Hugging Face's logo

tyzhu
/
lmind_nq_train6000_eval6489_v1_docidx_v3_5e-4_lora2

Safetensors
Generated from Trainer
Eval Results
Model card Files Community
tyzhu commited on
Commit
6633f9e
•
1 Parent(s): ba52384

Training in progress, epoch 24, checkpoint

Browse files
Files changed (17) hide show
  1. checkpoint-8196/README.md +9 -0
  2. checkpoint-8196/adapter_config.json +25 -0
  3. checkpoint-8196/adapter_model.safetensors +3 -0
  4. checkpoint-8196/added_tokens.json +5 -0
  5. checkpoint-8196/merges.txt +0 -0
  6. checkpoint-8196/optimizer.pt +3 -0
  7. checkpoint-8196/rng_state_0.pth +3 -0
  8. checkpoint-8196/rng_state_1.pth +3 -0
  9. checkpoint-8196/rng_state_2.pth +3 -0
  10. checkpoint-8196/rng_state_3.pth +3 -0
  11. checkpoint-8196/scheduler.pt +3 -0
  12. checkpoint-8196/special_tokens_map.json +14 -0
  13. checkpoint-8196/tokenizer.json +0 -0
  14. checkpoint-8196/tokenizer_config.json +43 -0
  15. checkpoint-8196/trainer_state.json +984 -0
  16. checkpoint-8196/training_args.bin +3 -0
  17. checkpoint-8196/vocab.json +0 -0
checkpoint-8196/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-8196/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-8196/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6a9f9b1f02b3e3747bc3eccd0af14177fe840a2c7dd8b0a8dcb88a8506bddb6e
3
+ size 112130416
checkpoint-8196/added_tokens.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "<|endoftext|>": 151643,
3
+ "<|im_end|>": 151645,
4
+ "<|im_start|>": 151644
5
+ }
checkpoint-8196/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-8196/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4707a9c82934453b73858997ef22b841c4a91049ba3576ddd6e45b044795e1cc
3
+ size 224537202
checkpoint-8196/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3700079582c3b8da2b04ac4369cf9dad95b579dbf08e14dcb12b009179265c21
3
+ size 15024
checkpoint-8196/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:293a40cd78722af015c6f3706088dc44be08fc3a0c188d928f53a88ae05d7585
3
+ size 15024
checkpoint-8196/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8e9dbd8c63f8e88e4c861ad8e6f3fcd8fdda5dd2deada91aa3a1be43b60cbfaa
3
+ size 15024
checkpoint-8196/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:54dd902a3ab2aeecca3a3977e1aa3fb6eacb0ec3baaf08733711974b5b1946e4
3
+ size 15024
checkpoint-8196/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:13088ca5501616d4de532a7d8dd14229a57c0c7766783da58603794926992809
3
+ size 1064
checkpoint-8196/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-8196/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-8196/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-8196/trainer_state.json ADDED
@@ -0,0 +1,984 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 24.0,
5
+ "eval_steps": 500,
6
+ "global_step": 8196,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.29282576866764276,
13
+ "grad_norm": 0.2425622195005417,
14
+ "learning_rate": 0.0005,
15
+ "loss": 1.972,
16
+ "step": 100
17
+ },
18
+ {
19
+ "epoch": 0.5856515373352855,
20
+ "grad_norm": 0.24936461448669434,
21
+ "learning_rate": 0.0005,
22
+ "loss": 1.954,
23
+ "step": 200
24
+ },
25
+ {
26
+ "epoch": 0.8784773060029283,
27
+ "grad_norm": 0.28534355759620667,
28
+ "learning_rate": 0.0005,
29
+ "loss": 1.9636,
30
+ "step": 300
31
+ },
32
+ {
33
+ "epoch": 0.9985358711566618,
34
+ "eval_accuracy": 0.4142051282051282,
35
+ "eval_loss": 4.36731481552124,
36
+ "eval_runtime": 5.2053,
37
+ "eval_samples_per_second": 96.056,
38
+ "eval_steps_per_second": 12.103,
39
+ "step": 341
40
+ },
41
+ {
42
+ "epoch": 0.9985358711566618,
43
+ "eval_bleu": 0.16271683142618718,
44
+ "eval_exact_match": 0.0,
45
+ "eval_prefix_exact_match": 0.032,
46
+ "step": 341
47
+ },
48
+ {
49
+ "epoch": 1.171303074670571,
50
+ "grad_norm": 0.39824551343917847,
51
+ "learning_rate": 0.0005,
52
+ "loss": 1.8098,
53
+ "step": 400
54
+ },
55
+ {
56
+ "epoch": 1.4641288433382138,
57
+ "grad_norm": 0.4451245069503784,
58
+ "learning_rate": 0.0005,
59
+ "loss": 1.6717,
60
+ "step": 500
61
+ },
62
+ {
63
+ "epoch": 1.7569546120058566,
64
+ "grad_norm": 0.49089929461479187,
65
+ "learning_rate": 0.0005,
66
+ "loss": 1.7006,
67
+ "step": 600
68
+ },
69
+ {
70
+ "epoch": 2.0,
71
+ "eval_accuracy": 0.43338461538461537,
72
+ "eval_loss": 4.586850643157959,
73
+ "eval_runtime": 5.9085,
74
+ "eval_samples_per_second": 84.624,
75
+ "eval_steps_per_second": 10.663,
76
+ "step": 683
77
+ },
78
+ {
79
+ "epoch": 2.0,
80
+ "eval_bleu": 0.2955390058097942,
81
+ "eval_exact_match": 0.0,
82
+ "eval_prefix_exact_match": 0.036,
83
+ "step": 683
84
+ },
85
+ {
86
+ "epoch": 2.049780380673499,
87
+ "grad_norm": 0.6150477528572083,
88
+ "learning_rate": 0.0005,
89
+ "loss": 1.6383,
90
+ "step": 700
91
+ },
92
+ {
93
+ "epoch": 2.342606149341142,
94
+ "grad_norm": 0.5455761551856995,
95
+ "learning_rate": 0.0005,
96
+ "loss": 1.2533,
97
+ "step": 800
98
+ },
99
+ {
100
+ "epoch": 2.6354319180087846,
101
+ "grad_norm": 0.6252492666244507,
102
+ "learning_rate": 0.0005,
103
+ "loss": 1.3256,
104
+ "step": 900
105
+ },
106
+ {
107
+ "epoch": 2.9282576866764276,
108
+ "grad_norm": 0.6721358299255371,
109
+ "learning_rate": 0.0005,
110
+ "loss": 1.3696,
111
+ "step": 1000
112
+ },
113
+ {
114
+ "epoch": 2.998535871156662,
115
+ "eval_accuracy": 0.42538461538461536,
116
+ "eval_loss": 4.720473766326904,
117
+ "eval_runtime": 5.6052,
118
+ "eval_samples_per_second": 89.204,
119
+ "eval_steps_per_second": 11.24,
120
+ "step": 1024
121
+ },
122
+ {
123
+ "epoch": 2.998535871156662,
124
+ "eval_bleu": 0.22572800415324107,
125
+ "eval_exact_match": 0.004,
126
+ "eval_prefix_exact_match": 0.044,
127
+ "step": 1024
128
+ },
129
+ {
130
+ "epoch": 3.22108345534407,
131
+ "grad_norm": 0.7449648380279541,
132
+ "learning_rate": 0.0005,
133
+ "loss": 1.0137,
134
+ "step": 1100
135
+ },
136
+ {
137
+ "epoch": 3.513909224011713,
138
+ "grad_norm": 0.7608182430267334,
139
+ "learning_rate": 0.0005,
140
+ "loss": 0.9535,
141
+ "step": 1200
142
+ },
143
+ {
144
+ "epoch": 3.8067349926793557,
145
+ "grad_norm": 0.8320977687835693,
146
+ "learning_rate": 0.0005,
147
+ "loss": 1.0297,
148
+ "step": 1300
149
+ },
150
+ {
151
+ "epoch": 4.0,
152
+ "eval_accuracy": 0.41717948717948716,
153
+ "eval_loss": 4.783276557922363,
154
+ "eval_runtime": 5.4398,
155
+ "eval_samples_per_second": 91.915,
156
+ "eval_steps_per_second": 11.581,
157
+ "step": 1366
158
+ },
159
+ {
160
+ "epoch": 4.0,
161
+ "eval_bleu": 0.28487206250718533,
162
+ "eval_exact_match": 0.008,
163
+ "eval_prefix_exact_match": 0.034,
164
+ "step": 1366
165
+ },
166
+ {
167
+ "epoch": 4.099560761346998,
168
+ "grad_norm": 0.7788870334625244,
169
+ "learning_rate": 0.0005,
170
+ "loss": 0.9222,
171
+ "step": 1400
172
+ },
173
+ {
174
+ "epoch": 4.392386530014641,
175
+ "grad_norm": 0.7981182336807251,
176
+ "learning_rate": 0.0005,
177
+ "loss": 0.6733,
178
+ "step": 1500
179
+ },
180
+ {
181
+ "epoch": 4.685212298682284,
182
+ "grad_norm": 0.938538670539856,
183
+ "learning_rate": 0.0005,
184
+ "loss": 0.7366,
185
+ "step": 1600
186
+ },
187
+ {
188
+ "epoch": 4.978038067349927,
189
+ "grad_norm": 0.9696006178855896,
190
+ "learning_rate": 0.0005,
191
+ "loss": 0.7991,
192
+ "step": 1700
193
+ },
194
+ {
195
+ "epoch": 4.998535871156662,
196
+ "eval_accuracy": 0.4169230769230769,
197
+ "eval_loss": 4.9334001541137695,
198
+ "eval_runtime": 5.1107,
199
+ "eval_samples_per_second": 97.835,
200
+ "eval_steps_per_second": 12.327,
201
+ "step": 1707
202
+ },
203
+ {
204
+ "epoch": 4.998535871156662,
205
+ "eval_bleu": 0.2463806011037282,
206
+ "eval_exact_match": 0.006,
207
+ "eval_prefix_exact_match": 0.034,
208
+ "step": 1707
209
+ },
210
+ {
211
+ "epoch": 5.270863836017569,
212
+ "grad_norm": 0.8532861471176147,
213
+ "learning_rate": 0.0005,
214
+ "loss": 0.5011,
215
+ "step": 1800
216
+ },
217
+ {
218
+ "epoch": 5.563689604685212,
219
+ "grad_norm": 0.8303696513175964,
220
+ "learning_rate": 0.0005,
221
+ "loss": 0.5186,
222
+ "step": 1900
223
+ },
224
+ {
225
+ "epoch": 5.856515373352855,
226
+ "grad_norm": 1.0248757600784302,
227
+ "learning_rate": 0.0005,
228
+ "loss": 0.5801,
229
+ "step": 2000
230
+ },
231
+ {
232
+ "epoch": 6.0,
233
+ "eval_accuracy": 0.4133846153846154,
234
+ "eval_loss": 5.196371078491211,
235
+ "eval_runtime": 5.382,
236
+ "eval_samples_per_second": 92.902,
237
+ "eval_steps_per_second": 11.706,
238
+ "step": 2049
239
+ },
240
+ {
241
+ "epoch": 6.0,
242
+ "eval_bleu": 0.24094851152622576,
243
+ "eval_exact_match": 0.008,
244
+ "eval_prefix_exact_match": 0.024,
245
+ "step": 2049
246
+ },
247
+ {
248
+ "epoch": 6.149341142020498,
249
+ "grad_norm": 0.8996496200561523,
250
+ "learning_rate": 0.0005,
251
+ "loss": 0.4772,
252
+ "step": 2100
253
+ },
254
+ {
255
+ "epoch": 6.44216691068814,
256
+ "grad_norm": 0.9010413885116577,
257
+ "learning_rate": 0.0005,
258
+ "loss": 0.377,
259
+ "step": 2200
260
+ },
261
+ {
262
+ "epoch": 6.734992679355783,
263
+ "grad_norm": 0.9476897716522217,
264
+ "learning_rate": 0.0005,
265
+ "loss": 0.4228,
266
+ "step": 2300
267
+ },
268
+ {
269
+ "epoch": 6.998535871156662,
270
+ "eval_accuracy": 0.4101025641025641,
271
+ "eval_loss": 5.403963565826416,
272
+ "eval_runtime": 5.2963,
273
+ "eval_samples_per_second": 94.406,
274
+ "eval_steps_per_second": 11.895,
275
+ "step": 2390
276
+ },
277
+ {
278
+ "epoch": 6.998535871156662,
279
+ "eval_bleu": 0.20794741740418707,
280
+ "eval_exact_match": 0.0,
281
+ "eval_prefix_exact_match": 0.024,
282
+ "step": 2390
283
+ },
284
+ {
285
+ "epoch": 7.027818448023426,
286
+ "grad_norm": 0.912433922290802,
287
+ "learning_rate": 0.0005,
288
+ "loss": 0.4456,
289
+ "step": 2400
290
+ },
291
+ {
292
+ "epoch": 7.320644216691068,
293
+ "grad_norm": 0.8055697679519653,
294
+ "learning_rate": 0.0005,
295
+ "loss": 0.2907,
296
+ "step": 2500
297
+ },
298
+ {
299
+ "epoch": 7.613469985358711,
300
+ "grad_norm": 0.9217767119407654,
301
+ "learning_rate": 0.0005,
302
+ "loss": 0.3285,
303
+ "step": 2600
304
+ },
305
+ {
306
+ "epoch": 7.906295754026354,
307
+ "grad_norm": 0.9473926424980164,
308
+ "learning_rate": 0.0005,
309
+ "loss": 0.3691,
310
+ "step": 2700
311
+ },
312
+ {
313
+ "epoch": 8.0,
314
+ "eval_accuracy": 0.40984615384615386,
315
+ "eval_loss": 5.655345439910889,
316
+ "eval_runtime": 5.2073,
317
+ "eval_samples_per_second": 96.019,
318
+ "eval_steps_per_second": 12.098,
319
+ "step": 2732
320
+ },
321
+ {
322
+ "epoch": 8.0,
323
+ "eval_bleu": 0.1614197770036361,
324
+ "eval_exact_match": 0.0,
325
+ "eval_prefix_exact_match": 0.028,
326
+ "step": 2732
327
+ },
328
+ {
329
+ "epoch": 8.199121522693996,
330
+ "grad_norm": 0.7216379642486572,
331
+ "learning_rate": 0.0005,
332
+ "loss": 0.296,
333
+ "step": 2800
334
+ },
335
+ {
336
+ "epoch": 8.49194729136164,
337
+ "grad_norm": 0.8579537868499756,
338
+ "learning_rate": 0.0005,
339
+ "loss": 0.2711,
340
+ "step": 2900
341
+ },
342
+ {
343
+ "epoch": 8.784773060029282,
344
+ "grad_norm": 0.8787992000579834,
345
+ "learning_rate": 0.0005,
346
+ "loss": 0.3052,
347
+ "step": 3000
348
+ },
349
+ {
350
+ "epoch": 8.998535871156662,
351
+ "eval_accuracy": 0.41215384615384615,
352
+ "eval_loss": 5.5592756271362305,
353
+ "eval_runtime": 5.4045,
354
+ "eval_samples_per_second": 92.515,
355
+ "eval_steps_per_second": 11.657,
356
+ "step": 3073
357
+ },
358
+ {
359
+ "epoch": 8.998535871156662,
360
+ "eval_bleu": 0.24598687767703112,
361
+ "eval_exact_match": 0.002,
362
+ "eval_prefix_exact_match": 0.018,
363
+ "step": 3073
364
+ },
365
+ {
366
+ "epoch": 9.077598828696924,
367
+ "grad_norm": 0.73407381772995,
368
+ "learning_rate": 0.0005,
369
+ "loss": 0.3065,
370
+ "step": 3100
371
+ },
372
+ {
373
+ "epoch": 9.370424597364568,
374
+ "grad_norm": 0.7941272854804993,
375
+ "learning_rate": 0.0005,
376
+ "loss": 0.2348,
377
+ "step": 3200
378
+ },
379
+ {
380
+ "epoch": 9.66325036603221,
381
+ "grad_norm": 0.8868365287780762,
382
+ "learning_rate": 0.0005,
383
+ "loss": 0.2645,
384
+ "step": 3300
385
+ },
386
+ {
387
+ "epoch": 9.956076134699854,
388
+ "grad_norm": 0.9728419184684753,
389
+ "learning_rate": 0.0005,
390
+ "loss": 0.2993,
391
+ "step": 3400
392
+ },
393
+ {
394
+ "epoch": 10.0,
395
+ "eval_accuracy": 0.4115384615384615,
396
+ "eval_loss": 5.659451484680176,
397
+ "eval_runtime": 5.4123,
398
+ "eval_samples_per_second": 92.383,
399
+ "eval_steps_per_second": 11.64,
400
+ "step": 3415
401
+ },
402
+ {
403
+ "epoch": 10.0,
404
+ "eval_bleu": 0.23027305956234542,
405
+ "eval_exact_match": 0.004,
406
+ "eval_prefix_exact_match": 0.024,
407
+ "step": 3415
408
+ },
409
+ {
410
+ "epoch": 10.248901903367496,
411
+ "grad_norm": 0.7543951869010925,
412
+ "learning_rate": 0.0005,
413
+ "loss": 0.2315,
414
+ "step": 3500
415
+ },
416
+ {
417
+ "epoch": 10.541727672035138,
418
+ "grad_norm": 0.824974536895752,
419
+ "learning_rate": 0.0005,
420
+ "loss": 0.2353,
421
+ "step": 3600
422
+ },
423
+ {
424
+ "epoch": 10.834553440702782,
425
+ "grad_norm": 0.8872408866882324,
426
+ "learning_rate": 0.0005,
427
+ "loss": 0.2639,
428
+ "step": 3700
429
+ },
430
+ {
431
+ "epoch": 10.998535871156662,
432
+ "eval_accuracy": 0.4052820512820513,
433
+ "eval_loss": 5.8112382888793945,
434
+ "eval_runtime": 5.3466,
435
+ "eval_samples_per_second": 93.518,
436
+ "eval_steps_per_second": 11.783,
437
+ "step": 3756
438
+ },
439
+ {
440
+ "epoch": 10.998535871156662,
441
+ "eval_bleu": 0.16709205929672397,
442
+ "eval_exact_match": 0.0,
443
+ "eval_prefix_exact_match": 0.024,
444
+ "step": 3756
445
+ },
446
+ {
447
+ "epoch": 11.127379209370424,
448
+ "grad_norm": 0.853175163269043,
449
+ "learning_rate": 0.0005,
450
+ "loss": 0.2524,
451
+ "step": 3800
452
+ },
453
+ {
454
+ "epoch": 11.420204978038067,
455
+ "grad_norm": 0.7671946883201599,
456
+ "learning_rate": 0.0005,
457
+ "loss": 0.2154,
458
+ "step": 3900
459
+ },
460
+ {
461
+ "epoch": 11.71303074670571,
462
+ "grad_norm": 0.945813775062561,
463
+ "learning_rate": 0.0005,
464
+ "loss": 0.2447,
465
+ "step": 4000
466
+ },
467
+ {
468
+ "epoch": 12.0,
469
+ "eval_accuracy": 0.40625641025641024,
470
+ "eval_loss": 5.811611175537109,
471
+ "eval_runtime": 5.2336,
472
+ "eval_samples_per_second": 95.536,
473
+ "eval_steps_per_second": 12.038,
474
+ "step": 4098
475
+ },
476
+ {
477
+ "epoch": 12.0,
478
+ "eval_bleu": 0.23356575996605206,
479
+ "eval_exact_match": 0.002,
480
+ "eval_prefix_exact_match": 0.028,
481
+ "step": 4098
482
+ },
483
+ {
484
+ "epoch": 12.005856515373353,
485
+ "grad_norm": 0.6562711000442505,
486
+ "learning_rate": 0.0005,
487
+ "loss": 0.2727,
488
+ "step": 4100
489
+ },
490
+ {
491
+ "epoch": 12.298682284040996,
492
+ "grad_norm": 0.8681979179382324,
493
+ "learning_rate": 0.0005,
494
+ "loss": 0.2047,
495
+ "step": 4200
496
+ },
497
+ {
498
+ "epoch": 12.591508052708638,
499
+ "grad_norm": 0.9272928237915039,
500
+ "learning_rate": 0.0005,
501
+ "loss": 0.2236,
502
+ "step": 4300
503
+ },
504
+ {
505
+ "epoch": 12.88433382137628,
506
+ "grad_norm": 0.9103503823280334,
507
+ "learning_rate": 0.0005,
508
+ "loss": 0.257,
509
+ "step": 4400
510
+ },
511
+ {
512
+ "epoch": 12.998535871156662,
513
+ "eval_accuracy": 0.4067179487179487,
514
+ "eval_loss": 5.797003269195557,
515
+ "eval_runtime": 5.1385,
516
+ "eval_samples_per_second": 97.304,
517
+ "eval_steps_per_second": 12.26,
518
+ "step": 4439
519
+ },
520
+ {
521
+ "epoch": 12.998535871156662,
522
+ "eval_bleu": 0.23987536271361076,
523
+ "eval_exact_match": 0.0,
524
+ "eval_prefix_exact_match": 0.018,
525
+ "step": 4439
526
+ },
527
+ {
528
+ "epoch": 13.177159590043924,
529
+ "grad_norm": 0.7556403875350952,
530
+ "learning_rate": 0.0005,
531
+ "loss": 0.2279,
532
+ "step": 4500
533
+ },
534
+ {
535
+ "epoch": 13.469985358711567,
536
+ "grad_norm": 0.8469985723495483,
537
+ "learning_rate": 0.0005,
538
+ "loss": 0.2122,
539
+ "step": 4600
540
+ },
541
+ {
542
+ "epoch": 13.762811127379209,
543
+ "grad_norm": 0.8414324522018433,
544
+ "learning_rate": 0.0005,
545
+ "loss": 0.2346,
546
+ "step": 4700
547
+ },
548
+ {
549
+ "epoch": 14.0,
550
+ "eval_accuracy": 0.40102564102564103,
551
+ "eval_loss": 5.798365592956543,
552
+ "eval_runtime": 5.1157,
553
+ "eval_samples_per_second": 97.738,
554
+ "eval_steps_per_second": 12.315,
555
+ "step": 4781
556
+ },
557
+ {
558
+ "epoch": 14.0,
559
+ "eval_bleu": 0.22627859689716945,
560
+ "eval_exact_match": 0.004,
561
+ "eval_prefix_exact_match": 0.034,
562
+ "step": 4781
563
+ },
564
+ {
565
+ "epoch": 14.055636896046853,
566
+ "grad_norm": 0.758216917514801,
567
+ "learning_rate": 0.0005,
568
+ "loss": 0.2479,
569
+ "step": 4800
570
+ },
571
+ {
572
+ "epoch": 14.348462664714495,
573
+ "grad_norm": 0.7591387629508972,
574
+ "learning_rate": 0.0005,
575
+ "loss": 0.1986,
576
+ "step": 4900
577
+ },
578
+ {
579
+ "epoch": 14.641288433382137,
580
+ "grad_norm": 0.8450336456298828,
581
+ "learning_rate": 0.0005,
582
+ "loss": 0.2208,
583
+ "step": 5000
584
+ },
585
+ {
586
+ "epoch": 14.93411420204978,
587
+ "grad_norm": 0.9157751202583313,
588
+ "learning_rate": 0.0005,
589
+ "loss": 0.2477,
590
+ "step": 5100
591
+ },
592
+ {
593
+ "epoch": 14.998535871156662,
594
+ "eval_accuracy": 0.4007179487179487,
595
+ "eval_loss": 5.948267936706543,
596
+ "eval_runtime": 5.2431,
597
+ "eval_samples_per_second": 95.363,
598
+ "eval_steps_per_second": 12.016,
599
+ "step": 5122
600
+ },
601
+ {
602
+ "epoch": 14.998535871156662,
603
+ "eval_bleu": 0.20358408351652205,
604
+ "eval_exact_match": 0.002,
605
+ "eval_prefix_exact_match": 0.03,
606
+ "step": 5122
607
+ },
608
+ {
609
+ "epoch": 15.226939970717423,
610
+ "grad_norm": 0.769048273563385,
611
+ "learning_rate": 0.0005,
612
+ "loss": 0.2087,
613
+ "step": 5200
614
+ },
615
+ {
616
+ "epoch": 15.519765739385067,
617
+ "grad_norm": 0.8749043345451355,
618
+ "learning_rate": 0.0005,
619
+ "loss": 0.2099,
620
+ "step": 5300
621
+ },
622
+ {
623
+ "epoch": 15.812591508052709,
624
+ "grad_norm": 1.0207552909851074,
625
+ "learning_rate": 0.0005,
626
+ "loss": 0.2341,
627
+ "step": 5400
628
+ },
629
+ {
630
+ "epoch": 16.0,
631
+ "eval_accuracy": 0.3996923076923077,
632
+ "eval_loss": 6.083972930908203,
633
+ "eval_runtime": 5.137,
634
+ "eval_samples_per_second": 97.333,
635
+ "eval_steps_per_second": 12.264,
636
+ "step": 5464
637
+ },
638
+ {
639
+ "epoch": 16.0,
640
+ "eval_bleu": 0.23385557447439007,
641
+ "eval_exact_match": 0.0,
642
+ "eval_prefix_exact_match": 0.022,
643
+ "step": 5464
644
+ },
645
+ {
646
+ "epoch": 16.105417276720353,
647
+ "grad_norm": 0.7007560729980469,
648
+ "learning_rate": 0.0005,
649
+ "loss": 0.2333,
650
+ "step": 5500
651
+ },
652
+ {
653
+ "epoch": 16.398243045387993,
654
+ "grad_norm": 0.7863999605178833,
655
+ "learning_rate": 0.0005,
656
+ "loss": 0.1969,
657
+ "step": 5600
658
+ },
659
+ {
660
+ "epoch": 16.691068814055637,
661
+ "grad_norm": 0.8659201860427856,
662
+ "learning_rate": 0.0005,
663
+ "loss": 0.2206,
664
+ "step": 5700
665
+ },
666
+ {
667
+ "epoch": 16.98389458272328,
668
+ "grad_norm": 1.0170942544937134,
669
+ "learning_rate": 0.0005,
670
+ "loss": 0.2471,
671
+ "step": 5800
672
+ },
673
+ {
674
+ "epoch": 16.998535871156662,
675
+ "eval_accuracy": 0.3975897435897436,
676
+ "eval_loss": 6.0254926681518555,
677
+ "eval_runtime": 5.1849,
678
+ "eval_samples_per_second": 96.434,
679
+ "eval_steps_per_second": 12.151,
680
+ "step": 5805
681
+ },
682
+ {
683
+ "epoch": 16.998535871156662,
684
+ "eval_bleu": 0.21616749091660492,
685
+ "eval_exact_match": 0.0,
686
+ "eval_prefix_exact_match": 0.028,
687
+ "step": 5805
688
+ },
689
+ {
690
+ "epoch": 17.27672035139092,
691
+ "grad_norm": 0.7882862687110901,
692
+ "learning_rate": 0.0005,
693
+ "loss": 0.1923,
694
+ "step": 5900
695
+ },
696
+ {
697
+ "epoch": 17.569546120058565,
698
+ "grad_norm": 0.8134691715240479,
699
+ "learning_rate": 0.0005,
700
+ "loss": 0.2085,
701
+ "step": 6000
702
+ },
703
+ {
704
+ "epoch": 17.86237188872621,
705
+ "grad_norm": 0.8864043354988098,
706
+ "learning_rate": 0.0005,
707
+ "loss": 0.2303,
708
+ "step": 6100
709
+ },
710
+ {
711
+ "epoch": 18.0,
712
+ "eval_accuracy": 0.4012307692307692,
713
+ "eval_loss": 5.9475321769714355,
714
+ "eval_runtime": 5.234,
715
+ "eval_samples_per_second": 95.529,
716
+ "eval_steps_per_second": 12.037,
717
+ "step": 6147
718
+ },
719
+ {
720
+ "epoch": 18.0,
721
+ "eval_bleu": 0.20913368282784864,
722
+ "eval_exact_match": 0.0,
723
+ "eval_prefix_exact_match": 0.032,
724
+ "step": 6147
725
+ },
726
+ {
727
+ "epoch": 18.15519765739385,
728
+ "grad_norm": 0.8333725333213806,
729
+ "learning_rate": 0.0005,
730
+ "loss": 0.2133,
731
+ "step": 6200
732
+ },
733
+ {
734
+ "epoch": 18.448023426061493,
735
+ "grad_norm": 0.7502799034118652,
736
+ "learning_rate": 0.0005,
737
+ "loss": 0.1958,
738
+ "step": 6300
739
+ },
740
+ {
741
+ "epoch": 18.740849194729137,
742
+ "grad_norm": 0.8406829237937927,
743
+ "learning_rate": 0.0005,
744
+ "loss": 0.2165,
745
+ "step": 6400
746
+ },
747
+ {
748
+ "epoch": 18.998535871156662,
749
+ "eval_accuracy": 0.396,
750
+ "eval_loss": 6.3112921714782715,
751
+ "eval_runtime": 5.9047,
752
+ "eval_samples_per_second": 84.678,
753
+ "eval_steps_per_second": 10.669,
754
+ "step": 6488
755
+ },
756
+ {
757
+ "epoch": 18.998535871156662,
758
+ "eval_bleu": 0.1825360355357865,
759
+ "eval_exact_match": 0.002,
760
+ "eval_prefix_exact_match": 0.02,
761
+ "step": 6488
762
+ },
763
+ {
764
+ "epoch": 19.03367496339678,
765
+ "grad_norm": 0.716468095779419,
766
+ "learning_rate": 0.0005,
767
+ "loss": 0.2323,
768
+ "step": 6500
769
+ },
770
+ {
771
+ "epoch": 19.32650073206442,
772
+ "grad_norm": 0.7107975482940674,
773
+ "learning_rate": 0.0005,
774
+ "loss": 0.1824,
775
+ "step": 6600
776
+ },
777
+ {
778
+ "epoch": 19.619326500732065,
779
+ "grad_norm": 0.7902608513832092,
780
+ "learning_rate": 0.0005,
781
+ "loss": 0.2039,
782
+ "step": 6700
783
+ },
784
+ {
785
+ "epoch": 19.91215226939971,
786
+ "grad_norm": 0.9088649749755859,
787
+ "learning_rate": 0.0005,
788
+ "loss": 0.2293,
789
+ "step": 6800
790
+ },
791
+ {
792
+ "epoch": 20.0,
793
+ "eval_accuracy": 0.39353846153846156,
794
+ "eval_loss": 6.162784099578857,
795
+ "eval_runtime": 5.2136,
796
+ "eval_samples_per_second": 95.902,
797
+ "eval_steps_per_second": 12.084,
798
+ "step": 6830
799
+ },
800
+ {
801
+ "epoch": 20.0,
802
+ "eval_bleu": 0.2067905686266933,
803
+ "eval_exact_match": 0.002,
804
+ "eval_prefix_exact_match": 0.028,
805
+ "step": 6830
806
+ },
807
+ {
808
+ "epoch": 20.20497803806735,
809
+ "grad_norm": 0.6533398628234863,
810
+ "learning_rate": 0.0005,
811
+ "loss": 0.1999,
812
+ "step": 6900
813
+ },
814
+ {
815
+ "epoch": 20.497803806734993,
816
+ "grad_norm": 0.9218800663948059,
817
+ "learning_rate": 0.0005,
818
+ "loss": 0.1961,
819
+ "step": 7000
820
+ },
821
+ {
822
+ "epoch": 20.790629575402637,
823
+ "grad_norm": 0.9343099594116211,
824
+ "learning_rate": 0.0005,
825
+ "loss": 0.2193,
826
+ "step": 7100
827
+ },
828
+ {
829
+ "epoch": 20.998535871156662,
830
+ "eval_accuracy": 0.3883076923076923,
831
+ "eval_loss": 6.313322067260742,
832
+ "eval_runtime": 5.9425,
833
+ "eval_samples_per_second": 84.139,
834
+ "eval_steps_per_second": 10.602,
835
+ "step": 7171
836
+ },
837
+ {
838
+ "epoch": 20.998535871156662,
839
+ "eval_bleu": 0.21669080510396269,
840
+ "eval_exact_match": 0.0,
841
+ "eval_prefix_exact_match": 0.028,
842
+ "step": 7171
843
+ },
844
+ {
845
+ "epoch": 21.083455344070277,
846
+ "grad_norm": 0.5941635966300964,
847
+ "learning_rate": 0.0005,
848
+ "loss": 0.2221,
849
+ "step": 7200
850
+ },
851
+ {
852
+ "epoch": 21.37628111273792,
853
+ "grad_norm": 0.8915172815322876,
854
+ "learning_rate": 0.0005,
855
+ "loss": 0.1894,
856
+ "step": 7300
857
+ },
858
+ {
859
+ "epoch": 21.669106881405565,
860
+ "grad_norm": 0.7834005951881409,
861
+ "learning_rate": 0.0005,
862
+ "loss": 0.2072,
863
+ "step": 7400
864
+ },
865
+ {
866
+ "epoch": 21.961932650073205,
867
+ "grad_norm": 0.9307728409767151,
868
+ "learning_rate": 0.0005,
869
+ "loss": 0.2313,
870
+ "step": 7500
871
+ },
872
+ {
873
+ "epoch": 22.0,
874
+ "eval_accuracy": 0.3914871794871795,
875
+ "eval_loss": 6.137081146240234,
876
+ "eval_runtime": 5.8625,
877
+ "eval_samples_per_second": 85.288,
878
+ "eval_steps_per_second": 10.746,
879
+ "step": 7513
880
+ },
881
+ {
882
+ "epoch": 22.0,
883
+ "eval_bleu": 0.21802779706463218,
884
+ "eval_exact_match": 0.002,
885
+ "eval_prefix_exact_match": 0.034,
886
+ "step": 7513
887
+ },
888
+ {
889
+ "epoch": 22.25475841874085,
890
+ "grad_norm": 0.6620280146598816,
891
+ "learning_rate": 0.0005,
892
+ "loss": 0.189,
893
+ "step": 7600
894
+ },
895
+ {
896
+ "epoch": 22.547584187408493,
897
+ "grad_norm": 0.8460494875907898,
898
+ "learning_rate": 0.0005,
899
+ "loss": 0.196,
900
+ "step": 7700
901
+ },
902
+ {
903
+ "epoch": 22.840409956076133,
904
+ "grad_norm": 1.0066401958465576,
905
+ "learning_rate": 0.0005,
906
+ "loss": 0.217,
907
+ "step": 7800
908
+ },
909
+ {
910
+ "epoch": 22.998535871156662,
911
+ "eval_accuracy": 0.3931794871794872,
912
+ "eval_loss": 6.032333850860596,
913
+ "eval_runtime": 5.8719,
914
+ "eval_samples_per_second": 85.151,
915
+ "eval_steps_per_second": 10.729,
916
+ "step": 7854
917
+ },
918
+ {
919
+ "epoch": 22.998535871156662,
920
+ "eval_bleu": 0.2132135919503725,
921
+ "eval_exact_match": 0.0,
922
+ "eval_prefix_exact_match": 0.012,
923
+ "step": 7854
924
+ },
925
+ {
926
+ "epoch": 23.133235724743777,
927
+ "grad_norm": 0.6808323860168457,
928
+ "learning_rate": 0.0005,
929
+ "loss": 0.2105,
930
+ "step": 7900
931
+ },
932
+ {
933
+ "epoch": 23.42606149341142,
934
+ "grad_norm": 0.8528187274932861,
935
+ "learning_rate": 0.0005,
936
+ "loss": 0.1866,
937
+ "step": 8000
938
+ },
939
+ {
940
+ "epoch": 23.718887262079065,
941
+ "grad_norm": 0.8722848296165466,
942
+ "learning_rate": 0.0005,
943
+ "loss": 0.205,
944
+ "step": 8100
945
+ },
946
+ {
947
+ "epoch": 24.0,
948
+ "eval_accuracy": 0.3891794871794872,
949
+ "eval_loss": 6.203835964202881,
950
+ "eval_runtime": 5.243,
951
+ "eval_samples_per_second": 95.365,
952
+ "eval_steps_per_second": 12.016,
953
+ "step": 8196
954
+ },
955
+ {
956
+ "epoch": 24.0,
957
+ "eval_bleu": 0.19430601075864046,
958
+ "eval_exact_match": 0.0,
959
+ "eval_prefix_exact_match": 0.014,
960
+ "step": 8196
961
+ }
962
+ ],
963
+ "logging_steps": 100,
964
+ "max_steps": 17050,
965
+ "num_input_tokens_seen": 0,
966
+ "num_train_epochs": 50,
967
+ "save_steps": 500,
968
+ "stateful_callbacks": {
969
+ "TrainerControl": {
970
+ "args": {
971
+ "should_epoch_stop": false,
972
+ "should_evaluate": false,
973
+ "should_log": false,
974
+ "should_save": true,
975
+ "should_training_stop": false
976
+ },
977
+ "attributes": {}
978
+ }
979
+ },
980
+ "total_flos": 7.038040180103578e+17,
981
+ "train_batch_size": 1,
982
+ "trial_name": null,
983
+ "trial_params": null
984
+ }
checkpoint-8196/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e09109d7fc8f49ffd654f0e1286ef00d8840620753fec453a60a103c55187142
3
+ size 5304
checkpoint-8196/vocab.json ADDED
The diff for this file is too large to render. See raw diff