robertou2 commited on
Commit
c6d5878
·
verified ·
1 Parent(s): 3d627c4

Upload folder using huggingface_hub

Browse files
adapter_config.json CHANGED
@@ -14,7 +14,7 @@
14
  "loftq_config": {},
15
  "lora_alpha": 112,
16
  "lora_bias": false,
17
- "lora_dropout": 0.05,
18
  "megatron_config": null,
19
  "megatron_core": "megatron.core",
20
  "modules_to_save": null,
@@ -23,10 +23,10 @@
23
  "rank_pattern": {},
24
  "revision": null,
25
  "target_modules": [
 
26
  "down_proj",
27
  "qkv_proj",
28
- "o_proj",
29
- "gate_up_proj"
30
  ],
31
  "task_type": "CAUSAL_LM",
32
  "use_dora": false,
 
14
  "loftq_config": {},
15
  "lora_alpha": 112,
16
  "lora_bias": false,
17
+ "lora_dropout": 0.1,
18
  "megatron_config": null,
19
  "megatron_core": "megatron.core",
20
  "modules_to_save": null,
 
23
  "rank_pattern": {},
24
  "revision": null,
25
  "target_modules": [
26
+ "gate_up_proj",
27
  "down_proj",
28
  "qkv_proj",
29
+ "o_proj"
 
30
  ],
31
  "task_type": "CAUSAL_LM",
32
  "use_dora": false,
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b119b381709119a0b888bea97fccc7d0584cfb2e5b9c98a2815c77e4003f0032
3
  size 322996232
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2ef79a75800f862f4e0fd92d66308bae954fcf321bde1c5dbf47d3acef07c03b
3
  size 322996232
optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:254a6199b0d6ba1480eaa9f60932c8df9951eabaf6f11f0e25e6937818e6ea0a
3
  size 646142667
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a27c659290fdac5e7ec3deaaf773d7831bb182d85a453783d32fb3b42f3b2526
3
  size 646142667
rng_state.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:051c2e5993afe9074ae7300a6dbc99d82340c681927373e494a44034e247675f
3
  size 14645
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f01e6fa000027d3e605db84f7de61a67a161a02e812c74c22a886d0bd96eb5d4
3
  size 14645
scheduler.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6f022a3f08a584f24ad0a681555fa9ed500c01a483e560f9305c7b86179a158a
3
  size 1465
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8785418585651bef269ecd0dbf2cde8c7100f2cba09bc9ef71d7b7bd8cb2c744
3
  size 1465
trainer_state.json CHANGED
@@ -2,106 +2,606 @@
2
  "best_global_step": null,
3
  "best_metric": null,
4
  "best_model_checkpoint": null,
5
- "epoch": 3.0,
6
  "eval_steps": 500,
7
- "global_step": 9,
8
  "is_hyper_param_search": false,
9
  "is_local_process_zero": true,
10
  "is_world_process_zero": true,
11
  "log_history": [
12
  {
13
- "entropy": 1.9974665641784668,
14
- "epoch": 0.3950617283950617,
15
- "grad_norm": 12.351557731628418,
16
  "learning_rate": 0.0,
17
- "loss": 2.2277,
18
- "mean_token_accuracy": 0.4728386988863349,
19
- "num_tokens": 19825.0,
20
  "step": 1
21
  },
22
  {
23
- "entropy": 2.0966694839298725,
24
- "epoch": 0.7901234567901234,
25
- "grad_norm": 14.322245597839355,
26
  "learning_rate": 3.3333333333333333e-06,
27
- "loss": 2.4305,
28
- "mean_token_accuracy": 0.4521916350349784,
29
- "num_tokens": 36505.0,
30
  "step": 2
31
  },
32
  {
33
- "entropy": 2.111897678936229,
34
- "epoch": 1.0,
35
- "grad_norm": 12.341168403625488,
36
  "learning_rate": 6.666666666666667e-06,
37
- "loss": 2.2828,
38
- "mean_token_accuracy": 0.46604182965615215,
39
- "num_tokens": 45620.0,
40
  "step": 3
41
  },
42
  {
43
- "entropy": 2.159341085702181,
44
- "epoch": 1.3950617283950617,
45
- "grad_norm": 9.26445484161377,
46
  "learning_rate": 1e-05,
47
- "loss": 2.1818,
48
- "mean_token_accuracy": 0.48002402670681477,
49
- "num_tokens": 62650.0,
50
  "step": 4
51
  },
52
  {
53
- "entropy": 2.2348405681550503,
54
- "epoch": 1.7901234567901234,
55
- "grad_norm": 7.206558704376221,
56
  "learning_rate": 1.3333333333333333e-05,
57
- "loss": 2.0205,
58
- "mean_token_accuracy": 0.495379738509655,
59
- "num_tokens": 80102.0,
60
  "step": 5
61
  },
62
  {
63
- "entropy": 2.109731309554156,
64
- "epoch": 2.0,
65
- "grad_norm": 4.857869625091553,
66
  "learning_rate": 1.6666666666666667e-05,
67
- "loss": 1.8422,
68
- "mean_token_accuracy": 0.5271621924989364,
69
- "num_tokens": 91240.0,
70
  "step": 6
71
  },
72
  {
73
- "entropy": 2.147480972111225,
74
- "epoch": 2.3950617283950617,
75
- "grad_norm": 4.716502666473389,
76
  "learning_rate": 2e-05,
77
- "loss": 1.8158,
78
- "mean_token_accuracy": 0.5386616727337241,
79
- "num_tokens": 108907.0,
80
  "step": 7
81
  },
82
  {
83
- "entropy": 2.111990097910166,
84
- "epoch": 2.7901234567901234,
85
- "grad_norm": 3.7848167419433594,
86
  "learning_rate": 2.3333333333333336e-05,
87
- "loss": 1.7498,
88
- "mean_token_accuracy": 0.546366567723453,
89
- "num_tokens": 126741.0,
90
  "step": 8
91
  },
92
  {
93
- "entropy": 2.050421209896312,
94
- "epoch": 3.0,
95
- "grad_norm": 4.006385326385498,
96
  "learning_rate": 2.6666666666666667e-05,
97
- "loss": 1.6883,
98
- "mean_token_accuracy": 0.5598909293904024,
99
- "num_tokens": 136860.0,
100
  "step": 9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
101
  }
102
  ],
103
  "logging_steps": 1,
104
- "max_steps": 9,
105
  "num_input_tokens_seen": 0,
106
  "num_train_epochs": 3,
107
  "save_steps": 1,
@@ -112,13 +612,13 @@
112
  "should_evaluate": false,
113
  "should_log": false,
114
  "should_save": true,
115
- "should_training_stop": true
116
  },
117
  "attributes": {}
118
  }
119
  },
120
- "total_flos": 4887449306167296.0,
121
- "train_batch_size": 4,
122
  "trial_name": null,
123
  "trial_params": null
124
  }
 
2
  "best_global_step": null,
3
  "best_metric": null,
4
  "best_model_checkpoint": null,
5
+ "epoch": 2.8395061728395063,
6
  "eval_steps": 500,
7
+ "global_step": 59,
8
  "is_hyper_param_search": false,
9
  "is_local_process_zero": true,
10
  "is_world_process_zero": true,
11
  "log_history": [
12
  {
13
+ "entropy": 2.1045966744422913,
14
+ "epoch": 0.04938271604938271,
15
+ "grad_norm": 17.07461929321289,
16
  "learning_rate": 0.0,
17
+ "loss": 2.4474,
18
+ "mean_token_accuracy": 0.44178610295057297,
19
+ "num_tokens": 2014.0,
20
  "step": 1
21
  },
22
  {
23
+ "entropy": 1.944445788860321,
24
+ "epoch": 0.09876543209876543,
25
+ "grad_norm": 13.418954849243164,
26
  "learning_rate": 3.3333333333333333e-06,
27
+ "loss": 2.153,
28
+ "mean_token_accuracy": 0.4885662570595741,
29
+ "num_tokens": 4384.0,
30
  "step": 2
31
  },
32
  {
33
+ "entropy": 2.1375502943992615,
34
+ "epoch": 0.14814814814814814,
35
+ "grad_norm": 13.263312339782715,
36
  "learning_rate": 6.666666666666667e-06,
37
+ "loss": 2.3265,
38
+ "mean_token_accuracy": 0.4580386020243168,
39
+ "num_tokens": 6448.0,
40
  "step": 3
41
  },
42
  {
43
+ "entropy": 2.0441721975803375,
44
+ "epoch": 0.19753086419753085,
45
+ "grad_norm": 12.733418464660645,
46
  "learning_rate": 1e-05,
47
+ "loss": 2.0155,
48
+ "mean_token_accuracy": 0.4843377247452736,
49
+ "num_tokens": 8532.0,
50
  "step": 4
51
  },
52
  {
53
+ "entropy": 2.224354177713394,
54
+ "epoch": 0.24691358024691357,
55
+ "grad_norm": 6.445976734161377,
56
  "learning_rate": 1.3333333333333333e-05,
57
+ "loss": 1.833,
58
+ "mean_token_accuracy": 0.4942014664411545,
59
+ "num_tokens": 11973.0,
60
  "step": 5
61
  },
62
  {
63
+ "entropy": 2.1060016453266144,
64
+ "epoch": 0.2962962962962963,
65
+ "grad_norm": 6.344198226928711,
66
  "learning_rate": 1.6666666666666667e-05,
67
+ "loss": 1.9109,
68
+ "mean_token_accuracy": 0.5272197648882866,
69
+ "num_tokens": 14982.0,
70
  "step": 6
71
  },
72
  {
73
+ "entropy": 2.2028936445713043,
74
+ "epoch": 0.345679012345679,
75
+ "grad_norm": 7.122912406921387,
76
  "learning_rate": 2e-05,
77
+ "loss": 1.8414,
78
+ "mean_token_accuracy": 0.5166565254330635,
79
+ "num_tokens": 17566.0,
80
  "step": 7
81
  },
82
  {
83
+ "entropy": 2.2383397221565247,
84
+ "epoch": 0.3950617283950617,
85
+ "grad_norm": 6.923185348510742,
86
  "learning_rate": 2.3333333333333336e-05,
87
+ "loss": 1.9645,
88
+ "mean_token_accuracy": 0.4945147931575775,
89
+ "num_tokens": 19825.0,
90
  "step": 8
91
  },
92
  {
93
+ "entropy": 2.2687272429466248,
94
+ "epoch": 0.4444444444444444,
95
+ "grad_norm": 6.454328536987305,
96
  "learning_rate": 2.6666666666666667e-05,
97
+ "loss": 1.9193,
98
+ "mean_token_accuracy": 0.520270973443985,
99
+ "num_tokens": 21855.0,
100
  "step": 9
101
+ },
102
+ {
103
+ "entropy": 2.137419432401657,
104
+ "epoch": 0.49382716049382713,
105
+ "grad_norm": 6.179715156555176,
106
+ "learning_rate": 3e-05,
107
+ "loss": 1.7229,
108
+ "mean_token_accuracy": 0.5469484068453312,
109
+ "num_tokens": 24405.0,
110
+ "step": 10
111
+ },
112
+ {
113
+ "entropy": 2.153471812605858,
114
+ "epoch": 0.5432098765432098,
115
+ "grad_norm": 6.750667572021484,
116
+ "learning_rate": 3.3333333333333335e-05,
117
+ "loss": 1.7749,
118
+ "mean_token_accuracy": 0.5488381646573544,
119
+ "num_tokens": 26247.0,
120
+ "step": 11
121
+ },
122
+ {
123
+ "entropy": 2.0210573077201843,
124
+ "epoch": 0.5925925925925926,
125
+ "grad_norm": 5.751638412475586,
126
+ "learning_rate": 3.6666666666666666e-05,
127
+ "loss": 1.7508,
128
+ "mean_token_accuracy": 0.5483265519142151,
129
+ "num_tokens": 28315.0,
130
+ "step": 12
131
+ },
132
+ {
133
+ "entropy": 2.0588729083538055,
134
+ "epoch": 0.6419753086419753,
135
+ "grad_norm": 6.749085903167725,
136
+ "learning_rate": 4e-05,
137
+ "loss": 1.78,
138
+ "mean_token_accuracy": 0.5516550652682781,
139
+ "num_tokens": 30028.0,
140
+ "step": 13
141
+ },
142
+ {
143
+ "entropy": 1.8380699753761292,
144
+ "epoch": 0.691358024691358,
145
+ "grad_norm": 4.831390380859375,
146
+ "learning_rate": 4.3333333333333334e-05,
147
+ "loss": 1.639,
148
+ "mean_token_accuracy": 0.5657348707318306,
149
+ "num_tokens": 33151.0,
150
+ "step": 14
151
+ },
152
+ {
153
+ "entropy": 2.002317950129509,
154
+ "epoch": 0.7407407407407407,
155
+ "grad_norm": 6.0173773765563965,
156
+ "learning_rate": 4.666666666666667e-05,
157
+ "loss": 1.6788,
158
+ "mean_token_accuracy": 0.5597697757184505,
159
+ "num_tokens": 35005.0,
160
+ "step": 15
161
+ },
162
+ {
163
+ "entropy": 2.2061654031276703,
164
+ "epoch": 0.7901234567901234,
165
+ "grad_norm": 6.6986517906188965,
166
+ "learning_rate": 5e-05,
167
+ "loss": 1.7774,
168
+ "mean_token_accuracy": 0.5585966594517231,
169
+ "num_tokens": 36505.0,
170
+ "step": 16
171
+ },
172
+ {
173
+ "entropy": 2.003757819533348,
174
+ "epoch": 0.8395061728395061,
175
+ "grad_norm": 6.035629749298096,
176
+ "learning_rate": 4.994647308096509e-05,
177
+ "loss": 1.657,
178
+ "mean_token_accuracy": 0.5558704920113087,
179
+ "num_tokens": 38467.0,
180
+ "step": 17
181
+ },
182
+ {
183
+ "entropy": 1.9430343508720398,
184
+ "epoch": 0.8888888888888888,
185
+ "grad_norm": 5.070599555969238,
186
+ "learning_rate": 4.9786121534345265e-05,
187
+ "loss": 1.6259,
188
+ "mean_token_accuracy": 0.5617526397109032,
189
+ "num_tokens": 40889.0,
190
+ "step": 18
191
+ },
192
+ {
193
+ "entropy": 2.037708967924118,
194
+ "epoch": 0.9382716049382716,
195
+ "grad_norm": 5.155577182769775,
196
+ "learning_rate": 4.951963201008076e-05,
197
+ "loss": 1.5762,
198
+ "mean_token_accuracy": 0.5841848030686378,
199
+ "num_tokens": 43002.0,
200
+ "step": 19
201
+ },
202
+ {
203
+ "entropy": 2.0400605648756027,
204
+ "epoch": 0.9876543209876543,
205
+ "grad_norm": 4.773712635040283,
206
+ "learning_rate": 4.914814565722671e-05,
207
+ "loss": 1.5847,
208
+ "mean_token_accuracy": 0.5751754343509674,
209
+ "num_tokens": 45100.0,
210
+ "step": 20
211
+ },
212
+ {
213
+ "entropy": 1.6524975299835205,
214
+ "epoch": 1.0,
215
+ "grad_norm": 10.611732482910156,
216
+ "learning_rate": 4.867325323737765e-05,
217
+ "loss": 1.3868,
218
+ "mean_token_accuracy": 0.6146115064620972,
219
+ "num_tokens": 45620.0,
220
+ "step": 21
221
+ },
222
+ {
223
+ "entropy": 1.9145531356334686,
224
+ "epoch": 1.0493827160493827,
225
+ "grad_norm": 4.789498805999756,
226
+ "learning_rate": 4.8096988312782174e-05,
227
+ "loss": 1.2313,
228
+ "mean_token_accuracy": 0.669404037296772,
229
+ "num_tokens": 47479.0,
230
+ "step": 22
231
+ },
232
+ {
233
+ "entropy": 1.7801827937364578,
234
+ "epoch": 1.0987654320987654,
235
+ "grad_norm": 4.739726543426514,
236
+ "learning_rate": 4.742181853831721e-05,
237
+ "loss": 1.26,
238
+ "mean_token_accuracy": 0.652519591152668,
239
+ "num_tokens": 49530.0,
240
+ "step": 23
241
+ },
242
+ {
243
+ "entropy": 1.6698480248451233,
244
+ "epoch": 1.1481481481481481,
245
+ "grad_norm": 4.931276798248291,
246
+ "learning_rate": 4.665063509461097e-05,
247
+ "loss": 1.2183,
248
+ "mean_token_accuracy": 0.6311887428164482,
249
+ "num_tokens": 51795.0,
250
+ "step": 24
251
+ },
252
+ {
253
+ "entropy": 1.6189886331558228,
254
+ "epoch": 1.1975308641975309,
255
+ "grad_norm": 4.8303542137146,
256
+ "learning_rate": 4.5786740307563636e-05,
257
+ "loss": 1.2323,
258
+ "mean_token_accuracy": 0.6584680005908012,
259
+ "num_tokens": 53775.0,
260
+ "step": 25
261
+ },
262
+ {
263
+ "entropy": 1.7246621996164322,
264
+ "epoch": 1.2469135802469136,
265
+ "grad_norm": 4.690735340118408,
266
+ "learning_rate": 4.4833833507280884e-05,
267
+ "loss": 1.2359,
268
+ "mean_token_accuracy": 0.6399939209222794,
269
+ "num_tokens": 56079.0,
270
+ "step": 26
271
+ },
272
+ {
273
+ "entropy": 1.5530678927898407,
274
+ "epoch": 1.2962962962962963,
275
+ "grad_norm": 4.586762428283691,
276
+ "learning_rate": 4.379599518697444e-05,
277
+ "loss": 1.2253,
278
+ "mean_token_accuracy": 0.6523294448852539,
279
+ "num_tokens": 58349.0,
280
+ "step": 27
281
+ },
282
+ {
283
+ "entropy": 1.551697015762329,
284
+ "epoch": 1.345679012345679,
285
+ "grad_norm": 4.633914947509766,
286
+ "learning_rate": 4.267766952966369e-05,
287
+ "loss": 1.1635,
288
+ "mean_token_accuracy": 0.6726551949977875,
289
+ "num_tokens": 60570.0,
290
+ "step": 28
291
+ },
292
+ {
293
+ "entropy": 1.6292523592710495,
294
+ "epoch": 1.3950617283950617,
295
+ "grad_norm": 5.112277507781982,
296
+ "learning_rate": 4.148364537750172e-05,
297
+ "loss": 1.2572,
298
+ "mean_token_accuracy": 0.6387738659977913,
299
+ "num_tokens": 62650.0,
300
+ "step": 29
301
+ },
302
+ {
303
+ "entropy": 1.4799976348876953,
304
+ "epoch": 1.4444444444444444,
305
+ "grad_norm": 4.31527042388916,
306
+ "learning_rate": 4.021903572521802e-05,
307
+ "loss": 1.1865,
308
+ "mean_token_accuracy": 0.6671764701604843,
309
+ "num_tokens": 65419.0,
310
+ "step": 30
311
+ },
312
+ {
313
+ "entropy": 1.549431949853897,
314
+ "epoch": 1.4938271604938271,
315
+ "grad_norm": 4.670032024383545,
316
+ "learning_rate": 3.888925582549006e-05,
317
+ "loss": 1.0756,
318
+ "mean_token_accuracy": 0.6905995309352875,
319
+ "num_tokens": 67644.0,
320
+ "step": 31
321
+ },
322
+ {
323
+ "entropy": 1.5958464741706848,
324
+ "epoch": 1.5432098765432098,
325
+ "grad_norm": 5.251377582550049,
326
+ "learning_rate": 3.7500000000000003e-05,
327
+ "loss": 1.1905,
328
+ "mean_token_accuracy": 0.6532313749194145,
329
+ "num_tokens": 69764.0,
330
+ "step": 32
331
+ },
332
+ {
333
+ "entropy": 1.6538591533899307,
334
+ "epoch": 1.5925925925925926,
335
+ "grad_norm": 5.467432022094727,
336
+ "learning_rate": 3.6057217255475034e-05,
337
+ "loss": 1.1792,
338
+ "mean_token_accuracy": 0.6628076210618019,
339
+ "num_tokens": 71383.0,
340
+ "step": 33
341
+ },
342
+ {
343
+ "entropy": 1.5740457475185394,
344
+ "epoch": 1.6419753086419753,
345
+ "grad_norm": 5.274950981140137,
346
+ "learning_rate": 3.456708580912725e-05,
347
+ "loss": 1.1839,
348
+ "mean_token_accuracy": 0.6636194065213203,
349
+ "num_tokens": 73330.0,
350
+ "step": 34
351
+ },
352
+ {
353
+ "entropy": 1.4796217530965805,
354
+ "epoch": 1.691358024691358,
355
+ "grad_norm": 4.721081256866455,
356
+ "learning_rate": 3.303598663257904e-05,
357
+ "loss": 1.0261,
358
+ "mean_token_accuracy": 0.6876147240400314,
359
+ "num_tokens": 75429.0,
360
+ "step": 35
361
+ },
362
+ {
363
+ "entropy": 1.4072065949440002,
364
+ "epoch": 1.7407407407407407,
365
+ "grad_norm": 4.086071014404297,
366
+ "learning_rate": 3.147047612756302e-05,
367
+ "loss": 1.0271,
368
+ "mean_token_accuracy": 0.6869702711701393,
369
+ "num_tokens": 78459.0,
370
+ "step": 36
371
+ },
372
+ {
373
+ "entropy": 1.5171602815389633,
374
+ "epoch": 1.7901234567901234,
375
+ "grad_norm": 5.711819648742676,
376
+ "learning_rate": 2.9877258050403212e-05,
377
+ "loss": 1.1432,
378
+ "mean_token_accuracy": 0.6735518574714661,
379
+ "num_tokens": 80102.0,
380
+ "step": 37
381
+ },
382
+ {
383
+ "entropy": 1.4996187686920166,
384
+ "epoch": 1.8395061728395061,
385
+ "grad_norm": 4.804397106170654,
386
+ "learning_rate": 2.8263154805501297e-05,
387
+ "loss": 1.1221,
388
+ "mean_token_accuracy": 0.6675888076424599,
389
+ "num_tokens": 82277.0,
390
+ "step": 38
391
+ },
392
+ {
393
+ "entropy": 1.277234360575676,
394
+ "epoch": 1.8888888888888888,
395
+ "grad_norm": 3.7578635215759277,
396
+ "learning_rate": 2.663507823075358e-05,
397
+ "loss": 0.9613,
398
+ "mean_token_accuracy": 0.6974349915981293,
399
+ "num_tokens": 85933.0,
400
+ "step": 39
401
+ },
402
+ {
403
+ "entropy": 1.4796784967184067,
404
+ "epoch": 1.9382716049382716,
405
+ "grad_norm": 4.373854160308838,
406
+ "learning_rate": 2.5e-05,
407
+ "loss": 1.1684,
408
+ "mean_token_accuracy": 0.6698590070009232,
409
+ "num_tokens": 88612.0,
410
+ "step": 40
411
+ },
412
+ {
413
+ "entropy": 1.4071252942085266,
414
+ "epoch": 1.9876543209876543,
415
+ "grad_norm": 5.221538543701172,
416
+ "learning_rate": 2.3364921769246423e-05,
417
+ "loss": 1.1439,
418
+ "mean_token_accuracy": 0.6814302429556847,
419
+ "num_tokens": 90823.0,
420
+ "step": 41
421
+ },
422
+ {
423
+ "entropy": 1.5970491170883179,
424
+ "epoch": 2.0,
425
+ "grad_norm": 10.636951446533203,
426
+ "learning_rate": 2.173684519449872e-05,
427
+ "loss": 1.0735,
428
+ "mean_token_accuracy": 0.7276268303394318,
429
+ "num_tokens": 91240.0,
430
+ "step": 42
431
+ },
432
+ {
433
+ "entropy": 1.39203742146492,
434
+ "epoch": 2.049382716049383,
435
+ "grad_norm": 3.8262007236480713,
436
+ "learning_rate": 2.0122741949596797e-05,
437
+ "loss": 0.7384,
438
+ "mean_token_accuracy": 0.7718699052929878,
439
+ "num_tokens": 93342.0,
440
+ "step": 43
441
+ },
442
+ {
443
+ "entropy": 1.4726413637399673,
444
+ "epoch": 2.0987654320987654,
445
+ "grad_norm": 4.435136795043945,
446
+ "learning_rate": 1.852952387243698e-05,
447
+ "loss": 0.7265,
448
+ "mean_token_accuracy": 0.7925485372543335,
449
+ "num_tokens": 95250.0,
450
+ "step": 44
451
+ },
452
+ {
453
+ "entropy": 1.3003774285316467,
454
+ "epoch": 2.148148148148148,
455
+ "grad_norm": 3.799497365951538,
456
+ "learning_rate": 1.6964013367420966e-05,
457
+ "loss": 0.7015,
458
+ "mean_token_accuracy": 0.7795400172472,
459
+ "num_tokens": 97864.0,
460
+ "step": 45
461
+ },
462
+ {
463
+ "entropy": 1.248021274805069,
464
+ "epoch": 2.197530864197531,
465
+ "grad_norm": 3.650768756866455,
466
+ "learning_rate": 1.5432914190872757e-05,
467
+ "loss": 0.6776,
468
+ "mean_token_accuracy": 0.8115019127726555,
469
+ "num_tokens": 100323.0,
470
+ "step": 46
471
+ },
472
+ {
473
+ "entropy": 1.2092759758234024,
474
+ "epoch": 2.246913580246914,
475
+ "grad_norm": 3.9729771614074707,
476
+ "learning_rate": 1.3942782744524973e-05,
477
+ "loss": 0.6091,
478
+ "mean_token_accuracy": 0.8202343359589577,
479
+ "num_tokens": 102615.0,
480
+ "step": 47
481
+ },
482
+ {
483
+ "entropy": 1.2965480834245682,
484
+ "epoch": 2.2962962962962963,
485
+ "grad_norm": 4.632357597351074,
486
+ "learning_rate": 1.2500000000000006e-05,
487
+ "loss": 0.7704,
488
+ "mean_token_accuracy": 0.7770233377814293,
489
+ "num_tokens": 104543.0,
490
+ "step": 48
491
+ },
492
+ {
493
+ "entropy": 1.2238061875104904,
494
+ "epoch": 2.3456790123456788,
495
+ "grad_norm": 4.053952217102051,
496
+ "learning_rate": 1.1110744174509952e-05,
497
+ "loss": 0.6047,
498
+ "mean_token_accuracy": 0.8104883581399918,
499
+ "num_tokens": 106982.0,
500
+ "step": 49
501
+ },
502
+ {
503
+ "entropy": 1.2117676883935928,
504
+ "epoch": 2.3950617283950617,
505
+ "grad_norm": 4.814194202423096,
506
+ "learning_rate": 9.780964274781984e-06,
507
+ "loss": 0.6945,
508
+ "mean_token_accuracy": 0.794559009373188,
509
+ "num_tokens": 108907.0,
510
+ "step": 50
511
+ },
512
+ {
513
+ "entropy": 1.2279776483774185,
514
+ "epoch": 2.4444444444444446,
515
+ "grad_norm": 5.624629974365234,
516
+ "learning_rate": 8.51635462249828e-06,
517
+ "loss": 0.7707,
518
+ "mean_token_accuracy": 0.7722515612840652,
519
+ "num_tokens": 110649.0,
520
+ "step": 51
521
+ },
522
+ {
523
+ "entropy": 1.130656473338604,
524
+ "epoch": 2.493827160493827,
525
+ "grad_norm": 4.165122985839844,
526
+ "learning_rate": 7.3223304703363135e-06,
527
+ "loss": 0.6238,
528
+ "mean_token_accuracy": 0.8032507002353668,
529
+ "num_tokens": 113000.0,
530
+ "step": 52
531
+ },
532
+ {
533
+ "entropy": 1.057595133781433,
534
+ "epoch": 2.5432098765432096,
535
+ "grad_norm": 4.7867817878723145,
536
+ "learning_rate": 6.204004813025568e-06,
537
+ "loss": 0.6951,
538
+ "mean_token_accuracy": 0.8001823052763939,
539
+ "num_tokens": 115252.0,
540
+ "step": 53
541
+ },
542
+ {
543
+ "entropy": 1.0031295642256737,
544
+ "epoch": 2.5925925925925926,
545
+ "grad_norm": 4.251824855804443,
546
+ "learning_rate": 5.166166492719124e-06,
547
+ "loss": 0.547,
548
+ "mean_token_accuracy": 0.8263749331235886,
549
+ "num_tokens": 117622.0,
550
+ "step": 54
551
+ },
552
+ {
553
+ "entropy": 0.9957862570881844,
554
+ "epoch": 2.6419753086419755,
555
+ "grad_norm": 4.6617279052734375,
556
+ "learning_rate": 4.213259692436367e-06,
557
+ "loss": 0.6031,
558
+ "mean_token_accuracy": 0.8142141699790955,
559
+ "num_tokens": 120361.0,
560
+ "step": 55
561
+ },
562
+ {
563
+ "entropy": 1.0976142808794975,
564
+ "epoch": 2.691358024691358,
565
+ "grad_norm": 4.915057182312012,
566
+ "learning_rate": 3.3493649053890326e-06,
567
+ "loss": 0.6949,
568
+ "mean_token_accuracy": 0.7822957187891006,
569
+ "num_tokens": 122610.0,
570
+ "step": 56
571
+ },
572
+ {
573
+ "entropy": 1.0492280274629593,
574
+ "epoch": 2.7407407407407405,
575
+ "grad_norm": 5.099884033203125,
576
+ "learning_rate": 2.578181461682794e-06,
577
+ "loss": 0.6004,
578
+ "mean_token_accuracy": 0.8302762731909752,
579
+ "num_tokens": 124571.0,
580
+ "step": 57
581
+ },
582
+ {
583
+ "entropy": 1.0981549099087715,
584
+ "epoch": 2.7901234567901234,
585
+ "grad_norm": 5.698803901672363,
586
+ "learning_rate": 1.9030116872178316e-06,
587
+ "loss": 0.8066,
588
+ "mean_token_accuracy": 0.7803784757852554,
589
+ "num_tokens": 126741.0,
590
+ "step": 58
591
+ },
592
+ {
593
+ "entropy": 1.0801162794232368,
594
+ "epoch": 2.8395061728395063,
595
+ "grad_norm": 4.8652567863464355,
596
+ "learning_rate": 1.3267467626223606e-06,
597
+ "loss": 0.6778,
598
+ "mean_token_accuracy": 0.8027013018727303,
599
+ "num_tokens": 129201.0,
600
+ "step": 59
601
  }
602
  ],
603
  "logging_steps": 1,
604
+ "max_steps": 63,
605
  "num_input_tokens_seen": 0,
606
  "num_train_epochs": 3,
607
  "save_steps": 1,
 
612
  "should_evaluate": false,
613
  "should_log": false,
614
  "should_save": true,
615
+ "should_training_stop": false
616
  },
617
  "attributes": {}
618
  }
619
  },
620
+ "total_flos": 3467511535116288.0,
621
+ "train_batch_size": 2,
622
  "trial_name": null,
623
  "trial_params": null
624
  }
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:4ef8ede9916fdd18e29704a7622df8d197558d4224803d9af4975589ca35ab78
3
  size 6353
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6450a7549e5c3d80ba33c4a8d5dfdf16bf8472ec89ae7d85c7f2650870f43667
3
  size 6353