+
"eval_steps_per_second": 32.16,
|
443 |
+
"step": 15500
|
444 |
+
},
|
445 |
+
{
|
446 |
+
"epoch": 2.02,
|
447 |
+
"learning_rate": 0.0002,
|
448 |
+
"loss": 0.5042,
|
449 |
+
"step": 16000
|
450 |
+
},
|
451 |
+
{
|
452 |
+
"epoch": 2.02,
|
453 |
+
"eval_loss": 0.712044894695282,
|
454 |
+
"eval_runtime": 15.4441,
|
455 |
+
"eval_samples_per_second": 64.75,
|
456 |
+
"eval_steps_per_second": 32.375,
|
457 |
+
"step": 16000
|
458 |
+
},
|
459 |
+
{
|
460 |
+
"epoch": 2.08,
|
461 |
+
"learning_rate": 0.000175,
|
462 |
+
"loss": 0.3784,
|
463 |
+
"step": 16500
|
464 |
+
},
|
465 |
+
{
|
466 |
+
"epoch": 2.08,
|
467 |
+
"eval_loss": 0.7175341844558716,
|
468 |
+
"eval_runtime": 15.0821,
|
469 |
+
"eval_samples_per_second": 66.304,
|
470 |
+
"eval_steps_per_second": 33.152,
|
471 |
+
"step": 16500
|
472 |
+
},
|
473 |
+
{
|
474 |
+
"epoch": 2.14,
|
475 |
+
"learning_rate": 0.00015,
|
476 |
+
"loss": 0.3755,
|
477 |
+
"step": 17000
|
478 |
+
},
|
479 |
+
{
|
480 |
+
"epoch": 2.14,
|
481 |
+
"eval_loss": 0.7178177833557129,
|
482 |
+
"eval_runtime": 15.1581,
|
483 |
+
"eval_samples_per_second": 65.972,
|
484 |
+
"eval_steps_per_second": 32.986,
|
485 |
+
"step": 17000
|
486 |
+
},
|
487 |
+
{
|
488 |
+
"epoch": 2.2,
|
489 |
+
"learning_rate": 0.000125,
|
490 |
+
"loss": 0.3632,
|
491 |
+
"step": 17500
|
492 |
+
},
|
493 |
+
{
|
494 |
+
"epoch": 2.2,
|
495 |
+
"eval_loss": 0.7224026918411255,
|
496 |
+
"eval_runtime": 15.2639,
|
497 |
+
"eval_samples_per_second": 65.514,
|
498 |
+
"eval_steps_per_second": 32.757,
|
499 |
+
"step": 17500
|
500 |
+
},
|
501 |
+
{
|
502 |
+
"epoch": 2.27,
|
503 |
+
"learning_rate": 0.0001,
|
504 |
+
"loss": 0.3632,
|
505 |
+
"step": 18000
|
506 |
+
},
|
507 |
+
{
|
508 |
+
"epoch": 2.27,
|
509 |
+
"eval_loss": 0.7143183350563049,
|
510 |
+
"eval_runtime": 15.506,
|
511 |
+
"eval_samples_per_second": 64.491,
|
512 |
+
"eval_steps_per_second": 32.246,
|
513 |
+
"step": 18000
|
514 |
+
},
|
515 |
+
{
|
516 |
+
"epoch": 2.33,
|
517 |
+
"learning_rate": 7.5e-05,
|
518 |
+
"loss": 0.3614,
|
519 |
+
"step": 18500
|
520 |
+
},
|
521 |
+
{
|
522 |
+
"epoch": 2.33,
|
523 |
+
"eval_loss": 0.7125980257987976,
|
524 |
+
"eval_runtime": 15.2549,
|
525 |
+
"eval_samples_per_second": 65.552,
|
526 |
+
"eval_steps_per_second": 32.776,
|
527 |
+
"step": 18500
|
528 |
+
},
|
529 |
+
{
|
530 |
+
"epoch": 2.39,
|
531 |
+
"learning_rate": 5e-05,
|
532 |
+
"loss": 0.3838,
|
533 |
+
"step": 19000
|
534 |
+
},
|
535 |
+
{
|
536 |
+
"epoch": 2.39,
|
537 |
+
"eval_loss": 0.7115333080291748,
|
538 |
+
"eval_runtime": 15.4705,
|
539 |
+
"eval_samples_per_second": 64.639,
|
540 |
+
"eval_steps_per_second": 32.32,
|
541 |
+
"step": 19000
|
542 |
+
},
|
543 |
+
{
|
544 |
+
"epoch": 2.46,
|
545 |
+
"learning_rate": 2.5e-05,
|
546 |
+
"loss": 0.3652,
|
547 |
+
"step": 19500
|
548 |
+
},
|
549 |
+
{
|
550 |
+
"epoch": 2.46,
|
551 |
+
"eval_loss": 0.7107402086257935,
|
552 |
+
"eval_runtime": 16.4678,
|
553 |
+
"eval_samples_per_second": 60.725,
|
554 |
+
"eval_steps_per_second": 30.362,
|
555 |
+
"step": 19500
|
556 |
+
},
|
557 |
+
{
|
558 |
+
"epoch": 2.52,
|
559 |
+
"learning_rate": 0.0,
|
560 |
+
"loss": 0.3608,
|
561 |
+
"step": 20000
|
562 |
+
},
|
563 |
+
{
|
564 |
+
"epoch": 2.52,
|
565 |
+
"eval_loss": 0.7085164189338684,
|
566 |
+
"eval_runtime": 15.3333,
|
567 |
+
"eval_samples_per_second": 65.218,
|
568 |
+
"eval_steps_per_second": 32.609,
|
569 |
+
"step": 20000
|
570 |
+
}
|
571 |
+
],
|
572 |
+
"logging_steps": 500,
|
573 |
+
"max_steps": 20000,
|
574 |
+
"num_train_epochs": 3,
|
575 |
+
"save_steps": 5000,
|
576 |
+
"total_flos": 1.0899423608893747e+17,
|
577 |
+
"trial_name": null,
|
578 |
+
"trial_params": null
|
579 |
+
}
|
function-base-flatten/checkpoint-20000/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:857168942562a5574ce0eadda06d15062fa7c087954ae9df9c3df06ec62f42c4
|
3 |
+
size 4283
|
function-base-flatten/checkpoint-5000/config.json
ADDED
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "google/flan-t5-base",
|
3 |
+
"architectures": [
|
4 |
+
"FiDT5"
|
5 |
+
],
|
6 |
+
"classifier_dropout": 0.0,
|
7 |
+
"d_ff": 2048,
|
8 |
+
"d_kv": 64,
|
9 |
+
"d_model": 768,
|
10 |
+
"decoder_start_token_id": 0,
|
11 |
+
"dense_act_fn": "gelu_new",
|
12 |
+
"dropout_rate": 0.1,
|
13 |
+
"eos_token_id": 1,
|
14 |
+
"feed_forward_proj": "gated-gelu",
|
15 |
+
"initializer_factor": 1.0,
|
16 |
+
"is_encoder_decoder": true,
|
17 |
+
"is_gated_act": true,
|
18 |
+
"layer_norm_epsilon": 1e-06,
|
19 |
+
"model_type": "t5",
|
20 |
+
"n_positions": 512,
|
21 |
+
"num_decoder_layers": 12,
|
22 |
+
"num_heads": 12,
|
23 |
+
"num_layers": 12,
|
24 |
+
"output_past": true,
|
25 |
+
"pad_token_id": 0,
|
26 |
+
"relative_attention_max_distance": 128,
|
27 |
+
"relative_attention_num_buckets": 32,
|
28 |
+
"task_specific_params": {
|
29 |
+
"summarization": {
|
30 |
+
"early_stopping": true,
|
31 |
+
"length_penalty": 2.0,
|
32 |
+
"max_length": 200,
|
33 |
+
"min_length": 30,
|
34 |
+
"no_repeat_ngram_size": 3,
|
35 |
+
"num_beams": 4,
|
36 |
+
"prefix": "summarize: "
|
37 |
+
},
|
38 |
+
"translation_en_to_de": {
|
39 |
+
"early_stopping": true,
|
40 |
+
"max_length": 300,
|
41 |
+
"num_beams": 4,
|
42 |
+
"prefix": "translate English to German: "
|
43 |
+
},
|
44 |
+
"translation_en_to_fr": {
|
45 |
+
"early_stopping": true,
|
46 |
+
"max_length": 300,
|
47 |
+
"num_beams": 4,
|
48 |
+
"prefix": "translate English to French: "
|
49 |
+
},
|
50 |
+
"translation_en_to_ro": {
|
51 |
+
"early_stopping": true,
|
52 |
+
"max_length": 300,
|
53 |
+
"num_beams": 4,
|
54 |
+
"prefix": "translate English to Romanian: "
|
55 |
+
}
|
56 |
+
},
|
57 |
+
"tie_word_embeddings": false,
|
58 |
+
"torch_dtype": "float32",
|
59 |
+
"transformers_version": "4.33.1",
|
60 |
+
"use_cache": true,
|
61 |
+
"vocab_size": 32128
|
62 |
+
}
|
function-base-flatten/checkpoint-5000/generation_config.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"decoder_start_token_id": 0,
|
4 |
+
"eos_token_id": 1,
|
5 |
+
"pad_token_id": 0,
|
6 |
+
"transformers_version": "4.33.1"
|
7 |
+
}
|
function-base-flatten/checkpoint-5000/optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f0908a23754ae755048af4506000784b7466bc0b906c915f3a3ecf011c299b63
|
3 |
+
size 2372293
|
function-base-flatten/checkpoint-5000/pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:72b274ec351daa7fada5fee63ca9860df026dabfc9939fcc1d6405f0e533689e
|
3 |
+
size 990410745
|
function-base-flatten/checkpoint-5000/rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9c9d06088956111c6e20d5851243172fc06211d12a6d319b3a987552ba61c609
|
3 |
+
size 14575
|
function-base-flatten/checkpoint-5000/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cf2e868687b4ae6ba785ff2503f15cde78e0501a53544557c44b05f9fae02479
|
3 |
+
size 627
|
function-base-flatten/checkpoint-5000/trainer_state.json
ADDED
@@ -0,0 +1,159 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.6298815822625347,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 5000,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.06,
|
13 |
+
"learning_rate": 0.000975,
|
14 |
+
"loss": 1.0322,
|
15 |
+
"step": 500
|
16 |
+
},
|
17 |
+
{
|
18 |
+
"epoch": 0.06,
|
19 |
+
"eval_loss": 0.8672707676887512,
|
20 |
+
"eval_runtime": 15.3736,
|
21 |
+
"eval_samples_per_second": 65.047,
|
22 |
+
"eval_steps_per_second": 32.523,
|
23 |
+
"step": 500
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.13,
|
27 |
+
"learning_rate": 0.00095,
|
28 |
+
"loss": 0.9884,
|
29 |
+
"step": 1000
|
30 |
+
},
|
31 |
+
{
|
32 |
+
"epoch": 0.13,
|
33 |
+
"eval_loss": 0.8287830948829651,
|
34 |
+
"eval_runtime": 15.4828,
|
35 |
+
"eval_samples_per_second": 64.588,
|
36 |
+
"eval_steps_per_second": 32.294,
|
37 |
+
"step": 1000
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.19,
|
41 |
+
"learning_rate": 0.000925,
|
42 |
+
"loss": 0.9418,
|
43 |
+
"step": 1500
|
44 |
+
},
|
45 |
+
{
|
46 |
+
"epoch": 0.19,
|
47 |
+
"eval_loss": 0.8468040823936462,
|
48 |
+
"eval_runtime": 15.4288,
|
49 |
+
"eval_samples_per_second": 64.814,
|
50 |
+
"eval_steps_per_second": 32.407,
|
51 |
+
"step": 1500
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.25,
|
55 |
+
"learning_rate": 0.0009000000000000001,
|
56 |
+
"loss": 0.9232,
|
57 |
+
"step": 2000
|
58 |
+
},
|
59 |
+
{
|
60 |
+
"epoch": 0.25,
|
61 |
+
"eval_loss": 0.8540903329849243,
|
62 |
+
"eval_runtime": 15.6504,
|
63 |
+
"eval_samples_per_second": 63.896,
|
64 |
+
"eval_steps_per_second": 31.948,
|
65 |
+
"step": 2000
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.31,
|
69 |
+
"learning_rate": 0.000875,
|
70 |
+
"loss": 0.9107,
|
71 |
+
"step": 2500
|
72 |
+
},
|
73 |
+
{
|
74 |
+
"epoch": 0.31,
|
75 |
+
"eval_loss": 0.809073269367218,
|
76 |
+
"eval_runtime": 15.3867,
|
77 |
+
"eval_samples_per_second": 64.991,
|
78 |
+
"eval_steps_per_second": 32.496,
|
79 |
+
"step": 2500
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 0.38,
|
83 |
+
"learning_rate": 0.00085,
|
84 |
+
"loss": 0.9233,
|
85 |
+
"step": 3000
|
86 |
+
},
|
87 |
+
{
|
88 |
+
"epoch": 0.38,
|
89 |
+
"eval_loss": 0.8151862621307373,
|
90 |
+
"eval_runtime": 15.0851,
|
91 |
+
"eval_samples_per_second": 66.291,
|
92 |
+
"eval_steps_per_second": 33.145,
|
93 |
+
"step": 3000
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 0.44,
|
97 |
+
"learning_rate": 0.000825,
|
98 |
+
"loss": 0.9038,
|
99 |
+
"step": 3500
|
100 |
+
},
|
101 |
+
{
|
102 |
+
"epoch": 0.44,
|
103 |
+
"eval_loss": 0.8232718706130981,
|
104 |
+
"eval_runtime": 15.5073,
|
105 |
+
"eval_samples_per_second": 64.486,
|
106 |
+
"eval_steps_per_second": 32.243,
|
107 |
+
"step": 3500
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 0.5,
|
111 |
+
"learning_rate": 0.0008,
|
112 |
+
"loss": 0.8747,
|
113 |
+
"step": 4000
|
114 |
+
},
|
115 |
+
{
|
116 |
+
"epoch": 0.5,
|
117 |
+
"eval_loss": 0.7864852547645569,
|
118 |
+
"eval_runtime": 15.3637,
|
119 |
+
"eval_samples_per_second": 65.089,
|
120 |
+
"eval_steps_per_second": 32.544,
|
121 |
+
"step": 4000
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"epoch": 0.57,
|
125 |
+
"learning_rate": 0.0007750000000000001,
|
126 |
+
"loss": 0.8981,
|
127 |
+
"step": 4500
|
128 |
+
},
|
129 |
+
{
|
130 |
+
"epoch": 0.57,
|
131 |
+
"eval_loss": 0.770944356918335,
|
132 |
+
"eval_runtime": 15.3602,
|
133 |
+
"eval_samples_per_second": 65.103,
|
134 |
+
"eval_steps_per_second": 32.552,
|
135 |
+
"step": 4500
|
136 |
+
},
|
137 |
+
{
|
138 |
+
"epoch": 0.63,
|
139 |
+
"learning_rate": 0.00075,
|
140 |
+
"loss": 0.8538,
|
141 |
+
"step": 5000
|
142 |
+
},
|
143 |
+
{
|
144 |
+
"epoch": 0.63,
|
145 |
+
"eval_loss": 0.7905715107917786,
|
146 |
+
"eval_runtime": 16.0451,
|
147 |
+
"eval_samples_per_second": 62.324,
|
148 |
+
"eval_steps_per_second": 31.162,
|
149 |
+
"step": 5000
|
150 |
+
}
|
151 |
+
],
|
152 |
+
"logging_steps": 500,
|
153 |
+
"max_steps": 20000,
|
154 |
+
"num_train_epochs": 3,
|
155 |
+
"save_steps": 5000,
|
156 |
+
"total_flos": 2.716206661221581e+16,
|
157 |
+
"trial_name": null,
|
158 |
+
"trial_params": null
|
159 |
+
}
|
function-base-flatten/checkpoint-5000/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:857168942562a5574ce0eadda06d15062fa7c087954ae9df9c3df06ec62f42c4
|
3 |
+
size 4283
|
function-base/checkpoint-10000/config.json
ADDED
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "google/flan-t5-base",
|
3 |
+
"architectures": [
|
4 |
+
"FiDT5"
|
5 |
+
],
|
6 |
+
"classifier_dropout": 0.0,
|
7 |
+
"d_ff": 2048,
|
8 |
+
"d_kv": 64,
|
9 |
+
"d_model": 768,
|
10 |
+
"decoder_start_token_id": 0,
|
11 |
+
"dense_act_fn": "gelu_new",
|
12 |
+
"dropout_rate": 0.1,
|
13 |
+
"eos_token_id": 1,
|
14 |
+
"feed_forward_proj": "gated-gelu",
|
15 |
+
"initializer_factor": 1.0,
|
16 |
+
"is_encoder_decoder": true,
|
17 |
+
"is_gated_act": true,
|
18 |
+
"layer_norm_epsilon": 1e-06,
|
19 |
+
"model_type": "t5",
|
20 |
+
"n_positions": 512,
|
21 |
+
"num_decoder_layers": 12,
|
22 |
+
"num_heads": 12,
|
23 |
+
"num_layers": 12,
|
24 |
+
"output_past": true,
|
25 |
+
"pad_token_id": 0,
|
26 |
+
"relative_attention_max_distance": 128,
|
27 |
+
"relative_attention_num_buckets": 32,
|
28 |
+
"task_specific_params": {
|
29 |
+
"summarization": {
|
30 |
+
"early_stopping": true,
|
31 |
+
"length_penalty": 2.0,
|
32 |
+
"max_length": 200,
|
33 |
+
"min_length": 30,
|
34 |
+
"no_repeat_ngram_size": 3,
|
35 |
+
"num_beams": 4,
|
36 |
+
"prefix": "summarize: "
|
37 |
+
},
|
38 |
+
"translation_en_to_de": {
|
39 |
+
"early_stopping": true,
|
40 |
+
"max_length": 300,
|
41 |
+
"num_beams": 4,
|
42 |
+
"prefix": "translate English to German: "
|
43 |
+
},
|
44 |
+
"translation_en_to_fr": {
|
45 |
+
"early_stopping": true,
|
46 |
+
"max_length": 300,
|
47 |
+
"num_beams": 4,
|
48 |
+
"prefix": "translate English to French: "
|
49 |
+
},
|
50 |
+
"translation_en_to_ro": {
|
51 |
+
"early_stopping": true,
|
52 |
+
"max_length": 300,
|
53 |
+
"num_beams": 4,
|
54 |
+
"prefix": "translate English to Romanian: "
|
55 |
+
}
|
56 |
+
},
|
57 |
+
"tie_word_embeddings": false,
|
58 |
+
"torch_dtype": "float32",
|
59 |
+
"transformers_version": "4.33.1",
|
60 |
+
"use_cache": true,
|
61 |
+
"vocab_size": 32128
|
62 |
+
}
|
function-base/checkpoint-10000/generation_config.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"decoder_start_token_id": 0,
|
4 |
+
"eos_token_id": 1,
|
5 |
+
"pad_token_id": 0,
|
6 |
+
"transformers_version": "4.33.1"
|
7 |
+
}
|
function-base/checkpoint-10000/optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:31a8a4d069a6841f5bc2ea4a8f73c35b1643780ee12aea2bacb523db2f0f8014
|
3 |
+
size 2372293
|
function-base/checkpoint-10000/pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:84906aa788bf8a1e878057086eb3a908712a1f7edc9948f79bf943f0a8211a7c
|
3 |
+
size 990410745
|
function-base/checkpoint-10000/rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0a1e8dfcc14f0e4c7687859bd95ad0362e76e36c0d62c4a58370f1dfda0a5a17
|
3 |
+
size 14575
|
function-base/checkpoint-10000/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5ca84735770f2a357642bb07a692869b77f0409ed01b19bf3c890697e28d947e
|
3 |
+
size 627
|
function-base/checkpoint-10000/trainer_state.json
ADDED
@@ -0,0 +1,299 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 1.2597631645250693,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 10000,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.06,
|
13 |
+
"learning_rate": 0.000975,
|
14 |
+
"loss": 0.9774,
|
15 |
+
"step": 500
|
16 |
+
},
|
17 |
+
{
|
18 |
+
"epoch": 0.06,
|
19 |
+
"eval_loss": 0.8144938945770264,
|
20 |
+
"eval_runtime": 14.6573,
|
21 |
+
"eval_samples_per_second": 68.225,
|
22 |
+
"eval_steps_per_second": 34.113,
|
23 |
+
"step": 500
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.13,
|
27 |
+
"learning_rate": 0.00095,
|
28 |
+
"loss": 0.9617,
|
29 |
+
"step": 1000
|
30 |
+
},
|
31 |
+
{
|
32 |
+
"epoch": 0.13,
|
33 |
+
"eval_loss": 0.8077166080474854,
|
34 |
+
"eval_runtime": 15.1765,
|
35 |
+
"eval_samples_per_second": 65.891,
|
36 |
+
"eval_steps_per_second": 32.946,
|
37 |
+
"step": 1000
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.19,
|
41 |
+
"learning_rate": 0.000925,
|
42 |
+
"loss": 0.911,
|
43 |
+
"step": 1500
|
44 |
+
},
|
45 |
+
{
|
46 |
+
"epoch": 0.19,
|
47 |
+
"eval_loss": 0.8064053058624268,
|
48 |
+
"eval_runtime": 15.3268,
|
49 |
+
"eval_samples_per_second": 65.245,
|
50 |
+
"eval_steps_per_second": 32.623,
|
51 |
+
"step": 1500
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.25,
|
55 |
+
"learning_rate": 0.0009000000000000001,
|
56 |
+
"loss": 0.8954,
|
57 |
+
"step": 2000
|
58 |
+
},
|
59 |
+
{
|
60 |
+
"epoch": 0.25,
|
61 |
+
"eval_loss": 0.8163686990737915,
|
62 |
+
"eval_runtime": 14.5944,
|
63 |
+
"eval_samples_per_second": 68.519,
|
64 |
+
"eval_steps_per_second": 34.26,
|
65 |
+
"step": 2000
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.31,
|
69 |
+
"learning_rate": 0.000875,
|
70 |
+
"loss": 0.883,
|
71 |
+
"step": 2500
|
72 |
+
},
|
73 |
+
{
|
74 |
+
"epoch": 0.31,
|
75 |
+
"eval_loss": 0.8068735003471375,
|
76 |
+
"eval_runtime": 14.8614,
|
77 |
+
"eval_samples_per_second": 67.288,
|
78 |
+
"eval_steps_per_second": 33.644,
|
79 |
+
"step": 2500
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 0.38,
|
83 |
+
"learning_rate": 0.00085,
|
84 |
+
"loss": 0.8867,
|
85 |
+
"step": 3000
|
86 |
+
},
|
87 |
+
{
|
88 |
+
"epoch": 0.38,
|
89 |
+
"eval_loss": 0.7819482088088989,
|
90 |
+
"eval_runtime": 14.6896,
|
91 |
+
"eval_samples_per_second": 68.076,
|
92 |
+
"eval_steps_per_second": 34.038,
|
93 |
+
"step": 3000
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 0.44,
|
97 |
+
"learning_rate": 0.000825,
|
98 |
+
"loss": 0.8688,
|
99 |
+
"step": 3500
|
100 |
+
},
|
101 |
+
{
|
102 |
+
"epoch": 0.44,
|
103 |
+
"eval_loss": 0.8062307238578796,
|
104 |
+
"eval_runtime": 14.6856,
|
105 |
+
"eval_samples_per_second": 68.094,
|
106 |
+
"eval_steps_per_second": 34.047,
|
107 |
+
"step": 3500
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 0.5,
|
111 |
+
"learning_rate": 0.0008,
|
112 |
+
"loss": 0.8446,
|
113 |
+
"step": 4000
|
114 |
+
},
|
115 |
+
{
|
116 |
+
"epoch": 0.5,
|
117 |
+
"eval_loss": 0.7707250714302063,
|
118 |
+
"eval_runtime": 15.0517,
|
119 |
+
"eval_samples_per_second": 66.438,
|
120 |
+
"eval_steps_per_second": 33.219,
|
121 |
+
"step": 4000
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"epoch": 0.57,
|
125 |
+
"learning_rate": 0.0007750000000000001,
|
126 |
+
"loss": 0.8617,
|
127 |
+
"step": 4500
|
128 |
+
},
|
129 |
+
{
|
130 |
+
"epoch": 0.57,
|
131 |
+
"eval_loss": 0.7528353333473206,
|
132 |
+
"eval_runtime": 14.7368,
|
133 |
+
"eval_samples_per_second": 67.857,
|
134 |
+
"eval_steps_per_second": 33.929,
|
135 |
+
"step": 4500
|
136 |
+
},
|
137 |
+
{
|
138 |
+
"epoch": 0.63,
|
139 |
+
"learning_rate": 0.00075,
|
140 |
+
"loss": 0.8158,
|
141 |
+
"step": 5000
|
142 |
+
},
|
143 |
+
{
|
144 |
+
"epoch": 0.63,
|
145 |
+
"eval_loss": 0.7551385760307312,
|
146 |
+
"eval_runtime": 15.3736,
|
147 |
+
"eval_samples_per_second": 65.047,
|
148 |
+
"eval_steps_per_second": 32.523,
|
149 |
+
"step": 5000
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 0.69,
|
153 |
+
"learning_rate": 0.000725,
|
154 |
+
"loss": 0.7889,
|
155 |
+
"step": 5500
|
156 |
+
},
|
157 |
+
{
|
158 |
+
"epoch": 0.69,
|
159 |
+
"eval_loss": 0.7405046820640564,
|
160 |
+
"eval_runtime": 15.4488,
|
161 |
+
"eval_samples_per_second": 64.73,
|
162 |
+
"eval_steps_per_second": 32.365,
|
163 |
+
"step": 5500
|
164 |
+
},
|
165 |
+
{
|
166 |
+
"epoch": 0.76,
|
167 |
+
"learning_rate": 0.0007,
|
168 |
+
"loss": 0.7992,
|
169 |
+
"step": 6000
|
170 |
+
},
|
171 |
+
{
|
172 |
+
"epoch": 0.76,
|
173 |
+
"eval_loss": 0.7292428016662598,
|
174 |
+
"eval_runtime": 15.892,
|
175 |
+
"eval_samples_per_second": 62.925,
|
176 |
+
"eval_steps_per_second": 31.462,
|
177 |
+
"step": 6000
|
178 |
+
},
|
179 |
+
{
|
180 |
+
"epoch": 0.82,
|
181 |
+
"learning_rate": 0.000675,
|
182 |
+
"loss": 0.8051,
|
183 |
+
"step": 6500
|
184 |
+
},
|
185 |
+
{
|
186 |
+
"epoch": 0.82,
|
187 |
+
"eval_loss": 0.7345249056816101,
|
188 |
+
"eval_runtime": 14.8049,
|
189 |
+
"eval_samples_per_second": 67.545,
|
190 |
+
"eval_steps_per_second": 33.773,
|
191 |
+
"step": 6500
|
192 |
+
},
|
193 |
+
{
|
194 |
+
"epoch": 0.88,
|
195 |
+
"learning_rate": 0.0006500000000000001,
|
196 |
+
"loss": 0.7684,
|
197 |
+
"step": 7000
|
198 |
+
},
|
199 |
+
{
|
200 |
+
"epoch": 0.88,
|
201 |
+
"eval_loss": 0.7357723712921143,
|
202 |
+
"eval_runtime": 14.7316,
|
203 |
+
"eval_samples_per_second": 67.881,
|
204 |
+
"eval_steps_per_second": 33.941,
|
205 |
+
"step": 7000
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 0.94,
|
209 |
+
"learning_rate": 0.000625,
|
210 |
+
"loss": 0.753,
|
211 |
+
"step": 7500
|
212 |
+
},
|
213 |
+
{
|
214 |
+
"epoch": 0.94,
|
215 |
+
"eval_loss": 0.7323009371757507,
|
216 |
+
"eval_runtime": 14.6239,
|
217 |
+
"eval_samples_per_second": 68.381,
|
218 |
+
"eval_steps_per_second": 34.191,
|
219 |
+
"step": 7500
|
220 |
+
},
|
221 |
+
{
|
222 |
+
"epoch": 1.01,
|
223 |
+
"learning_rate": 0.0006,
|
224 |
+
"loss": 0.7464,
|
225 |
+
"step": 8000
|
226 |
+
},
|
227 |
+
{
|
228 |
+
"epoch": 1.01,
|
229 |
+
"eval_loss": 0.7275989651679993,
|
230 |
+
"eval_runtime": 15.2815,
|
231 |
+
"eval_samples_per_second": 65.439,
|
232 |
+
"eval_steps_per_second": 32.719,
|
233 |
+
"step": 8000
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 1.07,
|
237 |
+
"learning_rate": 0.000575,
|
238 |
+
"loss": 0.5429,
|
239 |
+
"step": 8500
|
240 |
+
},
|
241 |
+
{
|
242 |
+
"epoch": 1.07,
|
243 |
+
"eval_loss": 0.7231326103210449,
|
244 |
+
"eval_runtime": 15.5099,
|
245 |
+
"eval_samples_per_second": 64.475,
|
246 |
+
"eval_steps_per_second": 32.238,
|
247 |
+
"step": 8500
|
248 |
+
},
|
249 |
+
{
|
250 |
+
"epoch": 1.13,
|
251 |
+
"learning_rate": 0.00055,
|
252 |
+
"loss": 0.5704,
|
253 |
+
"step": 9000
|
254 |
+
},
|
255 |
+
{
|
256 |
+
"epoch": 1.13,
|
257 |
+
"eval_loss": 0.717272162437439,
|
258 |
+
"eval_runtime": 14.9897,
|
259 |
+
"eval_samples_per_second": 66.712,
|
260 |
+
"eval_steps_per_second": 33.356,
|
261 |
+
"step": 9000
|
262 |
+
},
|
263 |
+
{
|
264 |
+
"epoch": 1.2,
|
265 |
+
"learning_rate": 0.0005250000000000001,
|
266 |
+
"loss": 0.5459,
|
267 |
+
"step": 9500
|
268 |
+
},
|
269 |
+
{
|
270 |
+
"epoch": 1.2,
|
271 |
+
"eval_loss": 0.7188604474067688,
|
272 |
+
"eval_runtime": 14.7366,
|
273 |
+
"eval_samples_per_second": 67.858,
|
274 |
+
"eval_steps_per_second": 33.929,
|
275 |
+
"step": 9500
|
276 |
+
},
|
277 |
+
{
|
278 |
+
"epoch": 1.26,
|
279 |
+
"learning_rate": 0.0005,
|
280 |
+
"loss": 0.5435,
|
281 |
+
"step": 10000
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 1.26,
|
285 |
+
"eval_loss": 0.7037996053695679,
|
286 |
+
"eval_runtime": 14.5588,
|
287 |
+
"eval_samples_per_second": 68.687,
|
288 |
+
"eval_steps_per_second": 34.343,
|
289 |
+
"step": 10000
|
290 |
+
}
|
291 |
+
],
|
292 |
+
"logging_steps": 500,
|
293 |
+
"max_steps": 20000,
|
294 |
+
"num_train_epochs": 3,
|
295 |
+
"save_steps": 5000,
|
296 |
+
"total_flos": 6.278100154500096e+16,
|
297 |
+
"trial_name": null,
|
298 |
+
"trial_params": null
|
299 |
+
}
|
training_args.bin β function-base/checkpoint-10000/training_args.bin
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4219
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:82ea3910cc2b6cda26544c9471813d33e4dddc44d4ed360de38519e745497679
|
3 |
size 4219
|
function-base/checkpoint-15000/config.json
ADDED
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "google/flan-t5-base",
|
3 |
+
"architectures": [
|
4 |
+
"FiDT5"
|
5 |
+
],
|
6 |
+
"classifier_dropout": 0.0,
|
7 |
+
"d_ff": 2048,
|
8 |
+
"d_kv": 64,
|
9 |
+
"d_model": 768,
|
10 |
+
"decoder_start_token_id": 0,
|
11 |
+
"dense_act_fn": "gelu_new",
|
12 |
+
"dropout_rate": 0.1,
|
13 |
+
"eos_token_id": 1,
|
14 |
+
"feed_forward_proj": "gated-gelu",
|
15 |
+
"initializer_factor": 1.0,
|
16 |
+
"is_encoder_decoder": true,
|
17 |
+
"is_gated_act": true,
|
18 |
+
"layer_norm_epsilon": 1e-06,
|
19 |
+
"model_type": "t5",
|
20 |
+
"n_positions": 512,
|
21 |
+
"num_decoder_layers": 12,
|
22 |
+
"num_heads": 12,
|
23 |
+
"num_layers": 12,
|
24 |
+
"output_past": true,
|
25 |
+
"pad_token_id": 0,
|
26 |
+
"relative_attention_max_distance": 128,
|
27 |
+
"relative_attention_num_buckets": 32,
|
28 |
+
"task_specific_params": {
|
29 |
+
"summarization": {
|
30 |
+
"early_stopping": true,
|
31 |
+
"length_penalty": 2.0,
|
32 |
+
"max_length": 200,
|
33 |
+
"min_length": 30,
|
34 |
+
"no_repeat_ngram_size": 3,
|
35 |
+
"num_beams": 4,
|
36 |
+
"prefix": "summarize: "
|
37 |
+
},
|
38 |
+
"translation_en_to_de": {
|
39 |
+
"early_stopping": true,
|
40 |
+
"max_length": 300,
|
41 |
+
"num_beams": 4,
|
42 |
+
"prefix": "translate English to German: "
|
43 |
+
},
|
44 |
+
"translation_en_to_fr": {
|
45 |
+
"early_stopping": true,
|
46 |
+
"max_length": 300,
|
47 |
+
"num_beams": 4,
|
48 |
+
"prefix": "translate English to French: "
|
49 |
+
},
|
50 |
+
"translation_en_to_ro": {
|
51 |
+
"early_stopping": true,
|
52 |
+
"max_length": 300,
|
53 |
+
"num_beams": 4,
|
54 |
+
"prefix": "translate English to Romanian: "
|
55 |
+
}
|
56 |
+
},
|
57 |
+
"tie_word_embeddings": false,
|
58 |
+
"torch_dtype": "float32",
|
59 |
+
"transformers_version": "4.33.1",
|
60 |
+
"use_cache": true,
|
61 |
+
"vocab_size": 32128
|
62 |
+
}
|
function-base/checkpoint-15000/generation_config.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"decoder_start_token_id": 0,
|
4 |
+
"eos_token_id": 1,
|
5 |
+
"pad_token_id": 0,
|
6 |
+
"transformers_version": "4.33.1"
|
7 |
+
}
|
function-base/checkpoint-15000/optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3138aeb908f29a6626871da58849ebc4a60358542679230032fbe02f4e3658e4
|
3 |
+
size 2372293
|
function-base/checkpoint-15000/pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9e85c9b3d2de2c3e436770a52af0d80543e90725ecee7e845d35843969e56ce2
|
3 |
+
size 990410745
|
function-base/checkpoint-15000/rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:941a40eace4874c8bc2176908a7a69bb49bf28a46f7d454a00dba777352366c7
|
3 |
+
size 14575
|
function-base/checkpoint-15000/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:62b793f9ea6eba39185d82063b1e7434411e2aeca1bab5a010024f955d1696b3
|
3 |
+
size 627
|
function-base/checkpoint-15000/trainer_state.json
ADDED
@@ -0,0 +1,439 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 1.8896447467876039,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 15000,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.06,
|
13 |
+
"learning_rate": 0.000975,
|
14 |
+
"loss": 0.9774,
|
15 |
+
"step": 500
|
16 |
+
},
|
17 |
+
{
|
18 |
+
"epoch": 0.06,
|
19 |
+
"eval_loss": 0.8144938945770264,
|
20 |
+
"eval_runtime": 14.6573,
|
21 |
+
"eval_samples_per_second": 68.225,
|
22 |
+
"eval_steps_per_second": 34.113,
|
23 |
+
"step": 500
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.13,
|
27 |
+
"learning_rate": 0.00095,
|
28 |
+
"loss": 0.9617,
|
29 |
+
"step": 1000
|
30 |
+
},
|
31 |
+
{
|
32 |
+
"epoch": 0.13,
|
33 |
+
"eval_loss": 0.8077166080474854,
|
34 |
+
"eval_runtime": 15.1765,
|
35 |
+
"eval_samples_per_second": 65.891,
|
36 |
+
"eval_steps_per_second": 32.946,
|
37 |
+
"step": 1000
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.19,
|
41 |
+
"learning_rate": 0.000925,
|
42 |
+
"loss": 0.911,
|
43 |
+
"step": 1500
|
44 |
+
},
|
45 |
+
{
|
46 |
+
"epoch": 0.19,
|
47 |
+
"eval_loss": 0.8064053058624268,
|
48 |
+
"eval_runtime": 15.3268,
|
49 |
+
"eval_samples_per_second": 65.245,
|
50 |
+
"eval_steps_per_second": 32.623,
|
51 |
+
"step": 1500
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.25,
|
55 |
+
"learning_rate": 0.0009000000000000001,
|
56 |
+
"loss": 0.8954,
|
57 |
+
"step": 2000
|
58 |
+
},
|
59 |
+
{
|
60 |
+
"epoch": 0.25,
|
61 |
+
"eval_loss": 0.8163686990737915,
|
62 |
+
"eval_runtime": 14.5944,
|
63 |
+
"eval_samples_per_second": 68.519,
|
64 |
+
"eval_steps_per_second": 34.26,
|
65 |
+
"step": 2000
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.31,
|
69 |
+
"learning_rate": 0.000875,
|
70 |
+
"loss": 0.883,
|
71 |
+
"step": 2500
|
72 |
+
},
|
73 |
+
{
|
74 |
+
"epoch": 0.31,
|
75 |
+
"eval_loss": 0.8068735003471375,
|
76 |
+
"eval_runtime": 14.8614,
|
77 |
+
"eval_samples_per_second": 67.288,
|
78 |
+
"eval_steps_per_second": 33.644,
|
79 |
+
"step": 2500
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 0.38,
|
83 |
+
"learning_rate": 0.00085,
|
84 |
+
"loss": 0.8867,
|
85 |
+
"step": 3000
|
86 |
+
},
|
87 |
+
{
|
88 |
+
"epoch": 0.38,
|
89 |
+
"eval_loss": 0.7819482088088989,
|
90 |
+
"eval_runtime": 14.6896,
|
91 |
+
"eval_samples_per_second": 68.076,
|
92 |
+
"eval_steps_per_second": 34.038,
|
93 |
+
"step": 3000
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 0.44,
|
97 |
+
"learning_rate": 0.000825,
|
98 |
+
"loss": 0.8688,
|
99 |
+
"step": 3500
|
100 |
+
},
|
101 |
+
{
|
102 |
+
"epoch": 0.44,
|
103 |
+
"eval_loss": 0.8062307238578796,
|
104 |
+
"eval_runtime": 14.6856,
|
105 |
+
"eval_samples_per_second": 68.094,
|
106 |
+
"eval_steps_per_second": 34.047,
|
107 |
+
"step": 3500
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 0.5,
|
111 |
+
"learning_rate": 0.0008,
|
112 |
+
"loss": 0.8446,
|
113 |
+
"step": 4000
|
114 |
+
},
|
115 |
+
{
|
116 |
+
"epoch": 0.5,
|
117 |
+
"eval_loss": 0.7707250714302063,
|
118 |
+
"eval_runtime": 15.0517,
|
119 |
+
"eval_samples_per_second": 66.438,
|
120 |
+
"eval_steps_per_second": 33.219,
|
121 |
+
"step": 4000
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"epoch": 0.57,
|
125 |
+
"learning_rate": 0.0007750000000000001,
|
126 |
+
"loss": 0.8617,
|
127 |
+
"step": 4500
|
128 |
+
},
|
129 |
+
{
|
130 |
+
"epoch": 0.57,
|
131 |
+
"eval_loss": 0.7528353333473206,
|
132 |
+
"eval_runtime": 14.7368,
|
133 |
+
"eval_samples_per_second": 67.857,
|
134 |
+
"eval_steps_per_second": 33.929,
|
135 |
+
"step": 4500
|
136 |
+
},
|
137 |
+
{
|
138 |
+
"epoch": 0.63,
|
139 |
+
"learning_rate": 0.00075,
|
140 |
+
"loss": 0.8158,
|
141 |
+
"step": 5000
|
142 |
+
},
|
143 |
+
{
|
144 |
+
"epoch": 0.63,
|
145 |
+
"eval_loss": 0.7551385760307312,
|
146 |
+
"eval_runtime": 15.3736,
|
147 |
+
"eval_samples_per_second": 65.047,
|
148 |
+
"eval_steps_per_second": 32.523,
|
149 |
+
"step": 5000
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 0.69,
|
153 |
+
"learning_rate": 0.000725,
|
154 |
+
"loss": 0.7889,
|
155 |
+
"step": 5500
|
156 |
+
},
|
157 |
+
{
|
158 |
+
"epoch": 0.69,
|
159 |
+
"eval_loss": 0.7405046820640564,
|
160 |
+
"eval_runtime": 15.4488,
|
161 |
+
"eval_samples_per_second": 64.73,
|
162 |
+
"eval_steps_per_second": 32.365,
|
163 |
+
"step": 5500
|
164 |
+
},
|
165 |
+
{
|
166 |
+
"epoch": 0.76,
|
167 |
+
"learning_rate": 0.0007,
|
168 |
+
"loss": 0.7992,
|
169 |
+
"step": 6000
|
170 |
+
},
|
171 |
+
{
|
172 |
+
"epoch": 0.76,
|
173 |
+
"eval_loss": 0.7292428016662598,
|
174 |
+
"eval_runtime": 15.892,
|
175 |
+
"eval_samples_per_second": 62.925,
|
176 |
+
"eval_steps_per_second": 31.462,
|
177 |
+
"step": 6000
|
178 |
+
},
|
179 |
+
{
|
180 |
+
"epoch": 0.82,
|
181 |
+
"learning_rate": 0.000675,
|
182 |
+
"loss": 0.8051,
|
183 |
+
"step": 6500
|
184 |
+
},
|
185 |
+
{
|
186 |
+
"epoch": 0.82,
|
187 |
+
"eval_loss": 0.7345249056816101,
|
188 |
+
"eval_runtime": 14.8049,
|
189 |
+
"eval_samples_per_second": 67.545,
|
190 |
+
"eval_steps_per_second": 33.773,
|
191 |
+
"step": 6500
|
192 |
+
},
|
193 |
+
{
|
194 |
+
"epoch": 0.88,
|
195 |
+
"learning_rate": 0.0006500000000000001,
|
196 |
+
"loss": 0.7684,
|
197 |
+
"step": 7000
|
198 |
+
},
|
199 |
+
{
|
200 |
+
"epoch": 0.88,
|
201 |
+
"eval_loss": 0.7357723712921143,
|
202 |
+
"eval_runtime": 14.7316,
|
203 |
+
"eval_samples_per_second": 67.881,
|
204 |
+
"eval_steps_per_second": 33.941,
|
205 |
+
"step": 7000
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 0.94,
|
209 |
+
"learning_rate": 0.000625,
|
210 |
+
"loss": 0.753,
|
211 |
+
"step": 7500
|
212 |
+
},
|
213 |
+
{
|
214 |
+
"epoch": 0.94,
|
215 |
+
"eval_loss": 0.7323009371757507,
|
216 |
+
"eval_runtime": 14.6239,
|
217 |
+
"eval_samples_per_second": 68.381,
|
218 |
+
"eval_steps_per_second": 34.191,
|
219 |
+
"step": 7500
|
220 |
+
},
|
221 |
+
{
|
222 |
+
"epoch": 1.01,
|
223 |
+
"learning_rate": 0.0006,
|
224 |
+
"loss": 0.7464,
|
225 |
+
"step": 8000
|
226 |
+
},
|
227 |
+
{
|
228 |
+
"epoch": 1.01,
|
229 |
+
"eval_loss": 0.7275989651679993,
|
230 |
+
"eval_runtime": 15.2815,
|
231 |
+
"eval_samples_per_second": 65.439,
|
232 |
+
"eval_steps_per_second": 32.719,
|
233 |
+
"step": 8000
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 1.07,
|
237 |
+
"learning_rate": 0.000575,
|
238 |
+
"loss": 0.5429,
|
239 |
+
"step": 8500
|
240 |
+
},
|
241 |
+
{
|
242 |
+
"epoch": 1.07,
|
243 |
+
"eval_loss": 0.7231326103210449,
|
244 |
+
"eval_runtime": 15.5099,
|
245 |
+
"eval_samples_per_second": 64.475,
|
246 |
+
"eval_steps_per_second": 32.238,
|
247 |
+
"step": 8500
|
248 |
+
},
|
249 |
+
{
|
250 |
+
"epoch": 1.13,
|
251 |
+
"learning_rate": 0.00055,
|
252 |
+
"loss": 0.5704,
|
253 |
+
"step": 9000
|
254 |
+
},
|
255 |
+
{
|
256 |
+
"epoch": 1.13,
|
257 |
+
"eval_loss": 0.717272162437439,
|
258 |
+
"eval_runtime": 14.9897,
|
259 |
+
"eval_samples_per_second": 66.712,
|
260 |
+
"eval_steps_per_second": 33.356,
|
261 |
+
"step": 9000
|
262 |
+
},
|
263 |
+
{
|
264 |
+
"epoch": 1.2,
|
265 |
+
"learning_rate": 0.0005250000000000001,
|
266 |
+
"loss": 0.5459,
|
267 |
+
"step": 9500
|
268 |
+
},
|
269 |
+
{
|
270 |
+
"epoch": 1.2,
|
271 |
+
"eval_loss": 0.7188604474067688,
|
272 |
+
"eval_runtime": 14.7366,
|
273 |
+
"eval_samples_per_second": 67.858,
|
274 |
+
"eval_steps_per_second": 33.929,
|
275 |
+
"step": 9500
|
276 |
+
},
|
277 |
+
{
|
278 |
+
"epoch": 1.26,
|
279 |
+
"learning_rate": 0.0005,
|
280 |
+
"loss": 0.5435,
|
281 |
+
"step": 10000
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 1.26,
|
285 |
+
"eval_loss": 0.7037996053695679,
|
286 |
+
"eval_runtime": 14.5588,
|
287 |
+
"eval_samples_per_second": 68.687,
|
288 |
+
"eval_steps_per_second": 34.343,
|
289 |
+
"step": 10000
|
290 |
+
},
|
291 |
+
{
|
292 |
+
"epoch": 1.32,
|
293 |
+
"learning_rate": 0.000475,
|
294 |
+
"loss": 0.5429,
|
295 |
+
"step": 10500
|
296 |
+
},
|
297 |
+
{
|
298 |
+
"epoch": 1.32,
|
299 |
+
"eval_loss": 0.7000067234039307,
|
300 |
+
"eval_runtime": 14.629,
|
301 |
+
"eval_samples_per_second": 68.357,
|
302 |
+
"eval_steps_per_second": 34.179,
|
303 |
+
"step": 10500
|
304 |
+
},
|
305 |
+
{
|
306 |
+
"epoch": 1.39,
|
307 |
+
"learning_rate": 0.00045000000000000004,
|
308 |
+
"loss": 0.5363,
|
309 |
+
"step": 11000
|
310 |
+
},
|
311 |
+
{
|
312 |
+
"epoch": 1.39,
|
313 |
+
"eval_loss": 0.7090610861778259,
|
314 |
+
"eval_runtime": 15.5146,
|
315 |
+
"eval_samples_per_second": 64.455,
|
316 |
+
"eval_steps_per_second": 32.228,
|
317 |
+
"step": 11000
|
318 |
+
},
|
319 |
+
{
|
320 |
+
"epoch": 1.45,
|
321 |
+
"learning_rate": 0.000425,
|
322 |
+
"loss": 0.551,
|
323 |
+
"step": 11500
|
324 |
+
},
|
325 |
+
{
|
326 |
+
"epoch": 1.45,
|
327 |
+
"eval_loss": 0.6937999129295349,
|
328 |
+
"eval_runtime": 15.2752,
|
329 |
+
"eval_samples_per_second": 65.466,
|
330 |
+
"eval_steps_per_second": 32.733,
|
331 |
+
"step": 11500
|
332 |
+
},
|
333 |
+
{
|
334 |
+
"epoch": 1.51,
|
335 |
+
"learning_rate": 0.0004,
|
336 |
+
"loss": 0.5345,
|
337 |
+
"step": 12000
|
338 |
+
},
|
339 |
+
{
|
340 |
+
"epoch": 1.51,
|
341 |
+
"eval_loss": 0.6926913261413574,
|
342 |
+
"eval_runtime": 14.5585,
|
343 |
+
"eval_samples_per_second": 68.688,
|
344 |
+
"eval_steps_per_second": 34.344,
|
345 |
+
"step": 12000
|
346 |
+
},
|
347 |
+
{
|
348 |
+
"epoch": 1.57,
|
349 |
+
"learning_rate": 0.000375,
|
350 |
+
"loss": 0.5519,
|
351 |
+
"step": 12500
|
352 |
+
},
|
353 |
+
{
|
354 |
+
"epoch": 1.57,
|
355 |
+
"eval_loss": 0.6763409972190857,
|
356 |
+
"eval_runtime": 14.6685,
|
357 |
+
"eval_samples_per_second": 68.173,
|
358 |
+
"eval_steps_per_second": 34.087,
|
359 |
+
"step": 12500
|
360 |
+
},
|
361 |
+
{
|
362 |
+
"epoch": 1.64,
|
363 |
+
"learning_rate": 0.00035,
|
364 |
+
"loss": 0.5324,
|
365 |
+
"step": 13000
|
366 |
+
},
|
367 |
+
{
|
368 |
+
"epoch": 1.64,
|
369 |
+
"eval_loss": 0.6778369545936584,
|
370 |
+
"eval_runtime": 17.3898,
|
371 |
+
"eval_samples_per_second": 57.505,
|
372 |
+
"eval_steps_per_second": 28.753,
|
373 |
+
"step": 13000
|
374 |
+
},
|
375 |
+
{
|
376 |
+
"epoch": 1.7,
|
377 |
+
"learning_rate": 0.00032500000000000004,
|
378 |
+
"loss": 0.5272,
|
379 |
+
"step": 13500
|
380 |
+
},
|
381 |
+
{
|
382 |
+
"epoch": 1.7,
|
383 |
+
"eval_loss": 0.6725330948829651,
|
384 |
+
"eval_runtime": 14.6869,
|
385 |
+
"eval_samples_per_second": 68.088,
|
386 |
+
"eval_steps_per_second": 34.044,
|
387 |
+
"step": 13500
|
388 |
+
},
|
389 |
+
{
|
390 |
+
"epoch": 1.76,
|
391 |
+
"learning_rate": 0.0003,
|
392 |
+
"loss": 0.5258,
|
393 |
+
"step": 14000
|
394 |
+
},
|
395 |
+
{
|
396 |
+
"epoch": 1.76,
|
397 |
+
"eval_loss": 0.6668800115585327,
|
398 |
+
"eval_runtime": 15.0231,
|
399 |
+
"eval_samples_per_second": 66.564,
|
400 |
+
"eval_steps_per_second": 33.282,
|
401 |
+
"step": 14000
|
402 |
+
},
|
403 |
+
{
|
404 |
+
"epoch": 1.83,
|
405 |
+
"learning_rate": 0.000275,
|
406 |
+
"loss": 0.5229,
|
407 |
+
"step": 14500
|
408 |
+
},
|
409 |
+
{
|
410 |
+
"epoch": 1.83,
|
411 |
+
"eval_loss": 0.6599903106689453,
|
412 |
+
"eval_runtime": 14.6661,
|
413 |
+
"eval_samples_per_second": 68.184,
|
414 |
+
"eval_steps_per_second": 34.092,
|
415 |
+
"step": 14500
|
416 |
+
},
|
417 |
+
{
|
418 |
+
"epoch": 1.89,
|
419 |
+
"learning_rate": 0.00025,
|
420 |
+
"loss": 0.5386,
|
421 |
+
"step": 15000
|
422 |
+
},
|
423 |
+
{
|
424 |
+
"epoch": 1.89,
|
425 |
+
"eval_loss": 0.659939706325531,
|
426 |
+
"eval_runtime": 14.6708,
|
427 |
+
"eval_samples_per_second": 68.163,
|
428 |
+
"eval_steps_per_second": 34.081,
|
429 |
+
"step": 15000
|
430 |
+
}
|
431 |
+
],
|
432 |
+
"logging_steps": 500,
|
433 |
+
"max_steps": 20000,
|
434 |
+
"num_train_epochs": 3,
|
435 |
+
"save_steps": 5000,
|
436 |
+
"total_flos": 9.418761150170112e+16,
|
437 |
+
"trial_name": null,
|
438 |
+
"trial_params": null
|
439 |
+
}
|
{archived/checkpoint-10000 β function-base/checkpoint-15000}/training_args.bin
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4219
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:82ea3910cc2b6cda26544c9471813d33e4dddc44d4ed360de38519e745497679
|
3 |
size 4219
|
function-base/checkpoint-20000/config.json
ADDED
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "google/flan-t5-base",
|
3 |
+
"architectures": [
|
4 |
+
"FiDT5"
|
5 |
+
],
|
6 |
+
"classifier_dropout": 0.0,
|
7 |
+
"d_ff": 2048,
|
8 |
+
"d_kv": 64,
|
9 |
+
"d_model": 768,
|
10 |
+
"decoder_start_token_id": 0,
|
11 |
+
"dense_act_fn": "gelu_new",
|
12 |
+
"dropout_rate": 0.1,
|
13 |
+
"eos_token_id": 1,
|
14 |
+
"feed_forward_proj": "gated-gelu",
|
15 |
+
"initializer_factor": 1.0,
|
16 |
+
"is_encoder_decoder": true,
|
17 |
+
"is_gated_act": true,
|
18 |
+
"layer_norm_epsilon": 1e-06,
|
19 |
+
"model_type": "t5",
|
20 |
+
"n_positions": 512,
|
21 |
+
"num_decoder_layers": 12,
|
22 |
+
"num_heads": 12,
|
23 |
+
"num_layers": 12,
|
24 |
+
"output_past": true,
|
25 |
+
"pad_token_id": 0,
|
26 |
+
"relative_attention_max_distance": 128,
|
27 |
+
"relative_attention_num_buckets": 32,
|
28 |
+
"task_specific_params": {
|
29 |
+
"summarization": {
|
30 |
+
"early_stopping": true,
|
31 |
+
"length_penalty": 2.0,
|
32 |
+
"max_length": 200,
|
33 |
+
"min_length": 30,
|
34 |
+
"no_repeat_ngram_size": 3,
|
35 |
+
"num_beams": 4,
|
36 |
+
"prefix": "summarize: "
|
37 |
+
},
|
38 |
+
"translation_en_to_de": {
|
39 |
+
"early_stopping": true,
|
40 |
+
"max_length": 300,
|
41 |
+
"num_beams": 4,
|
42 |
+
"prefix": "translate English to German: "
|
43 |
+
},
|
44 |
+
"translation_en_to_fr": {
|
45 |
+
"early_stopping": true,
|
46 |
+
"max_length": 300,
|
47 |
+
"num_beams": 4,
|
48 |
+
"prefix": "translate English to French: "
|
49 |
+
},
|
50 |
+
"translation_en_to_ro": {
|
51 |
+
"early_stopping": true,
|
52 |
+
"max_length": 300,
|
53 |
+
"num_beams": 4,
|
54 |
+
"prefix": "translate English to Romanian: "
|
55 |
+
}
|
56 |
+
},
|
57 |
+
"tie_word_embeddings": false,
|
58 |
+
"torch_dtype": "float32",
|
59 |
+
"transformers_version": "4.33.1",
|
60 |
+
"use_cache": true,
|
61 |
+
"vocab_size": 32128
|
62 |
+
}
|