Hugging Face's logo

tyzhu
/
lmind_nq_train6000_eval6489_v1_reciteonly_qa_v3_5e-5_lora2

Safetensors
Generated from Trainer
Eval Results
Model card Files Community
tyzhu commited on
Commit
86299b1
•
1 Parent(s): 7d6bc13

Training in progress, epoch 20, checkpoint

Browse files
Files changed (17) hide show
  1. checkpoint-3937/README.md +9 -0
  2. checkpoint-3937/adapter_config.json +25 -0
  3. checkpoint-3937/adapter_model.safetensors +3 -0
  4. checkpoint-3937/added_tokens.json +5 -0
  5. checkpoint-3937/merges.txt +0 -0
  6. checkpoint-3937/optimizer.pt +3 -0
  7. checkpoint-3937/rng_state_0.pth +3 -0
  8. checkpoint-3937/rng_state_1.pth +3 -0
  9. checkpoint-3937/rng_state_2.pth +3 -0
  10. checkpoint-3937/rng_state_3.pth +3 -0
  11. checkpoint-3937/scheduler.pt +3 -0
  12. checkpoint-3937/special_tokens_map.json +14 -0
  13. checkpoint-3937/tokenizer.json +0 -0
  14. checkpoint-3937/tokenizer_config.json +43 -0
  15. checkpoint-3937/trainer_state.json +705 -0
  16. checkpoint-3937/training_args.bin +3 -0
  17. checkpoint-3937/vocab.json +0 -0
checkpoint-3937/README.md ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: peft
3
+ ---
4
+ ## Training procedure
5
+
6
+ ### Framework versions
7
+
8
+
9
+ - PEFT 0.5.0
checkpoint-3937/adapter_config.json ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_mapping": null,
3
+ "base_model_name_or_path": "Qwen/Qwen1.5-4B",
4
+ "bias": "none",
5
+ "fan_in_fan_out": false,
6
+ "inference_mode": true,
7
+ "init_lora_weights": true,
8
+ "layers_pattern": null,
9
+ "layers_to_transform": null,
10
+ "lora_alpha": 16,
11
+ "lora_dropout": 0.05,
12
+ "modules_to_save": null,
13
+ "peft_type": "LORA",
14
+ "r": 16,
15
+ "revision": null,
16
+ "target_modules": [
17
+ "q_proj",
18
+ "k_proj",
19
+ "v_proj",
20
+ "down_proj",
21
+ "gate_proj",
22
+ "up_proj"
23
+ ],
24
+ "task_type": "CAUSAL_LM"
25
+ }
checkpoint-3937/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:96c64ecf7931488de033557de49f626ffeb9e176ac8a3f81c5ed2d58a37317e4
3
+ size 112130416
checkpoint-3937/added_tokens.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "<|endoftext|>": 151643,
3
+ "<|im_end|>": 151645,
4
+ "<|im_start|>": 151644
5
+ }
checkpoint-3937/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-3937/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4bc5b73ce04b6c7eda2ace2f1a8e41cfc19b3c9249978ce102f7c073b928fdbe
3
+ size 224537202
checkpoint-3937/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f2922eab2234858c759703f7dc1eee41e4fa536b26459812f67b5e661826f38
3
+ size 15024
checkpoint-3937/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c31fdb54bc5da50de85a410b6200a75ebfc7bc55c59d3d96c539a3616bd732ec
3
+ size 15024
checkpoint-3937/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:99314abe9cd73f608c96211861b8a1afade4f5ad47f1e748048c250914bb6295
3
+ size 15024
checkpoint-3937/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c4697dd007ab2c2b65d267b556b561e17531e27beaa1edd4c40348a0c6de55eb
3
+ size 15024
checkpoint-3937/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f3cb260a6dd2c9049eaf25fca3c1c0c77bf7bd1e0b17f11d445867040dd832c5
3
+ size 1064
checkpoint-3937/special_tokens_map.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>"
5
+ ],
6
+ "eos_token": {
7
+ "content": "<|endoftext|>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false
12
+ },
13
+ "pad_token": "<|endoftext|>"
14
+ }
checkpoint-3937/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-3937/tokenizer_config.json ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "151643": {
5
+ "content": "<|endoftext|>",
6
+ "lstrip": false,
7
+ "normalized": false,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "151644": {
13
+ "content": "<|im_start|>",
14
+ "lstrip": false,
15
+ "normalized": false,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ },
20
+ "151645": {
21
+ "content": "<|im_end|>",
22
+ "lstrip": false,
23
+ "normalized": false,
24
+ "rstrip": false,
25
+ "single_word": false,
26
+ "special": true
27
+ }
28
+ },
29
+ "additional_special_tokens": [
30
+ "<|im_start|>",
31
+ "<|im_end|>"
32
+ ],
33
+ "bos_token": null,
34
+ "chat_template": "{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful assistant<|im_end|>\n' }}{% endif %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
35
+ "clean_up_tokenization_spaces": false,
36
+ "eos_token": "<|endoftext|>",
37
+ "errors": "replace",
38
+ "model_max_length": 32768,
39
+ "pad_token": "<|endoftext|>",
40
+ "split_special_tokens": false,
41
+ "tokenizer_class": "Qwen2Tokenizer",
42
+ "unk_token": null
43
+ }
checkpoint-3937/trainer_state.json ADDED
@@ -0,0 +1,705 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 20.997333333333334,
5
+ "eval_steps": 500,
6
+ "global_step": 3937,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.5333333333333333,
13
+ "grad_norm": 0.271308958530426,
14
+ "learning_rate": 5e-05,
15
+ "loss": 1.8262,
16
+ "step": 100
17
+ },
18
+ {
19
+ "epoch": 0.9973333333333333,
20
+ "eval_accuracy": 0.6047174887892377,
21
+ "eval_loss": 1.7027028799057007,
22
+ "eval_runtime": 6.6338,
23
+ "eval_samples_per_second": 75.372,
24
+ "eval_steps_per_second": 9.497,
25
+ "step": 187
26
+ },
27
+ {
28
+ "epoch": 0.9973333333333333,
29
+ "eval_exact_match": 14.6,
30
+ "eval_f1": 23.665714285714298,
31
+ "eval_qa_bleu": 10.24028390049623,
32
+ "eval_qa_exact_match": 0.114,
33
+ "eval_recite_bleu": 12.525534079250063,
34
+ "eval_recite_exact_match": 0.0,
35
+ "step": 187
36
+ },
37
+ {
38
+ "epoch": 1.0666666666666667,
39
+ "grad_norm": 0.301651269197464,
40
+ "learning_rate": 5e-05,
41
+ "loss": 1.7424,
42
+ "step": 200
43
+ },
44
+ {
45
+ "epoch": 1.6,
46
+ "grad_norm": 0.35949045419692993,
47
+ "learning_rate": 5e-05,
48
+ "loss": 1.6978,
49
+ "step": 300
50
+ },
51
+ {
52
+ "epoch": 2.0,
53
+ "eval_accuracy": 0.6068699551569506,
54
+ "eval_loss": 1.6845285892486572,
55
+ "eval_runtime": 5.7524,
56
+ "eval_samples_per_second": 86.92,
57
+ "eval_steps_per_second": 10.952,
58
+ "step": 375
59
+ },
60
+ {
61
+ "epoch": 2.0,
62
+ "eval_exact_match": 14.6,
63
+ "eval_f1": 23.74190476190477,
64
+ "eval_qa_bleu": 9.70432698897927,
65
+ "eval_qa_exact_match": 0.114,
66
+ "eval_recite_bleu": 13.014822566317982,
67
+ "eval_recite_exact_match": 0.006,
68
+ "step": 375
69
+ },
70
+ {
71
+ "epoch": 2.1333333333333333,
72
+ "grad_norm": 0.41301965713500977,
73
+ "learning_rate": 5e-05,
74
+ "loss": 1.694,
75
+ "step": 400
76
+ },
77
+ {
78
+ "epoch": 2.6666666666666665,
79
+ "grad_norm": 0.5057485699653625,
80
+ "learning_rate": 5e-05,
81
+ "loss": 1.6324,
82
+ "step": 500
83
+ },
84
+ {
85
+ "epoch": 2.997333333333333,
86
+ "eval_accuracy": 0.6079551569506726,
87
+ "eval_loss": 1.6851012706756592,
88
+ "eval_runtime": 5.5314,
89
+ "eval_samples_per_second": 90.394,
90
+ "eval_steps_per_second": 11.39,
91
+ "step": 562
92
+ },
93
+ {
94
+ "epoch": 2.997333333333333,
95
+ "eval_exact_match": 14.2,
96
+ "eval_f1": 23.29190476190477,
97
+ "eval_qa_bleu": 10.724307586340888,
98
+ "eval_qa_exact_match": 0.104,
99
+ "eval_recite_bleu": 12.83531066630853,
100
+ "eval_recite_exact_match": 0.004,
101
+ "step": 562
102
+ },
103
+ {
104
+ "epoch": 3.2,
105
+ "grad_norm": 0.6363334655761719,
106
+ "learning_rate": 5e-05,
107
+ "loss": 1.6063,
108
+ "step": 600
109
+ },
110
+ {
111
+ "epoch": 3.7333333333333334,
112
+ "grad_norm": 0.6466453671455383,
113
+ "learning_rate": 5e-05,
114
+ "loss": 1.5562,
115
+ "step": 700
116
+ },
117
+ {
118
+ "epoch": 4.0,
119
+ "eval_accuracy": 0.6071121076233184,
120
+ "eval_loss": 1.6969172954559326,
121
+ "eval_runtime": 6.2612,
122
+ "eval_samples_per_second": 79.857,
123
+ "eval_steps_per_second": 10.062,
124
+ "step": 750
125
+ },
126
+ {
127
+ "epoch": 4.0,
128
+ "eval_exact_match": 14.0,
129
+ "eval_f1": 21.34333333333334,
130
+ "eval_qa_bleu": 10.250340106635626,
131
+ "eval_qa_exact_match": 0.084,
132
+ "eval_recite_bleu": 12.741830585824024,
133
+ "eval_recite_exact_match": 0.004,
134
+ "step": 750
135
+ },
136
+ {
137
+ "epoch": 4.266666666666667,
138
+ "grad_norm": 0.8162885904312134,
139
+ "learning_rate": 5e-05,
140
+ "loss": 1.5052,
141
+ "step": 800
142
+ },
143
+ {
144
+ "epoch": 4.8,
145
+ "grad_norm": 0.8799439072608948,
146
+ "learning_rate": 5e-05,
147
+ "loss": 1.469,
148
+ "step": 900
149
+ },
150
+ {
151
+ "epoch": 4.997333333333334,
152
+ "eval_accuracy": 0.6059282511210762,
153
+ "eval_loss": 1.7216957807540894,
154
+ "eval_runtime": 6.3012,
155
+ "eval_samples_per_second": 79.35,
156
+ "eval_steps_per_second": 9.998,
157
+ "step": 937
158
+ },
159
+ {
160
+ "epoch": 4.997333333333334,
161
+ "eval_exact_match": 12.6,
162
+ "eval_f1": 21.041428571428575,
163
+ "eval_qa_bleu": 8.239984727774104,
164
+ "eval_qa_exact_match": 0.094,
165
+ "eval_recite_bleu": 12.327373697720192,
166
+ "eval_recite_exact_match": 0.002,
167
+ "step": 937
168
+ },
169
+ {
170
+ "epoch": 5.333333333333333,
171
+ "grad_norm": 0.9834449291229248,
172
+ "learning_rate": 5e-05,
173
+ "loss": 1.3845,
174
+ "step": 1000
175
+ },
176
+ {
177
+ "epoch": 5.866666666666667,
178
+ "grad_norm": 0.9377098679542542,
179
+ "learning_rate": 5e-05,
180
+ "loss": 1.3927,
181
+ "step": 1100
182
+ },
183
+ {
184
+ "epoch": 6.0,
185
+ "eval_accuracy": 0.6047354260089686,
186
+ "eval_loss": 1.7564114332199097,
187
+ "eval_runtime": 6.3447,
188
+ "eval_samples_per_second": 78.806,
189
+ "eval_steps_per_second": 9.93,
190
+ "step": 1125
191
+ },
192
+ {
193
+ "epoch": 6.0,
194
+ "eval_exact_match": 13.4,
195
+ "eval_f1": 21.70317460317461,
196
+ "eval_qa_bleu": 10.522554916579404,
197
+ "eval_qa_exact_match": 0.104,
198
+ "eval_recite_bleu": 12.383604385812577,
199
+ "eval_recite_exact_match": 0.006,
200
+ "step": 1125
201
+ },
202
+ {
203
+ "epoch": 6.4,
204
+ "grad_norm": 1.1172072887420654,
205
+ "learning_rate": 5e-05,
206
+ "loss": 1.2924,
207
+ "step": 1200
208
+ },
209
+ {
210
+ "epoch": 6.933333333333334,
211
+ "grad_norm": 1.0980473756790161,
212
+ "learning_rate": 5e-05,
213
+ "loss": 1.286,
214
+ "step": 1300
215
+ },
216
+ {
217
+ "epoch": 6.997333333333334,
218
+ "eval_accuracy": 0.6026995515695067,
219
+ "eval_loss": 1.7977122068405151,
220
+ "eval_runtime": 5.9021,
221
+ "eval_samples_per_second": 84.716,
222
+ "eval_steps_per_second": 10.674,
223
+ "step": 1312
224
+ },
225
+ {
226
+ "epoch": 6.997333333333334,
227
+ "eval_exact_match": 12.0,
228
+ "eval_f1": 20.07142857142858,
229
+ "eval_qa_bleu": 9.89685444999411,
230
+ "eval_qa_exact_match": 0.102,
231
+ "eval_recite_bleu": 12.555118604776576,
232
+ "eval_recite_exact_match": 0.008,
233
+ "step": 1312
234
+ },
235
+ {
236
+ "epoch": 7.466666666666667,
237
+ "grad_norm": 1.1534159183502197,
238
+ "learning_rate": 5e-05,
239
+ "loss": 1.1956,
240
+ "step": 1400
241
+ },
242
+ {
243
+ "epoch": 8.0,
244
+ "grad_norm": 1.1209707260131836,
245
+ "learning_rate": 5e-05,
246
+ "loss": 1.1891,
247
+ "step": 1500
248
+ },
249
+ {
250
+ "epoch": 8.0,
251
+ "eval_accuracy": 0.6010403587443947,
252
+ "eval_loss": 1.840914249420166,
253
+ "eval_runtime": 6.3016,
254
+ "eval_samples_per_second": 79.345,
255
+ "eval_steps_per_second": 9.997,
256
+ "step": 1500
257
+ },
258
+ {
259
+ "epoch": 8.0,
260
+ "eval_exact_match": 12.8,
261
+ "eval_f1": 21.598730158730167,
262
+ "eval_qa_bleu": 9.863863368837224,
263
+ "eval_qa_exact_match": 0.104,
264
+ "eval_recite_bleu": 12.09730367063489,
265
+ "eval_recite_exact_match": 0.01,
266
+ "step": 1500
267
+ },
268
+ {
269
+ "epoch": 8.533333333333333,
270
+ "grad_norm": 1.3646217584609985,
271
+ "learning_rate": 5e-05,
272
+ "loss": 1.0861,
273
+ "step": 1600
274
+ },
275
+ {
276
+ "epoch": 8.997333333333334,
277
+ "eval_accuracy": 0.5981255605381166,
278
+ "eval_loss": 1.9223005771636963,
279
+ "eval_runtime": 5.5092,
280
+ "eval_samples_per_second": 90.757,
281
+ "eval_steps_per_second": 11.435,
282
+ "step": 1687
283
+ },
284
+ {
285
+ "epoch": 8.997333333333334,
286
+ "eval_exact_match": 14.8,
287
+ "eval_f1": 22.160000000000004,
288
+ "eval_qa_bleu": 8.357776258354589,
289
+ "eval_qa_exact_match": 0.116,
290
+ "eval_recite_bleu": 12.76267532884902,
291
+ "eval_recite_exact_match": 0.01,
292
+ "step": 1687
293
+ },
294
+ {
295
+ "epoch": 9.066666666666666,
296
+ "grad_norm": 1.6948473453521729,
297
+ "learning_rate": 5e-05,
298
+ "loss": 1.0935,
299
+ "step": 1700
300
+ },
301
+ {
302
+ "epoch": 9.6,
303
+ "grad_norm": 1.6024364233016968,
304
+ "learning_rate": 5e-05,
305
+ "loss": 1.0127,
306
+ "step": 1800
307
+ },
308
+ {
309
+ "epoch": 10.0,
310
+ "eval_accuracy": 0.5973632286995516,
311
+ "eval_loss": 1.9414385557174683,
312
+ "eval_runtime": 5.5824,
313
+ "eval_samples_per_second": 89.568,
314
+ "eval_steps_per_second": 11.286,
315
+ "step": 1875
316
+ },
317
+ {
318
+ "epoch": 10.0,
319
+ "eval_exact_match": 12.4,
320
+ "eval_f1": 20.75746031746032,
321
+ "eval_qa_bleu": 8.760424886550503,
322
+ "eval_qa_exact_match": 0.088,
323
+ "eval_recite_bleu": 12.754935192029142,
324
+ "eval_recite_exact_match": 0.016,
325
+ "step": 1875
326
+ },
327
+ {
328
+ "epoch": 10.133333333333333,
329
+ "grad_norm": 1.6970032453536987,
330
+ "learning_rate": 5e-05,
331
+ "loss": 0.9788,
332
+ "step": 1900
333
+ },
334
+ {
335
+ "epoch": 10.666666666666666,
336
+ "grad_norm": 1.559332251548767,
337
+ "learning_rate": 5e-05,
338
+ "loss": 0.926,
339
+ "step": 2000
340
+ },
341
+ {
342
+ "epoch": 10.997333333333334,
343
+ "eval_accuracy": 0.5945201793721973,
344
+ "eval_loss": 2.0455663204193115,
345
+ "eval_runtime": 5.5259,
346
+ "eval_samples_per_second": 90.484,
347
+ "eval_steps_per_second": 11.401,
348
+ "step": 2062
349
+ },
350
+ {
351
+ "epoch": 10.997333333333334,
352
+ "eval_exact_match": 13.0,
353
+ "eval_f1": 20.56190476190476,
354
+ "eval_qa_bleu": 11.437402190353293,
355
+ "eval_qa_exact_match": 0.106,
356
+ "eval_recite_bleu": 13.600786696850818,
357
+ "eval_recite_exact_match": 0.022,
358
+ "step": 2062
359
+ },
360
+ {
361
+ "epoch": 11.2,
362
+ "grad_norm": 1.667855143547058,
363
+ "learning_rate": 5e-05,
364
+ "loss": 0.8872,
365
+ "step": 2100
366
+ },
367
+ {
368
+ "epoch": 11.733333333333333,
369
+ "grad_norm": 1.6040037870407104,
370
+ "learning_rate": 5e-05,
371
+ "loss": 0.8347,
372
+ "step": 2200
373
+ },
374
+ {
375
+ "epoch": 12.0,
376
+ "eval_accuracy": 0.5923946188340807,
377
+ "eval_loss": 2.116677761077881,
378
+ "eval_runtime": 6.4038,
379
+ "eval_samples_per_second": 78.078,
380
+ "eval_steps_per_second": 9.838,
381
+ "step": 2250
382
+ },
383
+ {
384
+ "epoch": 12.0,
385
+ "eval_exact_match": 13.6,
386
+ "eval_f1": 20.470981240981256,
387
+ "eval_qa_bleu": 8.968260765274712,
388
+ "eval_qa_exact_match": 0.106,
389
+ "eval_recite_bleu": 12.793280320213494,
390
+ "eval_recite_exact_match": 0.014,
391
+ "step": 2250
392
+ },
393
+ {
394
+ "epoch": 12.266666666666667,
395
+ "grad_norm": 1.794092059135437,
396
+ "learning_rate": 5e-05,
397
+ "loss": 0.8032,
398
+ "step": 2300
399
+ },
400
+ {
401
+ "epoch": 12.8,
402
+ "grad_norm": 1.8774526119232178,
403
+ "learning_rate": 5e-05,
404
+ "loss": 0.7736,
405
+ "step": 2400
406
+ },
407
+ {
408
+ "epoch": 12.997333333333334,
409
+ "eval_accuracy": 0.5902690582959641,
410
+ "eval_loss": 2.179492950439453,
411
+ "eval_runtime": 5.9945,
412
+ "eval_samples_per_second": 83.409,
413
+ "eval_steps_per_second": 10.51,
414
+ "step": 2437
415
+ },
416
+ {
417
+ "epoch": 12.997333333333334,
418
+ "eval_exact_match": 12.8,
419
+ "eval_f1": 20.780476190476197,
420
+ "eval_qa_bleu": 11.176823944321667,
421
+ "eval_qa_exact_match": 0.094,
422
+ "eval_recite_bleu": 13.292972390103905,
423
+ "eval_recite_exact_match": 0.02,
424
+ "step": 2437
425
+ },
426
+ {
427
+ "epoch": 13.333333333333334,
428
+ "grad_norm": 1.9102168083190918,
429
+ "learning_rate": 5e-05,
430
+ "loss": 0.7168,
431
+ "step": 2500
432
+ },
433
+ {
434
+ "epoch": 13.866666666666667,
435
+ "grad_norm": 1.9183365106582642,
436
+ "learning_rate": 5e-05,
437
+ "loss": 0.6903,
438
+ "step": 2600
439
+ },
440
+ {
441
+ "epoch": 14.0,
442
+ "eval_accuracy": 0.5875874439461883,
443
+ "eval_loss": 2.274404287338257,
444
+ "eval_runtime": 6.2502,
445
+ "eval_samples_per_second": 79.998,
446
+ "eval_steps_per_second": 10.08,
447
+ "step": 2625
448
+ },
449
+ {
450
+ "epoch": 14.0,
451
+ "eval_exact_match": 11.2,
452
+ "eval_f1": 18.26571428571429,
453
+ "eval_qa_bleu": 10.347091943612007,
454
+ "eval_qa_exact_match": 0.082,
455
+ "eval_recite_bleu": 12.720454352082847,
456
+ "eval_recite_exact_match": 0.014,
457
+ "step": 2625
458
+ },
459
+ {
460
+ "epoch": 14.4,
461
+ "grad_norm": 2.326387405395508,
462
+ "learning_rate": 5e-05,
463
+ "loss": 0.6312,
464
+ "step": 2700
465
+ },
466
+ {
467
+ "epoch": 14.933333333333334,
468
+ "grad_norm": 2.1555840969085693,
469
+ "learning_rate": 5e-05,
470
+ "loss": 0.6267,
471
+ "step": 2800
472
+ },
473
+ {
474
+ "epoch": 14.997333333333334,
475
+ "eval_accuracy": 0.5870582959641255,
476
+ "eval_loss": 2.338937997817993,
477
+ "eval_runtime": 6.643,
478
+ "eval_samples_per_second": 75.267,
479
+ "eval_steps_per_second": 9.484,
480
+ "step": 2812
481
+ },
482
+ {
483
+ "epoch": 14.997333333333334,
484
+ "eval_exact_match": 10.6,
485
+ "eval_f1": 18.311471861471865,
486
+ "eval_qa_bleu": 9.56012066081651,
487
+ "eval_qa_exact_match": 0.086,
488
+ "eval_recite_bleu": 12.790292247765784,
489
+ "eval_recite_exact_match": 0.022,
490
+ "step": 2812
491
+ },
492
+ {
493
+ "epoch": 15.466666666666667,
494
+ "grad_norm": 1.8364830017089844,
495
+ "learning_rate": 5e-05,
496
+ "loss": 0.5582,
497
+ "step": 2900
498
+ },
499
+ {
500
+ "epoch": 16.0,
501
+ "grad_norm": 1.9223560094833374,
502
+ "learning_rate": 5e-05,
503
+ "loss": 0.5673,
504
+ "step": 3000
505
+ },
506
+ {
507
+ "epoch": 16.0,
508
+ "eval_accuracy": 0.585067264573991,
509
+ "eval_loss": 2.4404428005218506,
510
+ "eval_runtime": 6.2254,
511
+ "eval_samples_per_second": 80.316,
512
+ "eval_steps_per_second": 10.12,
513
+ "step": 3000
514
+ },
515
+ {
516
+ "epoch": 16.0,
517
+ "eval_exact_match": 11.4,
518
+ "eval_f1": 18.553809523809527,
519
+ "eval_qa_bleu": 7.054735448183399,
520
+ "eval_qa_exact_match": 0.088,
521
+ "eval_recite_bleu": 12.41180854100743,
522
+ "eval_recite_exact_match": 0.018,
523
+ "step": 3000
524
+ },
525
+ {
526
+ "epoch": 16.533333333333335,
527
+ "grad_norm": 2.403259515762329,
528
+ "learning_rate": 5e-05,
529
+ "loss": 0.4886,
530
+ "step": 3100
531
+ },
532
+ {
533
+ "epoch": 16.997333333333334,
534
+ "eval_accuracy": 0.5844035874439462,
535
+ "eval_loss": 2.5225181579589844,
536
+ "eval_runtime": 5.9008,
537
+ "eval_samples_per_second": 84.734,
538
+ "eval_steps_per_second": 10.677,
539
+ "step": 3187
540
+ },
541
+ {
542
+ "epoch": 16.997333333333334,
543
+ "eval_exact_match": 10.6,
544
+ "eval_f1": 17.309567099567108,
545
+ "eval_qa_bleu": 7.6942420610670945,
546
+ "eval_qa_exact_match": 0.086,
547
+ "eval_recite_bleu": 11.974979252669918,
548
+ "eval_recite_exact_match": 0.016,
549
+ "step": 3187
550
+ },
551
+ {
552
+ "epoch": 17.066666666666666,
553
+ "grad_norm": 2.1279492378234863,
554
+ "learning_rate": 5e-05,
555
+ "loss": 0.4985,
556
+ "step": 3200
557
+ },
558
+ {
559
+ "epoch": 17.6,
560
+ "grad_norm": 2.377486228942871,
561
+ "learning_rate": 5e-05,
562
+ "loss": 0.4357,
563
+ "step": 3300
564
+ },
565
+ {
566
+ "epoch": 18.0,
567
+ "eval_accuracy": 0.5838026905829596,
568
+ "eval_loss": 2.590667724609375,
569
+ "eval_runtime": 5.984,
570
+ "eval_samples_per_second": 83.556,
571
+ "eval_steps_per_second": 10.528,
572
+ "step": 3375
573
+ },
574
+ {
575
+ "epoch": 18.0,
576
+ "eval_exact_match": 10.8,
577
+ "eval_f1": 18.150000000000006,
578
+ "eval_qa_bleu": 10.081752461945443,
579
+ "eval_qa_exact_match": 0.078,
580
+ "eval_recite_bleu": 12.17977433983274,
581
+ "eval_recite_exact_match": 0.018,
582
+ "step": 3375
583
+ },
584
+ {
585
+ "epoch": 18.133333333333333,
586
+ "grad_norm": 2.425731897354126,
587
+ "learning_rate": 5e-05,
588
+ "loss": 0.4424,
589
+ "step": 3400
590
+ },
591
+ {
592
+ "epoch": 18.666666666666668,
593
+ "grad_norm": 2.4585390090942383,
594
+ "learning_rate": 5e-05,
595
+ "loss": 0.3873,
596
+ "step": 3500
597
+ },
598
+ {
599
+ "epoch": 18.997333333333334,
600
+ "eval_accuracy": 0.5823228699551569,
601
+ "eval_loss": 2.6423120498657227,
602
+ "eval_runtime": 5.5641,
603
+ "eval_samples_per_second": 89.861,
604
+ "eval_steps_per_second": 11.323,
605
+ "step": 3562
606
+ },
607
+ {
608
+ "epoch": 18.997333333333334,
609
+ "eval_exact_match": 11.0,
610
+ "eval_f1": 18.006349206349213,
611
+ "eval_qa_bleu": 8.890762850928647,
612
+ "eval_qa_exact_match": 0.092,
613
+ "eval_recite_bleu": 12.489755382776915,
614
+ "eval_recite_exact_match": 0.022,
615
+ "step": 3562
616
+ },
617
+ {
618
+ "epoch": 19.2,
619
+ "grad_norm": 2.00604510307312,
620
+ "learning_rate": 5e-05,
621
+ "loss": 0.383,
622
+ "step": 3600
623
+ },
624
+ {
625
+ "epoch": 19.733333333333334,
626
+ "grad_norm": 2.393627405166626,
627
+ "learning_rate": 5e-05,
628
+ "loss": 0.3497,
629
+ "step": 3700
630
+ },
631
+ {
632
+ "epoch": 20.0,
633
+ "eval_accuracy": 0.5803587443946189,
634
+ "eval_loss": 2.7260048389434814,
635
+ "eval_runtime": 6.546,
636
+ "eval_samples_per_second": 76.382,
637
+ "eval_steps_per_second": 9.624,
638
+ "step": 3750
639
+ },
640
+ {
641
+ "epoch": 20.0,
642
+ "eval_exact_match": 9.6,
643
+ "eval_f1": 15.837460317460323,
644
+ "eval_qa_bleu": 7.780402070269215,
645
+ "eval_qa_exact_match": 0.078,
646
+ "eval_recite_bleu": 12.344573606229106,
647
+ "eval_recite_exact_match": 0.024,
648
+ "step": 3750
649
+ },
650
+ {
651
+ "epoch": 20.266666666666666,
652
+ "grad_norm": 2.1227269172668457,
653
+ "learning_rate": 5e-05,
654
+ "loss": 0.3321,
655
+ "step": 3800
656
+ },
657
+ {
658
+ "epoch": 20.8,
659
+ "grad_norm": 2.2578742504119873,
660
+ "learning_rate": 5e-05,
661
+ "loss": 0.314,
662
+ "step": 3900
663
+ },
664
+ {
665
+ "epoch": 20.997333333333334,
666
+ "eval_accuracy": 0.5801883408071749,
667
+ "eval_loss": 2.820314884185791,
668
+ "eval_runtime": 5.6037,
669
+ "eval_samples_per_second": 89.227,
670
+ "eval_steps_per_second": 11.243,
671
+ "step": 3937
672
+ },
673
+ {
674
+ "epoch": 20.997333333333334,
675
+ "eval_exact_match": 12.4,
676
+ "eval_f1": 17.80337662337662,
677
+ "eval_qa_bleu": 9.273311361506353,
678
+ "eval_qa_exact_match": 0.096,
679
+ "eval_recite_bleu": 12.39691014320001,
680
+ "eval_recite_exact_match": 0.018,
681
+ "step": 3937
682
+ }
683
+ ],
684
+ "logging_steps": 100,
685
+ "max_steps": 9350,
686
+ "num_input_tokens_seen": 0,
687
+ "num_train_epochs": 50,
688
+ "save_steps": 500,
689
+ "stateful_callbacks": {
690
+ "TrainerControl": {
691
+ "args": {
692
+ "should_epoch_stop": false,
693
+ "should_evaluate": false,
694
+ "should_log": false,
695
+ "should_save": true,
696
+ "should_training_stop": false
697
+ },
698
+ "attributes": {}
699
+ }
700
+ },
701
+ "total_flos": 3.966174565268193e+17,
702
+ "train_batch_size": 1,
703
+ "trial_name": null,
704
+ "trial_params": null
705
+ }
checkpoint-3937/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:01f9f0dc44c57fb44afb03320f4d02bf48066009e3bbcd96d8e4377b0171c5a5
3
+ size 5304
checkpoint-3937/vocab.json ADDED
The diff for this file is too large to render. See raw diff