Hugging Face's logo

bishaltwr
/
final_m2m100

Safetensors
m2m_100
Model card Files Community
bishaltwr commited on
Commit
a8bf27e
•
1 Parent(s): c0af94e

Upload 13 files

Browse files
Files changed (13) hide show
  1. added_tokens.json +102 -0
  2. config.json +37 -0
  3. generation_config.json +11 -0
  4. model.safetensors +3 -0
  5. optimizer.pt +3 -0
  6. rng_state.pth +3 -0
  7. scheduler.pt +3 -0
  8. sentencepiece.bpe.model +3 -0
  9. special_tokens_map.json +109 -0
  10. tokenizer_config.json +951 -0
  11. trainer_state.json +2993 -0
  12. training_args.bin +3 -0
  13. vocab.json +0 -0
added_tokens.json ADDED
@@ -0,0 +1,102 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "__af__": 128004,
3
+ "__am__": 128005,
4
+ "__ar__": 128006,
5
+ "__ast__": 128007,
6
+ "__az__": 128008,
7
+ "__ba__": 128009,
8
+ "__be__": 128010,
9
+ "__bg__": 128011,
10
+ "__bn__": 128012,
11
+ "__br__": 128013,
12
+ "__bs__": 128014,
13
+ "__ca__": 128015,
14
+ "__ceb__": 128016,
15
+ "__cs__": 128017,
16
+ "__cy__": 128018,
17
+ "__da__": 128019,
18
+ "__de__": 128020,
19
+ "__el__": 128021,
20
+ "__en__": 128022,
21
+ "__es__": 128023,
22
+ "__et__": 128024,
23
+ "__fa__": 128025,
24
+ "__ff__": 128026,
25
+ "__fi__": 128027,
26
+ "__fr__": 128028,
27
+ "__fy__": 128029,
28
+ "__ga__": 128030,
29
+ "__gd__": 128031,
30
+ "__gl__": 128032,
31
+ "__gu__": 128033,
32
+ "__ha__": 128034,
33
+ "__he__": 128035,
34
+ "__hi__": 128036,
35
+ "__hr__": 128037,
36
+ "__ht__": 128038,
37
+ "__hu__": 128039,
38
+ "__hy__": 128040,
39
+ "__id__": 128041,
40
+ "__ig__": 128042,
41
+ "__ilo__": 128043,
42
+ "__is__": 128044,
43
+ "__it__": 128045,
44
+ "__ja__": 128046,
45
+ "__jv__": 128047,
46
+ "__ka__": 128048,
47
+ "__kk__": 128049,
48
+ "__km__": 128050,
49
+ "__kn__": 128051,
50
+ "__ko__": 128052,
51
+ "__lb__": 128053,
52
+ "__lg__": 128054,
53
+ "__ln__": 128055,
54
+ "__lo__": 128056,
55
+ "__lt__": 128057,
56
+ "__lv__": 128058,
57
+ "__mg__": 128059,
58
+ "__mk__": 128060,
59
+ "__ml__": 128061,
60
+ "__mn__": 128062,
61
+ "__mr__": 128063,
62
+ "__ms__": 128064,
63
+ "__my__": 128065,
64
+ "__ne__": 128066,
65
+ "__nl__": 128067,
66
+ "__no__": 128068,
67
+ "__ns__": 128069,
68
+ "__oc__": 128070,
69
+ "__or__": 128071,
70
+ "__pa__": 128072,
71
+ "__pl__": 128073,
72
+ "__ps__": 128074,
73
+ "__pt__": 128075,
74
+ "__ro__": 128076,
75
+ "__ru__": 128077,
76
+ "__sd__": 128078,
77
+ "__si__": 128079,
78
+ "__sk__": 128080,
79
+ "__sl__": 128081,
80
+ "__so__": 128082,
81
+ "__sq__": 128083,
82
+ "__sr__": 128084,
83
+ "__ss__": 128085,
84
+ "__su__": 128086,
85
+ "__sv__": 128087,
86
+ "__sw__": 128088,
87
+ "__ta__": 128089,
88
+ "__th__": 128090,
89
+ "__tl__": 128091,
90
+ "__tn__": 128092,
91
+ "__tr__": 128093,
92
+ "__uk__": 128094,
93
+ "__ur__": 128095,
94
+ "__uz__": 128096,
95
+ "__vi__": 128097,
96
+ "__wo__": 128098,
97
+ "__xh__": 128099,
98
+ "__yi__": 128100,
99
+ "__yo__": 128101,
100
+ "__zh__": 128102,
101
+ "__zu__": 128103
102
+ }
config.json ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/m2m100_418M",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "relu",
5
+ "architectures": [
6
+ "M2M100ForConditionalGeneration"
7
+ ],
8
+ "attention_dropout": 0.1,
9
+ "bos_token_id": 0,
10
+ "d_model": 1024,
11
+ "decoder_attention_heads": 16,
12
+ "decoder_ffn_dim": 4096,
13
+ "decoder_layerdrop": 0.05,
14
+ "decoder_layers": 12,
15
+ "decoder_start_token_id": 2,
16
+ "dropout": 0.1,
17
+ "early_stopping": true,
18
+ "encoder_attention_heads": 16,
19
+ "encoder_ffn_dim": 4096,
20
+ "encoder_layerdrop": 0.05,
21
+ "encoder_layers": 12,
22
+ "eos_token_id": 2,
23
+ "gradient_checkpointing": false,
24
+ "init_std": 0.02,
25
+ "is_encoder_decoder": true,
26
+ "max_length": 200,
27
+ "max_position_embeddings": 1024,
28
+ "model_type": "m2m_100",
29
+ "num_beams": 5,
30
+ "num_hidden_layers": 12,
31
+ "pad_token_id": 1,
32
+ "scale_embedding": true,
33
+ "torch_dtype": "float32",
34
+ "transformers_version": "4.45.0.dev0",
35
+ "use_cache": true,
36
+ "vocab_size": 128112
37
+ }
generation_config.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 0,
4
+ "decoder_start_token_id": 2,
5
+ "early_stopping": true,
6
+ "eos_token_id": 2,
7
+ "max_length": 200,
8
+ "num_beams": 5,
9
+ "pad_token_id": 1,
10
+ "transformers_version": "4.45.0.dev0"
11
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:969c6b05e7f300c0d4675d0840de6660f2f9ff3e868caf697625946fd6500e69
3
+ size 1935681888
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cfe933b9cc968f55e094f1f52177627479b420d6847ea543bc666a93a2b9bd83
3
+ size 3871670921
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:06df9c86539ea91466ad77c0fc538d3171651954fb38abbe5e5fe131fc42b3e8
3
+ size 14244
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cec8709d312f77ca0b51474743f6f1f65a6d92476993dfb10b2ad88e4f74e69f
3
+ size 1064
sentencepiece.bpe.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d8f7c76ed2a5e0822be39f0a4f95a55eb19c78f4593ce609e2edbc2aea4d380a
3
+ size 2423393
special_tokens_map.json ADDED
@@ -0,0 +1,109 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "__af__",
4
+ "__am__",
5
+ "__ar__",
6
+ "__ast__",
7
+ "__az__",
8
+ "__ba__",
9
+ "__be__",
10
+ "__bg__",
11
+ "__bn__",
12
+ "__br__",
13
+ "__bs__",
14
+ "__ca__",
15
+ "__ceb__",
16
+ "__cs__",
17
+ "__cy__",
18
+ "__da__",
19
+ "__de__",
20
+ "__el__",
21
+ "__en__",
22
+ "__es__",
23
+ "__et__",
24
+ "__fa__",
25
+ "__ff__",
26
+ "__fi__",
27
+ "__fr__",
28
+ "__fy__",
29
+ "__ga__",
30
+ "__gd__",
31
+ "__gl__",
32
+ "__gu__",
33
+ "__ha__",
34
+ "__he__",
35
+ "__hi__",
36
+ "__hr__",
37
+ "__ht__",
38
+ "__hu__",
39
+ "__hy__",
40
+ "__id__",
41
+ "__ig__",
42
+ "__ilo__",
43
+ "__is__",
44
+ "__it__",
45
+ "__ja__",
46
+ "__jv__",
47
+ "__ka__",
48
+ "__kk__",
49
+ "__km__",
50
+ "__kn__",
51
+ "__ko__",
52
+ "__lb__",
53
+ "__lg__",
54
+ "__ln__",
55
+ "__lo__",
56
+ "__lt__",
57
+ "__lv__",
58
+ "__mg__",
59
+ "__mk__",
60
+ "__ml__",
61
+ "__mn__",
62
+ "__mr__",
63
+ "__ms__",
64
+ "__my__",
65
+ "__ne__",
66
+ "__nl__",
67
+ "__no__",
68
+ "__ns__",
69
+ "__oc__",
70
+ "__or__",
71
+ "__pa__",
72
+ "__pl__",
73
+ "__ps__",
74
+ "__pt__",
75
+ "__ro__",
76
+ "__ru__",
77
+ "__sd__",
78
+ "__si__",
79
+ "__sk__",
80
+ "__sl__",
81
+ "__so__",
82
+ "__sq__",
83
+ "__sr__",
84
+ "__ss__",
85
+ "__su__",
86
+ "__sv__",
87
+ "__sw__",
88
+ "__ta__",
89
+ "__th__",
90
+ "__tl__",
91
+ "__tn__",
92
+ "__tr__",
93
+ "__uk__",
94
+ "__ur__",
95
+ "__uz__",
96
+ "__vi__",
97
+ "__wo__",
98
+ "__xh__",
99
+ "__yi__",
100
+ "__yo__",
101
+ "__zh__",
102
+ "__zu__"
103
+ ],
104
+ "bos_token": "<s>",
105
+ "eos_token": "</s>",
106
+ "pad_token": "<pad>",
107
+ "sep_token": "</s>",
108
+ "unk_token": "<unk>"
109
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,951 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "</s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "128004": {
36
+ "content": "__af__",
37
+ "lstrip": false,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ },
43
+ "128005": {
44
+ "content": "__am__",
45
+ "lstrip": false,
46
+ "normalized": false,
47
+ "rstrip": false,
48
+ "single_word": false,
49
+ "special": true
50
+ },
51
+ "128006": {
52
+ "content": "__ar__",
53
+ "lstrip": false,
54
+ "normalized": false,
55
+ "rstrip": false,
56
+ "single_word": false,
57
+ "special": true
58
+ },
59
+ "128007": {
60
+ "content": "__ast__",
61
+ "lstrip": false,
62
+ "normalized": false,
63
+ "rstrip": false,
64
+ "single_word": false,
65
+ "special": true
66
+ },
67
+ "128008": {
68
+ "content": "__az__",
69
+ "lstrip": false,
70
+ "normalized": false,
71
+ "rstrip": false,
72
+ "single_word": false,
73
+ "special": true
74
+ },
75
+ "128009": {
76
+ "content": "__ba__",
77
+ "lstrip": false,
78
+ "normalized": false,
79
+ "rstrip": false,
80
+ "single_word": false,
81
+ "special": true
82
+ },
83
+ "128010": {
84
+ "content": "__be__",
85
+ "lstrip": false,
86
+ "normalized": false,
87
+ "rstrip": false,
88
+ "single_word": false,
89
+ "special": true
90
+ },
91
+ "128011": {
92
+ "content": "__bg__",
93
+ "lstrip": false,
94
+ "normalized": false,
95
+ "rstrip": false,
96
+ "single_word": false,
97
+ "special": true
98
+ },
99
+ "128012": {
100
+ "content": "__bn__",
101
+ "lstrip": false,
102
+ "normalized": false,
103
+ "rstrip": false,
104
+ "single_word": false,
105
+ "special": true
106
+ },
107
+ "128013": {
108
+ "content": "__br__",
109
+ "lstrip": false,
110
+ "normalized": false,
111
+ "rstrip": false,
112
+ "single_word": false,
113
+ "special": true
114
+ },
115
+ "128014": {
116
+ "content": "__bs__",
117
+ "lstrip": false,
118
+ "normalized": false,
119
+ "rstrip": false,
120
+ "single_word": false,
121
+ "special": true
122
+ },
123
+ "128015": {
124
+ "content": "__ca__",
125
+ "lstrip": false,
126
+ "normalized": false,
127
+ "rstrip": false,
128
+ "single_word": false,
129
+ "special": true
130
+ },
131
+ "128016": {
132
+ "content": "__ceb__",
133
+ "lstrip": false,
134
+ "normalized": false,
135
+ "rstrip": false,
136
+ "single_word": false,
137
+ "special": true
138
+ },
139
+ "128017": {
140
+ "content": "__cs__",
141
+ "lstrip": false,
142
+ "normalized": false,
143
+ "rstrip": false,
144
+ "single_word": false,
145
+ "special": true
146
+ },
147
+ "128018": {
148
+ "content": "__cy__",
149
+ "lstrip": false,
150
+ "normalized": false,
151
+ "rstrip": false,
152
+ "single_word": false,
153
+ "special": true
154
+ },
155
+ "128019": {
156
+ "content": "__da__",
157
+ "lstrip": false,
158
+ "normalized": false,
159
+ "rstrip": false,
160
+ "single_word": false,
161
+ "special": true
162
+ },
163
+ "128020": {
164
+ "content": "__de__",
165
+ "lstrip": false,
166
+ "normalized": false,
167
+ "rstrip": false,
168
+ "single_word": false,
169
+ "special": true
170
+ },
171
+ "128021": {
172
+ "content": "__el__",
173
+ "lstrip": false,
174
+ "normalized": false,
175
+ "rstrip": false,
176
+ "single_word": false,
177
+ "special": true
178
+ },
179
+ "128022": {
180
+ "content": "__en__",
181
+ "lstrip": false,
182
+ "normalized": false,
183
+ "rstrip": false,
184
+ "single_word": false,
185
+ "special": true
186
+ },
187
+ "128023": {
188
+ "content": "__es__",
189
+ "lstrip": false,
190
+ "normalized": false,
191
+ "rstrip": false,
192
+ "single_word": false,
193
+ "special": true
194
+ },
195
+ "128024": {
196
+ "content": "__et__",
197
+ "lstrip": false,
198
+ "normalized": false,
199
+ "rstrip": false,
200
+ "single_word": false,
201
+ "special": true
202
+ },
203
+ "128025": {
204
+ "content": "__fa__",
205
+ "lstrip": false,
206
+ "normalized": false,
207
+ "rstrip": false,
208
+ "single_word": false,
209
+ "special": true
210
+ },
211
+ "128026": {
212
+ "content": "__ff__",
213
+ "lstrip": false,
214
+ "normalized": false,
215
+ "rstrip": false,
216
+ "single_word": false,
217
+ "special": true
218
+ },
219
+ "128027": {
220
+ "content": "__fi__",
221
+ "lstrip": false,
222
+ "normalized": false,
223
+ "rstrip": false,
224
+ "single_word": false,
225
+ "special": true
226
+ },
227
+ "128028": {
228
+ "content": "__fr__",
229
+ "lstrip": false,
230
+ "normalized": false,
231
+ "rstrip": false,
232
+ "single_word": false,
233
+ "special": true
234
+ },
235
+ "128029": {
236
+ "content": "__fy__",
237
+ "lstrip": false,
238
+ "normalized": false,
239
+ "rstrip": false,
240
+ "single_word": false,
241
+ "special": true
242
+ },
243
+ "128030": {
244
+ "content": "__ga__",
245
+ "lstrip": false,
246
+ "normalized": false,
247
+ "rstrip": false,
248
+ "single_word": false,
249
+ "special": true
250
+ },
251
+ "128031": {
252
+ "content": "__gd__",
253
+ "lstrip": false,
254
+ "normalized": false,
255
+ "rstrip": false,
256
+ "single_word": false,
257
+ "special": true
258
+ },
259
+ "128032": {
260
+ "content": "__gl__",
261
+ "lstrip": false,
262
+ "normalized": false,
263
+ "rstrip": false,
264
+ "single_word": false,
265
+ "special": true
266
+ },
267
+ "128033": {
268
+ "content": "__gu__",
269
+ "lstrip": false,
270
+ "normalized": false,
271
+ "rstrip": false,
272
+ "single_word": false,
273
+ "special": true
274
+ },
275
+ "128034": {
276
+ "content": "__ha__",
277
+ "lstrip": false,
278
+ "normalized": false,
279
+ "rstrip": false,
280
+ "single_word": false,
281
+ "special": true
282
+ },
283
+ "128035": {
284
+ "content": "__he__",
285
+ "lstrip": false,
286
+ "normalized": false,
287
+ "rstrip": false,
288
+ "single_word": false,
289
+ "special": true
290
+ },
291
+ "128036": {
292
+ "content": "__hi__",
293
+ "lstrip": false,
294
+ "normalized": false,
295
+ "rstrip": false,
296
+ "single_word": false,
297
+ "special": true
298
+ },
299
+ "128037": {
300
+ "content": "__hr__",
301
+ "lstrip": false,
302
+ "normalized": false,
303
+ "rstrip": false,
304
+ "single_word": false,
305
+ "special": true
306
+ },
307
+ "128038": {
308
+ "content": "__ht__",
309
+ "lstrip": false,
310
+ "normalized": false,
311
+ "rstrip": false,
312
+ "single_word": false,
313
+ "special": true
314
+ },
315
+ "128039": {
316
+ "content": "__hu__",
317
+ "lstrip": false,
318
+ "normalized": false,
319
+ "rstrip": false,
320
+ "single_word": false,
321
+ "special": true
322
+ },
323
+ "128040": {
324
+ "content": "__hy__",
325
+ "lstrip": false,
326
+ "normalized": false,
327
+ "rstrip": false,
328
+ "single_word": false,
329
+ "special": true
330
+ },
331
+ "128041": {
332
+ "content": "__id__",
333
+ "lstrip": false,
334
+ "normalized": false,
335
+ "rstrip": false,
336
+ "single_word": false,
337
+ "special": true
338
+ },
339
+ "128042": {
340
+ "content": "__ig__",
341
+ "lstrip": false,
342
+ "normalized": false,
343
+ "rstrip": false,
344
+ "single_word": false,
345
+ "special": true
346
+ },
347
+ "128043": {
348
+ "content": "__ilo__",
349
+ "lstrip": false,
350
+ "normalized": false,
351
+ "rstrip": false,
352
+ "single_word": false,
353
+ "special": true
354
+ },
355
+ "128044": {
356
+ "content": "__is__",
357
+ "lstrip": false,
358
+ "normalized": false,
359
+ "rstrip": false,
360
+ "single_word": false,
361
+ "special": true
362
+ },
363
+ "128045": {
364
+ "content": "__it__",
365
+ "lstrip": false,
366
+ "normalized": false,
367
+ "rstrip": false,
368
+ "single_word": false,
369
+ "special": true
370
+ },
371
+ "128046": {
372
+ "content": "__ja__",
373
+ "lstrip": false,
374
+ "normalized": false,
375
+ "rstrip": false,
376
+ "single_word": false,
377
+ "special": true
378
+ },
379
+ "128047": {
380
+ "content": "__jv__",
381
+ "lstrip": false,
382
+ "normalized": false,
383
+ "rstrip": false,
384
+ "single_word": false,
385
+ "special": true
386
+ },
387
+ "128048": {
388
+ "content": "__ka__",
389
+ "lstrip": false,
390
+ "normalized": false,
391
+ "rstrip": false,
392
+ "single_word": false,
393
+ "special": true
394
+ },
395
+ "128049": {
396
+ "content": "__kk__",
397
+ "lstrip": false,
398
+ "normalized": false,
399
+ "rstrip": false,
400
+ "single_word": false,
401
+ "special": true
402
+ },
403
+ "128050": {
404
+ "content": "__km__",
405
+ "lstrip": false,
406
+ "normalized": false,
407
+ "rstrip": false,
408
+ "single_word": false,
409
+ "special": true
410
+ },
411
+ "128051": {
412
+ "content": "__kn__",
413
+ "lstrip": false,
414
+ "normalized": false,
415
+ "rstrip": false,
416
+ "single_word": false,
417
+ "special": true
418
+ },
419
+ "128052": {
420
+ "content": "__ko__",
421
+ "lstrip": false,
422
+ "normalized": false,
423
+ "rstrip": false,
424
+ "single_word": false,
425
+ "special": true
426
+ },
427
+ "128053": {
428
+ "content": "__lb__",
429
+ "lstrip": false,
430
+ "normalized": false,
431
+ "rstrip": false,
432
+ "single_word": false,
433
+ "special": true
434
+ },
435
+ "128054": {
436
+ "content": "__lg__",
437
+ "lstrip": false,
438
+ "normalized": false,
439
+ "rstrip": false,
440
+ "single_word": false,
441
+ "special": true
442
+ },
443
+ "128055": {
444
+ "content": "__ln__",
445
+ "lstrip": false,
446
+ "normalized": false,
447
+ "rstrip": false,
448
+ "single_word": false,
449
+ "special": true
450
+ },
451
+ "128056": {
452
+ "content": "__lo__",
453
+ "lstrip": false,
454
+ "normalized": false,
455
+ "rstrip": false,
456
+ "single_word": false,
457
+ "special": true
458
+ },
459
+ "128057": {
460
+ "content": "__lt__",
461
+ "lstrip": false,
462
+ "normalized": false,
463
+ "rstrip": false,
464
+ "single_word": false,
465
+ "special": true
466
+ },
467
+ "128058": {
468
+ "content": "__lv__",
469
+ "lstrip": false,
470
+ "normalized": false,
471
+ "rstrip": false,
472
+ "single_word": false,
473
+ "special": true
474
+ },
475
+ "128059": {
476
+ "content": "__mg__",
477
+ "lstrip": false,
478
+ "normalized": false,
479
+ "rstrip": false,
480
+ "single_word": false,
481
+ "special": true
482
+ },
483
+ "128060": {
484
+ "content": "__mk__",
485
+ "lstrip": false,
486
+ "normalized": false,
487
+ "rstrip": false,
488
+ "single_word": false,
489
+ "special": true
490
+ },
491
+ "128061": {
492
+ "content": "__ml__",
493
+ "lstrip": false,
494
+ "normalized": false,
495
+ "rstrip": false,
496
+ "single_word": false,
497
+ "special": true
498
+ },
499
+ "128062": {
500
+ "content": "__mn__",
501
+ "lstrip": false,
502
+ "normalized": false,
503
+ "rstrip": false,
504
+ "single_word": false,
505
+ "special": true
506
+ },
507
+ "128063": {
508
+ "content": "__mr__",
509
+ "lstrip": false,
510
+ "normalized": false,
511
+ "rstrip": false,
512
+ "single_word": false,
513
+ "special": true
514
+ },
515
+ "128064": {
516
+ "content": "__ms__",
517
+ "lstrip": false,
518
+ "normalized": false,
519
+ "rstrip": false,
520
+ "single_word": false,
521
+ "special": true
522
+ },
523
+ "128065": {
524
+ "content": "__my__",
525
+ "lstrip": false,
526
+ "normalized": false,
527
+ "rstrip": false,
528
+ "single_word": false,
529
+ "special": true
530
+ },
531
+ "128066": {
532
+ "content": "__ne__",
533
+ "lstrip": false,
534
+ "normalized": false,
535
+ "rstrip": false,
536
+ "single_word": false,
537
+ "special": true
538
+ },
539
+ "128067": {
540
+ "content": "__nl__",
541
+ "lstrip": false,
542
+ "normalized": false,
543
+ "rstrip": false,
544
+ "single_word": false,
545
+ "special": true
546
+ },
547
+ "128068": {
548
+ "content": "__no__",
549
+ "lstrip": false,
550
+ "normalized": false,
551
+ "rstrip": false,
552
+ "single_word": false,
553
+ "special": true
554
+ },
555
+ "128069": {
556
+ "content": "__ns__",
557
+ "lstrip": false,
558
+ "normalized": false,
559
+ "rstrip": false,
560
+ "single_word": false,
561
+ "special": true
562
+ },
563
+ "128070": {
564
+ "content": "__oc__",
565
+ "lstrip": false,
566
+ "normalized": false,
567
+ "rstrip": false,
568
+ "single_word": false,
569
+ "special": true
570
+ },
571
+ "128071": {
572
+ "content": "__or__",
573
+ "lstrip": false,
574
+ "normalized": false,
575
+ "rstrip": false,
576
+ "single_word": false,
577
+ "special": true
578
+ },
579
+ "128072": {
580
+ "content": "__pa__",
581
+ "lstrip": false,
582
+ "normalized": false,
583
+ "rstrip": false,
584
+ "single_word": false,
585
+ "special": true
586
+ },
587
+ "128073": {
588
+ "content": "__pl__",
589
+ "lstrip": false,
590
+ "normalized": false,
591
+ "rstrip": false,
592
+ "single_word": false,
593
+ "special": true
594
+ },
595
+ "128074": {
596
+ "content": "__ps__",
597
+ "lstrip": false,
598
+ "normalized": false,
599
+ "rstrip": false,
600
+ "single_word": false,
601
+ "special": true
602
+ },
603
+ "128075": {
604
+ "content": "__pt__",
605
+ "lstrip": false,
606
+ "normalized": false,
607
+ "rstrip": false,
608
+ "single_word": false,
609
+ "special": true
610
+ },
611
+ "128076": {
612
+ "content": "__ro__",
613
+ "lstrip": false,
614
+ "normalized": false,
615
+ "rstrip": false,
616
+ "single_word": false,
617
+ "special": true
618
+ },
619
+ "128077": {
620
+ "content": "__ru__",
621
+ "lstrip": false,
622
+ "normalized": false,
623
+ "rstrip": false,
624
+ "single_word": false,
625
+ "special": true
626
+ },
627
+ "128078": {
628
+ "content": "__sd__",
629
+ "lstrip": false,
630
+ "normalized": false,
631
+ "rstrip": false,
632
+ "single_word": false,
633
+ "special": true
634
+ },
635
+ "128079": {
636
+ "content": "__si__",
637
+ "lstrip": false,
638
+ "normalized": false,
639
+ "rstrip": false,
640
+ "single_word": false,
641
+ "special": true
642
+ },
643
+ "128080": {
644
+ "content": "__sk__",
645
+ "lstrip": false,
646
+ "normalized": false,
647
+ "rstrip": false,
648
+ "single_word": false,
649
+ "special": true
650
+ },
651
+ "128081": {
652
+ "content": "__sl__",
653
+ "lstrip": false,
654
+ "normalized": false,
655
+ "rstrip": false,
656
+ "single_word": false,
657
+ "special": true
658
+ },
659
+ "128082": {
660
+ "content": "__so__",
661
+ "lstrip": false,
662
+ "normalized": false,
663
+ "rstrip": false,
664
+ "single_word": false,
665
+ "special": true
666
+ },
667
+ "128083": {
668
+ "content": "__sq__",
669
+ "lstrip": false,
670
+ "normalized": false,
671
+ "rstrip": false,
672
+ "single_word": false,
673
+ "special": true
674
+ },
675
+ "128084": {
676
+ "content": "__sr__",
677
+ "lstrip": false,
678
+ "normalized": false,
679
+ "rstrip": false,
680
+ "single_word": false,
681
+ "special": true
682
+ },
683
+ "128085": {
684
+ "content": "__ss__",
685
+ "lstrip": false,
686
+ "normalized": false,
687
+ "rstrip": false,
688
+ "single_word": false,
689
+ "special": true
690
+ },
691
+ "128086": {
692
+ "content": "__su__",
693
+ "lstrip": false,
694
+ "normalized": false,
695
+ "rstrip": false,
696
+ "single_word": false,
697
+ "special": true
698
+ },
699
+ "128087": {
700
+ "content": "__sv__",
701
+ "lstrip": false,
702
+ "normalized": false,
703
+ "rstrip": false,
704
+ "single_word": false,
705
+ "special": true
706
+ },
707
+ "128088": {
708
+ "content": "__sw__",
709
+ "lstrip": false,
710
+ "normalized": false,
711
+ "rstrip": false,
712
+ "single_word": false,
713
+ "special": true
714
+ },
715
+ "128089": {
716
+ "content": "__ta__",
717
+ "lstrip": false,
718
+ "normalized": false,
719
+ "rstrip": false,
720
+ "single_word": false,
721
+ "special": true
722
+ },
723
+ "128090": {
724
+ "content": "__th__",
725
+ "lstrip": false,
726
+ "normalized": false,
727
+ "rstrip": false,
728
+ "single_word": false,
729
+ "special": true
730
+ },
731
+ "128091": {
732
+ "content": "__tl__",
733
+ "lstrip": false,
734
+ "normalized": false,
735
+ "rstrip": false,
736
+ "single_word": false,
737
+ "special": true
738
+ },
739
+ "128092": {
740
+ "content": "__tn__",
741
+ "lstrip": false,
742
+ "normalized": false,
743
+ "rstrip": false,
744
+ "single_word": false,
745
+ "special": true
746
+ },
747
+ "128093": {
748
+ "content": "__tr__",
749
+ "lstrip": false,
750
+ "normalized": false,
751
+ "rstrip": false,
752
+ "single_word": false,
753
+ "special": true
754
+ },
755
+ "128094": {
756
+ "content": "__uk__",
757
+ "lstrip": false,
758
+ "normalized": false,
759
+ "rstrip": false,
760
+ "single_word": false,
761
+ "special": true
762
+ },
763
+ "128095": {
764
+ "content": "__ur__",
765
+ "lstrip": false,
766
+ "normalized": false,
767
+ "rstrip": false,
768
+ "single_word": false,
769
+ "special": true
770
+ },
771
+ "128096": {
772
+ "content": "__uz__",
773
+ "lstrip": false,
774
+ "normalized": false,
775
+ "rstrip": false,
776
+ "single_word": false,
777
+ "special": true
778
+ },
779
+ "128097": {
780
+ "content": "__vi__",
781
+ "lstrip": false,
782
+ "normalized": false,
783
+ "rstrip": false,
784
+ "single_word": false,
785
+ "special": true
786
+ },
787
+ "128098": {
788
+ "content": "__wo__",
789
+ "lstrip": false,
790
+ "normalized": false,
791
+ "rstrip": false,
792
+ "single_word": false,
793
+ "special": true
794
+ },
795
+ "128099": {
796
+ "content": "__xh__",
797
+ "lstrip": false,
798
+ "normalized": false,
799
+ "rstrip": false,
800
+ "single_word": false,
801
+ "special": true
802
+ },
803
+ "128100": {
804
+ "content": "__yi__",
805
+ "lstrip": false,
806
+ "normalized": false,
807
+ "rstrip": false,
808
+ "single_word": false,
809
+ "special": true
810
+ },
811
+ "128101": {
812
+ "content": "__yo__",
813
+ "lstrip": false,
814
+ "normalized": false,
815
+ "rstrip": false,
816
+ "single_word": false,
817
+ "special": true
818
+ },
819
+ "128102": {
820
+ "content": "__zh__",
821
+ "lstrip": false,
822
+ "normalized": false,
823
+ "rstrip": false,
824
+ "single_word": false,
825
+ "special": true
826
+ },
827
+ "128103": {
828
+ "content": "__zu__",
829
+ "lstrip": false,
830
+ "normalized": false,
831
+ "rstrip": false,
832
+ "single_word": false,
833
+ "special": true
834
+ }
835
+ },
836
+ "additional_special_tokens": [
837
+ "__af__",
838
+ "__am__",
839
+ "__ar__",
840
+ "__ast__",
841
+ "__az__",
842
+ "__ba__",
843
+ "__be__",
844
+ "__bg__",
845
+ "__bn__",
846
+ "__br__",
847
+ "__bs__",
848
+ "__ca__",
849
+ "__ceb__",
850
+ "__cs__",
851
+ "__cy__",
852
+ "__da__",
853
+ "__de__",
854
+ "__el__",
855
+ "__en__",
856
+ "__es__",
857
+ "__et__",
858
+ "__fa__",
859
+ "__ff__",
860
+ "__fi__",
861
+ "__fr__",
862
+ "__fy__",
863
+ "__ga__",
864
+ "__gd__",
865
+ "__gl__",
866
+ "__gu__",
867
+ "__ha__",
868
+ "__he__",
869
+ "__hi__",
870
+ "__hr__",
871
+ "__ht__",
872
+ "__hu__",
873
+ "__hy__",
874
+ "__id__",
875
+ "__ig__",
876
+ "__ilo__",
877
+ "__is__",
878
+ "__it__",
879
+ "__ja__",
880
+ "__jv__",
881
+ "__ka__",
882
+ "__kk__",
883
+ "__km__",
884
+ "__kn__",
885
+ "__ko__",
886
+ "__lb__",
887
+ "__lg__",
888
+ "__ln__",
889
+ "__lo__",
890
+ "__lt__",
891
+ "__lv__",
892
+ "__mg__",
893
+ "__mk__",
894
+ "__ml__",
895
+ "__mn__",
896
+ "__mr__",
897
+ "__ms__",
898
+ "__my__",
899
+ "__ne__",
900
+ "__nl__",
901
+ "__no__",
902
+ "__ns__",
903
+ "__oc__",
904
+ "__or__",
905
+ "__pa__",
906
+ "__pl__",
907
+ "__ps__",
908
+ "__pt__",
909
+ "__ro__",
910
+ "__ru__",
911
+ "__sd__",
912
+ "__si__",
913
+ "__sk__",
914
+ "__sl__",
915
+ "__so__",
916
+ "__sq__",
917
+ "__sr__",
918
+ "__ss__",
919
+ "__su__",
920
+ "__sv__",
921
+ "__sw__",
922
+ "__ta__",
923
+ "__th__",
924
+ "__tl__",
925
+ "__tn__",
926
+ "__tr__",
927
+ "__uk__",
928
+ "__ur__",
929
+ "__uz__",
930
+ "__vi__",
931
+ "__wo__",
932
+ "__xh__",
933
+ "__yi__",
934
+ "__yo__",
935
+ "__zh__",
936
+ "__zu__"
937
+ ],
938
+ "bos_token": "<s>",
939
+ "clean_up_tokenization_spaces": true,
940
+ "eos_token": "</s>",
941
+ "language_codes": "m2m100",
942
+ "model_max_length": 1024,
943
+ "num_madeup_words": 8,
944
+ "pad_token": "<pad>",
945
+ "sep_token": "</s>",
946
+ "sp_model_kwargs": {},
947
+ "src_lang": "en",
948
+ "tgt_lang": "ne",
949
+ "tokenizer_class": "M2M100Tokenizer",
950
+ "unk_token": "<unk>"
951
+ }
trainer_state.json ADDED
@@ -0,0 +1,2993 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.24187567830085754,
3
+ "best_model_checkpoint": "./results\\checkpoint-150000",
4
+ "epoch": 5.318552300921432,
5
+ "eval_steps": 10000,
6
+ "global_step": 200000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.013296281030195855,
13
+ "grad_norm": 1.8945389986038208,
14
+ "learning_rate": 1.9986783320923308e-05,
15
+ "loss": 2.1575,
16
+ "step": 500
17
+ },
18
+ {
19
+ "epoch": 0.02659256206039171,
20
+ "grad_norm": 1.8132811784744263,
21
+ "learning_rate": 1.997348686309967e-05,
22
+ "loss": 0.6486,
23
+ "step": 1000
24
+ },
25
+ {
26
+ "epoch": 0.039888843090587565,
27
+ "grad_norm": 1.9245736598968506,
28
+ "learning_rate": 1.9960190405276038e-05,
29
+ "loss": 0.6164,
30
+ "step": 1500
31
+ },
32
+ {
33
+ "epoch": 0.05318512412078342,
34
+ "grad_norm": 1.4257850646972656,
35
+ "learning_rate": 1.99468939474524e-05,
36
+ "loss": 0.6067,
37
+ "step": 2000
38
+ },
39
+ {
40
+ "epoch": 0.06648140515097926,
41
+ "grad_norm": 1.434894323348999,
42
+ "learning_rate": 1.9933597489628764e-05,
43
+ "loss": 0.5671,
44
+ "step": 2500
45
+ },
46
+ {
47
+ "epoch": 0.07977768618117513,
48
+ "grad_norm": 1.8151600360870361,
49
+ "learning_rate": 1.992030103180513e-05,
50
+ "loss": 0.5535,
51
+ "step": 3000
52
+ },
53
+ {
54
+ "epoch": 0.09307396721137098,
55
+ "grad_norm": 1.9667155742645264,
56
+ "learning_rate": 1.990700457398149e-05,
57
+ "loss": 0.5628,
58
+ "step": 3500
59
+ },
60
+ {
61
+ "epoch": 0.10637024824156684,
62
+ "grad_norm": 2.0493931770324707,
63
+ "learning_rate": 1.9893708116157857e-05,
64
+ "loss": 0.5465,
65
+ "step": 4000
66
+ },
67
+ {
68
+ "epoch": 0.11966652927176269,
69
+ "grad_norm": 2.201901435852051,
70
+ "learning_rate": 1.988041165833422e-05,
71
+ "loss": 0.5472,
72
+ "step": 4500
73
+ },
74
+ {
75
+ "epoch": 0.13296281030195853,
76
+ "grad_norm": 1.8538507223129272,
77
+ "learning_rate": 1.9867115200510584e-05,
78
+ "loss": 0.5451,
79
+ "step": 5000
80
+ },
81
+ {
82
+ "epoch": 0.1462590913321544,
83
+ "grad_norm": 1.35111403465271,
84
+ "learning_rate": 1.985381874268695e-05,
85
+ "loss": 0.5333,
86
+ "step": 5500
87
+ },
88
+ {
89
+ "epoch": 0.15955537236235026,
90
+ "grad_norm": 1.2387117147445679,
91
+ "learning_rate": 1.9840522284863314e-05,
92
+ "loss": 0.5096,
93
+ "step": 6000
94
+ },
95
+ {
96
+ "epoch": 0.1728516533925461,
97
+ "grad_norm": 1.2727868556976318,
98
+ "learning_rate": 1.9827225827039677e-05,
99
+ "loss": 0.5256,
100
+ "step": 6500
101
+ },
102
+ {
103
+ "epoch": 0.18614793442274197,
104
+ "grad_norm": 1.7770028114318848,
105
+ "learning_rate": 1.9813929369216044e-05,
106
+ "loss": 0.5064,
107
+ "step": 7000
108
+ },
109
+ {
110
+ "epoch": 0.1994442154529378,
111
+ "grad_norm": 2.1460671424865723,
112
+ "learning_rate": 1.9800632911392407e-05,
113
+ "loss": 0.5066,
114
+ "step": 7500
115
+ },
116
+ {
117
+ "epoch": 0.21274049648313367,
118
+ "grad_norm": 1.315503716468811,
119
+ "learning_rate": 1.978733645356877e-05,
120
+ "loss": 0.4974,
121
+ "step": 8000
122
+ },
123
+ {
124
+ "epoch": 0.2260367775133295,
125
+ "grad_norm": 1.6402277946472168,
126
+ "learning_rate": 1.9774039995745137e-05,
127
+ "loss": 0.5019,
128
+ "step": 8500
129
+ },
130
+ {
131
+ "epoch": 0.23933305854352538,
132
+ "grad_norm": 1.4794352054595947,
133
+ "learning_rate": 1.97607435379215e-05,
134
+ "loss": 0.5072,
135
+ "step": 9000
136
+ },
137
+ {
138
+ "epoch": 0.25262933957372125,
139
+ "grad_norm": 1.432265043258667,
140
+ "learning_rate": 1.9747447080097863e-05,
141
+ "loss": 0.4844,
142
+ "step": 9500
143
+ },
144
+ {
145
+ "epoch": 0.26592562060391706,
146
+ "grad_norm": 1.7145007848739624,
147
+ "learning_rate": 1.973415062227423e-05,
148
+ "loss": 0.488,
149
+ "step": 10000
150
+ },
151
+ {
152
+ "epoch": 0.26592562060391706,
153
+ "eval_loss": 0.4845176041126251,
154
+ "eval_runtime": 45.8707,
155
+ "eval_samples_per_second": 66.251,
156
+ "eval_steps_per_second": 16.568,
157
+ "step": 10000
158
+ },
159
+ {
160
+ "epoch": 0.2792219016341129,
161
+ "grad_norm": 1.404540777206421,
162
+ "learning_rate": 1.972085416445059e-05,
163
+ "loss": 0.4931,
164
+ "step": 10500
165
+ },
166
+ {
167
+ "epoch": 0.2925181826643088,
168
+ "grad_norm": 1.3520866632461548,
169
+ "learning_rate": 1.9707557706626957e-05,
170
+ "loss": 0.4868,
171
+ "step": 11000
172
+ },
173
+ {
174
+ "epoch": 0.30581446369450466,
175
+ "grad_norm": 1.4363296031951904,
176
+ "learning_rate": 1.969426124880332e-05,
177
+ "loss": 0.4798,
178
+ "step": 11500
179
+ },
180
+ {
181
+ "epoch": 0.3191107447247005,
182
+ "grad_norm": 1.4917619228363037,
183
+ "learning_rate": 1.9680964790979683e-05,
184
+ "loss": 0.4736,
185
+ "step": 12000
186
+ },
187
+ {
188
+ "epoch": 0.33240702575489633,
189
+ "grad_norm": 1.7679712772369385,
190
+ "learning_rate": 1.966766833315605e-05,
191
+ "loss": 0.4778,
192
+ "step": 12500
193
+ },
194
+ {
195
+ "epoch": 0.3457033067850922,
196
+ "grad_norm": 1.4026507139205933,
197
+ "learning_rate": 1.9654371875332413e-05,
198
+ "loss": 0.4761,
199
+ "step": 13000
200
+ },
201
+ {
202
+ "epoch": 0.35899958781528807,
203
+ "grad_norm": 1.8736493587493896,
204
+ "learning_rate": 1.9641075417508776e-05,
205
+ "loss": 0.4871,
206
+ "step": 13500
207
+ },
208
+ {
209
+ "epoch": 0.37229586884548393,
210
+ "grad_norm": 1.2222760915756226,
211
+ "learning_rate": 1.9627778959685143e-05,
212
+ "loss": 0.4745,
213
+ "step": 14000
214
+ },
215
+ {
216
+ "epoch": 0.38559214987567975,
217
+ "grad_norm": 1.4624443054199219,
218
+ "learning_rate": 1.9614509094777154e-05,
219
+ "loss": 0.4573,
220
+ "step": 14500
221
+ },
222
+ {
223
+ "epoch": 0.3988884309058756,
224
+ "grad_norm": 1.8538755178451538,
225
+ "learning_rate": 1.9601212636953517e-05,
226
+ "loss": 0.4633,
227
+ "step": 15000
228
+ },
229
+ {
230
+ "epoch": 0.4121847119360715,
231
+ "grad_norm": 1.47249174118042,
232
+ "learning_rate": 1.9587916179129884e-05,
233
+ "loss": 0.4764,
234
+ "step": 15500
235
+ },
236
+ {
237
+ "epoch": 0.42548099296626735,
238
+ "grad_norm": 1.7755292654037476,
239
+ "learning_rate": 1.9574619721306244e-05,
240
+ "loss": 0.4699,
241
+ "step": 16000
242
+ },
243
+ {
244
+ "epoch": 0.4387772739964632,
245
+ "grad_norm": 1.4009391069412231,
246
+ "learning_rate": 1.956132326348261e-05,
247
+ "loss": 0.4666,
248
+ "step": 16500
249
+ },
250
+ {
251
+ "epoch": 0.452073555026659,
252
+ "grad_norm": 1.520727276802063,
253
+ "learning_rate": 1.9548026805658973e-05,
254
+ "loss": 0.4697,
255
+ "step": 17000
256
+ },
257
+ {
258
+ "epoch": 0.4653698360568549,
259
+ "grad_norm": 1.7645255327224731,
260
+ "learning_rate": 1.9534730347835337e-05,
261
+ "loss": 0.46,
262
+ "step": 17500
263
+ },
264
+ {
265
+ "epoch": 0.47866611708705076,
266
+ "grad_norm": 1.7914655208587646,
267
+ "learning_rate": 1.9521433890011703e-05,
268
+ "loss": 0.4601,
269
+ "step": 18000
270
+ },
271
+ {
272
+ "epoch": 0.4919623981172466,
273
+ "grad_norm": 1.4558111429214478,
274
+ "learning_rate": 1.9508164025103714e-05,
275
+ "loss": 0.4601,
276
+ "step": 18500
277
+ },
278
+ {
279
+ "epoch": 0.5052586791474425,
280
+ "grad_norm": 1.1980383396148682,
281
+ "learning_rate": 1.9494894160195725e-05,
282
+ "loss": 0.4563,
283
+ "step": 19000
284
+ },
285
+ {
286
+ "epoch": 0.5185549601776384,
287
+ "grad_norm": 1.8653793334960938,
288
+ "learning_rate": 1.9481597702372092e-05,
289
+ "loss": 0.4434,
290
+ "step": 19500
291
+ },
292
+ {
293
+ "epoch": 0.5318512412078341,
294
+ "grad_norm": 1.3913460969924927,
295
+ "learning_rate": 1.946830124454845e-05,
296
+ "loss": 0.45,
297
+ "step": 20000
298
+ },
299
+ {
300
+ "epoch": 0.5318512412078341,
301
+ "eval_loss": 0.4461110234260559,
302
+ "eval_runtime": 46.6164,
303
+ "eval_samples_per_second": 65.192,
304
+ "eval_steps_per_second": 16.303,
305
+ "step": 20000
306
+ },
307
+ {
308
+ "epoch": 0.54514752223803,
309
+ "grad_norm": 1.4441711902618408,
310
+ "learning_rate": 1.9455004786724818e-05,
311
+ "loss": 0.4536,
312
+ "step": 20500
313
+ },
314
+ {
315
+ "epoch": 0.5584438032682258,
316
+ "grad_norm": 1.2844038009643555,
317
+ "learning_rate": 1.9441708328901185e-05,
318
+ "loss": 0.447,
319
+ "step": 21000
320
+ },
321
+ {
322
+ "epoch": 0.5717400842984217,
323
+ "grad_norm": 2.10949444770813,
324
+ "learning_rate": 1.9428438463993192e-05,
325
+ "loss": 0.4484,
326
+ "step": 21500
327
+ },
328
+ {
329
+ "epoch": 0.5850363653286176,
330
+ "grad_norm": 0.925681471824646,
331
+ "learning_rate": 1.941514200616956e-05,
332
+ "loss": 0.4374,
333
+ "step": 22000
334
+ },
335
+ {
336
+ "epoch": 0.5983326463588134,
337
+ "grad_norm": 1.49970543384552,
338
+ "learning_rate": 1.9401845548345922e-05,
339
+ "loss": 0.4533,
340
+ "step": 22500
341
+ },
342
+ {
343
+ "epoch": 0.6116289273890093,
344
+ "grad_norm": 1.2436000108718872,
345
+ "learning_rate": 1.9388549090522286e-05,
346
+ "loss": 0.4455,
347
+ "step": 23000
348
+ },
349
+ {
350
+ "epoch": 0.6249252084192052,
351
+ "grad_norm": 1.1918786764144897,
352
+ "learning_rate": 1.9375252632698652e-05,
353
+ "loss": 0.4389,
354
+ "step": 23500
355
+ },
356
+ {
357
+ "epoch": 0.638221489449401,
358
+ "grad_norm": 1.2042317390441895,
359
+ "learning_rate": 1.9361956174875015e-05,
360
+ "loss": 0.4446,
361
+ "step": 24000
362
+ },
363
+ {
364
+ "epoch": 0.6515177704795968,
365
+ "grad_norm": 1.3460853099822998,
366
+ "learning_rate": 1.934865971705138e-05,
367
+ "loss": 0.4464,
368
+ "step": 24500
369
+ },
370
+ {
371
+ "epoch": 0.6648140515097927,
372
+ "grad_norm": 1.1541383266448975,
373
+ "learning_rate": 1.9335363259227745e-05,
374
+ "loss": 0.4404,
375
+ "step": 25000
376
+ },
377
+ {
378
+ "epoch": 0.6781103325399885,
379
+ "grad_norm": 1.2001688480377197,
380
+ "learning_rate": 1.9322093394319753e-05,
381
+ "loss": 0.4409,
382
+ "step": 25500
383
+ },
384
+ {
385
+ "epoch": 0.6914066135701844,
386
+ "grad_norm": 1.38569974899292,
387
+ "learning_rate": 1.930879693649612e-05,
388
+ "loss": 0.4412,
389
+ "step": 26000
390
+ },
391
+ {
392
+ "epoch": 0.7047028946003803,
393
+ "grad_norm": 1.6519745588302612,
394
+ "learning_rate": 1.9295500478672483e-05,
395
+ "loss": 0.4293,
396
+ "step": 26500
397
+ },
398
+ {
399
+ "epoch": 0.7179991756305761,
400
+ "grad_norm": 1.1891002655029297,
401
+ "learning_rate": 1.9282230613764494e-05,
402
+ "loss": 0.4478,
403
+ "step": 27000
404
+ },
405
+ {
406
+ "epoch": 0.731295456660772,
407
+ "grad_norm": 1.2890375852584839,
408
+ "learning_rate": 1.926893415594086e-05,
409
+ "loss": 0.4422,
410
+ "step": 27500
411
+ },
412
+ {
413
+ "epoch": 0.7445917376909679,
414
+ "grad_norm": 1.2893614768981934,
415
+ "learning_rate": 1.9255637698117224e-05,
416
+ "loss": 0.4379,
417
+ "step": 28000
418
+ },
419
+ {
420
+ "epoch": 0.7578880187211637,
421
+ "grad_norm": 1.2751696109771729,
422
+ "learning_rate": 1.9242341240293587e-05,
423
+ "loss": 0.4259,
424
+ "step": 28500
425
+ },
426
+ {
427
+ "epoch": 0.7711842997513595,
428
+ "grad_norm": 1.263238549232483,
429
+ "learning_rate": 1.9229044782469953e-05,
430
+ "loss": 0.4285,
431
+ "step": 29000
432
+ },
433
+ {
434
+ "epoch": 0.7844805807815554,
435
+ "grad_norm": 1.6520916223526,
436
+ "learning_rate": 1.921577491756196e-05,
437
+ "loss": 0.4278,
438
+ "step": 29500
439
+ },
440
+ {
441
+ "epoch": 0.7977768618117512,
442
+ "grad_norm": 1.4022353887557983,
443
+ "learning_rate": 1.9202478459738328e-05,
444
+ "loss": 0.4277,
445
+ "step": 30000
446
+ },
447
+ {
448
+ "epoch": 0.7977768618117512,
449
+ "eval_loss": 0.4259246587753296,
450
+ "eval_runtime": 45.6826,
451
+ "eval_samples_per_second": 66.524,
452
+ "eval_steps_per_second": 16.637,
453
+ "step": 30000
454
+ },
455
+ {
456
+ "epoch": 0.8110731428419471,
457
+ "grad_norm": 1.3312253952026367,
458
+ "learning_rate": 1.918920859483034e-05,
459
+ "loss": 0.4262,
460
+ "step": 30500
461
+ },
462
+ {
463
+ "epoch": 0.824369423872143,
464
+ "grad_norm": 1.5275497436523438,
465
+ "learning_rate": 1.91759121370067e-05,
466
+ "loss": 0.4281,
467
+ "step": 31000
468
+ },
469
+ {
470
+ "epoch": 0.8376657049023388,
471
+ "grad_norm": 1.7679028511047363,
472
+ "learning_rate": 1.916261567918307e-05,
473
+ "loss": 0.44,
474
+ "step": 31500
475
+ },
476
+ {
477
+ "epoch": 0.8509619859325347,
478
+ "grad_norm": 1.1467547416687012,
479
+ "learning_rate": 1.914931922135943e-05,
480
+ "loss": 0.4153,
481
+ "step": 32000
482
+ },
483
+ {
484
+ "epoch": 0.8642582669627306,
485
+ "grad_norm": 1.2685679197311401,
486
+ "learning_rate": 1.9136022763535795e-05,
487
+ "loss": 0.4232,
488
+ "step": 32500
489
+ },
490
+ {
491
+ "epoch": 0.8775545479929264,
492
+ "grad_norm": 0.9204725623130798,
493
+ "learning_rate": 1.912272630571216e-05,
494
+ "loss": 0.425,
495
+ "step": 33000
496
+ },
497
+ {
498
+ "epoch": 0.8908508290231222,
499
+ "grad_norm": 0.7971585988998413,
500
+ "learning_rate": 1.9109429847888525e-05,
501
+ "loss": 0.431,
502
+ "step": 33500
503
+ },
504
+ {
505
+ "epoch": 0.904147110053318,
506
+ "grad_norm": 1.6358304023742676,
507
+ "learning_rate": 1.9096133390064888e-05,
508
+ "loss": 0.4361,
509
+ "step": 34000
510
+ },
511
+ {
512
+ "epoch": 0.9174433910835139,
513
+ "grad_norm": 1.7240855693817139,
514
+ "learning_rate": 1.9082836932241255e-05,
515
+ "loss": 0.432,
516
+ "step": 34500
517
+ },
518
+ {
519
+ "epoch": 0.9307396721137098,
520
+ "grad_norm": 1.1937910318374634,
521
+ "learning_rate": 1.9069567067333262e-05,
522
+ "loss": 0.4324,
523
+ "step": 35000
524
+ },
525
+ {
526
+ "epoch": 0.9440359531439056,
527
+ "grad_norm": 1.1422160863876343,
528
+ "learning_rate": 1.905627060950963e-05,
529
+ "loss": 0.4286,
530
+ "step": 35500
531
+ },
532
+ {
533
+ "epoch": 0.9573322341741015,
534
+ "grad_norm": 1.0084590911865234,
535
+ "learning_rate": 1.9042974151685992e-05,
536
+ "loss": 0.4401,
537
+ "step": 36000
538
+ },
539
+ {
540
+ "epoch": 0.9706285152042974,
541
+ "grad_norm": 1.7879573106765747,
542
+ "learning_rate": 1.9029677693862355e-05,
543
+ "loss": 0.4259,
544
+ "step": 36500
545
+ },
546
+ {
547
+ "epoch": 0.9839247962344932,
548
+ "grad_norm": 1.1148507595062256,
549
+ "learning_rate": 1.9016381236038722e-05,
550
+ "loss": 0.3995,
551
+ "step": 37000
552
+ },
553
+ {
554
+ "epoch": 0.9972210772646891,
555
+ "grad_norm": 1.2119601964950562,
556
+ "learning_rate": 1.9003084778215085e-05,
557
+ "loss": 0.4228,
558
+ "step": 37500
559
+ },
560
+ {
561
+ "epoch": 1.010517358294885,
562
+ "grad_norm": 1.2811840772628784,
563
+ "learning_rate": 1.898978832039145e-05,
564
+ "loss": 0.3945,
565
+ "step": 38000
566
+ },
567
+ {
568
+ "epoch": 1.0238136393250807,
569
+ "grad_norm": 1.199591040611267,
570
+ "learning_rate": 1.8976491862567815e-05,
571
+ "loss": 0.3724,
572
+ "step": 38500
573
+ },
574
+ {
575
+ "epoch": 1.0371099203552767,
576
+ "grad_norm": 1.4783821105957031,
577
+ "learning_rate": 1.8963221997659823e-05,
578
+ "loss": 0.3869,
579
+ "step": 39000
580
+ },
581
+ {
582
+ "epoch": 1.0504062013854725,
583
+ "grad_norm": 1.1598395109176636,
584
+ "learning_rate": 1.894992553983619e-05,
585
+ "loss": 0.3888,
586
+ "step": 39500
587
+ },
588
+ {
589
+ "epoch": 1.0637024824156684,
590
+ "grad_norm": 1.0254014730453491,
591
+ "learning_rate": 1.8936629082012556e-05,
592
+ "loss": 0.3741,
593
+ "step": 40000
594
+ },
595
+ {
596
+ "epoch": 1.0637024824156684,
597
+ "eval_loss": 0.4110487699508667,
598
+ "eval_runtime": 45.9661,
599
+ "eval_samples_per_second": 66.114,
600
+ "eval_steps_per_second": 16.534,
601
+ "step": 40000
602
+ },
603
+ {
604
+ "epoch": 1.0769987634458642,
605
+ "grad_norm": 1.4140042066574097,
606
+ "learning_rate": 1.8923332624188916e-05,
607
+ "loss": 0.3838,
608
+ "step": 40500
609
+ },
610
+ {
611
+ "epoch": 1.09029504447606,
612
+ "grad_norm": 1.4062950611114502,
613
+ "learning_rate": 1.8910089352196578e-05,
614
+ "loss": 0.368,
615
+ "step": 41000
616
+ },
617
+ {
618
+ "epoch": 1.103591325506256,
619
+ "grad_norm": 1.2314609289169312,
620
+ "learning_rate": 1.889681948728859e-05,
621
+ "loss": 0.4033,
622
+ "step": 41500
623
+ },
624
+ {
625
+ "epoch": 1.1168876065364517,
626
+ "grad_norm": 1.4966250658035278,
627
+ "learning_rate": 1.8883523029464952e-05,
628
+ "loss": 0.3878,
629
+ "step": 42000
630
+ },
631
+ {
632
+ "epoch": 1.1301838875666477,
633
+ "grad_norm": 1.1920274496078491,
634
+ "learning_rate": 1.887022657164132e-05,
635
+ "loss": 0.375,
636
+ "step": 42500
637
+ },
638
+ {
639
+ "epoch": 1.1434801685968434,
640
+ "grad_norm": 1.0075514316558838,
641
+ "learning_rate": 1.8856930113817678e-05,
642
+ "loss": 0.3821,
643
+ "step": 43000
644
+ },
645
+ {
646
+ "epoch": 1.1567764496270394,
647
+ "grad_norm": 1.2424709796905518,
648
+ "learning_rate": 1.8843633655994045e-05,
649
+ "loss": 0.3793,
650
+ "step": 43500
651
+ },
652
+ {
653
+ "epoch": 1.1700727306572352,
654
+ "grad_norm": 0.7054141163825989,
655
+ "learning_rate": 1.8830337198170408e-05,
656
+ "loss": 0.3821,
657
+ "step": 44000
658
+ },
659
+ {
660
+ "epoch": 1.183369011687431,
661
+ "grad_norm": 0.7828935384750366,
662
+ "learning_rate": 1.881704074034677e-05,
663
+ "loss": 0.3737,
664
+ "step": 44500
665
+ },
666
+ {
667
+ "epoch": 1.196665292717627,
668
+ "grad_norm": 1.3097379207611084,
669
+ "learning_rate": 1.8803744282523138e-05,
670
+ "loss": 0.3747,
671
+ "step": 45000
672
+ },
673
+ {
674
+ "epoch": 1.2099615737478226,
675
+ "grad_norm": 1.2207602262496948,
676
+ "learning_rate": 1.87904478246995e-05,
677
+ "loss": 0.3799,
678
+ "step": 45500
679
+ },
680
+ {
681
+ "epoch": 1.2232578547780186,
682
+ "grad_norm": 1.5833189487457275,
683
+ "learning_rate": 1.8777151366875865e-05,
684
+ "loss": 0.387,
685
+ "step": 46000
686
+ },
687
+ {
688
+ "epoch": 1.2365541358082144,
689
+ "grad_norm": 1.3836820125579834,
690
+ "learning_rate": 1.876385490905223e-05,
691
+ "loss": 0.3853,
692
+ "step": 46500
693
+ },
694
+ {
695
+ "epoch": 1.2498504168384104,
696
+ "grad_norm": 0.9267215728759766,
697
+ "learning_rate": 1.8750558451228595e-05,
698
+ "loss": 0.3763,
699
+ "step": 47000
700
+ },
701
+ {
702
+ "epoch": 1.2631466978686061,
703
+ "grad_norm": 1.4826302528381348,
704
+ "learning_rate": 1.8737288586320605e-05,
705
+ "loss": 0.3753,
706
+ "step": 47500
707
+ },
708
+ {
709
+ "epoch": 1.2764429788988019,
710
+ "grad_norm": 1.7470266819000244,
711
+ "learning_rate": 1.8723992128496972e-05,
712
+ "loss": 0.389,
713
+ "step": 48000
714
+ },
715
+ {
716
+ "epoch": 1.2897392599289978,
717
+ "grad_norm": 0.9160776734352112,
718
+ "learning_rate": 1.8710695670673332e-05,
719
+ "loss": 0.3732,
720
+ "step": 48500
721
+ },
722
+ {
723
+ "epoch": 1.3030355409591938,
724
+ "grad_norm": 1.5541558265686035,
725
+ "learning_rate": 1.86973992128497e-05,
726
+ "loss": 0.3815,
727
+ "step": 49000
728
+ },
729
+ {
730
+ "epoch": 1.3163318219893896,
731
+ "grad_norm": 1.3254436254501343,
732
+ "learning_rate": 1.8684102755026062e-05,
733
+ "loss": 0.3786,
734
+ "step": 49500
735
+ },
736
+ {
737
+ "epoch": 1.3296281030195853,
738
+ "grad_norm": 1.5244851112365723,
739
+ "learning_rate": 1.8670806297202425e-05,
740
+ "loss": 0.3886,
741
+ "step": 50000
742
+ },
743
+ {
744
+ "epoch": 1.3296281030195853,
745
+ "eval_loss": 0.4009094834327698,
746
+ "eval_runtime": 45.5454,
747
+ "eval_samples_per_second": 66.725,
748
+ "eval_steps_per_second": 16.687,
749
+ "step": 50000
750
+ },
751
+ {
752
+ "epoch": 1.3429376803308115,
753
+ "grad_norm": 0.867463231086731,
754
+ "learning_rate": 1.8657509839378792e-05,
755
+ "loss": 0.3681,
756
+ "step": 50500
757
+ },
758
+ {
759
+ "epoch": 1.3562339613610073,
760
+ "grad_norm": 1.2311478853225708,
761
+ "learning_rate": 1.8644213381555155e-05,
762
+ "loss": 0.347,
763
+ "step": 51000
764
+ },
765
+ {
766
+ "epoch": 1.3695302423912032,
767
+ "grad_norm": 1.4698890447616577,
768
+ "learning_rate": 1.8630916923731518e-05,
769
+ "loss": 0.3666,
770
+ "step": 51500
771
+ },
772
+ {
773
+ "epoch": 1.382826523421399,
774
+ "grad_norm": 1.3057785034179688,
775
+ "learning_rate": 1.8617620465907885e-05,
776
+ "loss": 0.3716,
777
+ "step": 52000
778
+ },
779
+ {
780
+ "epoch": 1.3961228044515948,
781
+ "grad_norm": 0.9626977443695068,
782
+ "learning_rate": 1.8604324008084248e-05,
783
+ "loss": 0.3633,
784
+ "step": 52500
785
+ },
786
+ {
787
+ "epoch": 1.4094190854817907,
788
+ "grad_norm": 1.055158257484436,
789
+ "learning_rate": 1.859102755026061e-05,
790
+ "loss": 0.364,
791
+ "step": 53000
792
+ },
793
+ {
794
+ "epoch": 1.4227153665119867,
795
+ "grad_norm": 1.7065026760101318,
796
+ "learning_rate": 1.8577731092436978e-05,
797
+ "loss": 0.365,
798
+ "step": 53500
799
+ },
800
+ {
801
+ "epoch": 1.4360116475421825,
802
+ "grad_norm": 1.1983647346496582,
803
+ "learning_rate": 1.856443463461334e-05,
804
+ "loss": 0.3584,
805
+ "step": 54000
806
+ },
807
+ {
808
+ "epoch": 1.4493079285723782,
809
+ "grad_norm": 1.465380072593689,
810
+ "learning_rate": 1.8551138176789705e-05,
811
+ "loss": 0.3698,
812
+ "step": 54500
813
+ },
814
+ {
815
+ "epoch": 1.4626042096025742,
816
+ "grad_norm": 0.9600203037261963,
817
+ "learning_rate": 1.853784171896607e-05,
818
+ "loss": 0.3686,
819
+ "step": 55000
820
+ },
821
+ {
822
+ "epoch": 1.47590049063277,
823
+ "grad_norm": 0.7194857597351074,
824
+ "learning_rate": 1.852454526114243e-05,
825
+ "loss": 0.3691,
826
+ "step": 55500
827
+ },
828
+ {
829
+ "epoch": 1.489196771662966,
830
+ "grad_norm": 1.2525266408920288,
831
+ "learning_rate": 1.8511248803318798e-05,
832
+ "loss": 0.3716,
833
+ "step": 56000
834
+ },
835
+ {
836
+ "epoch": 1.5024930526931617,
837
+ "grad_norm": 1.4390616416931152,
838
+ "learning_rate": 1.849795234549516e-05,
839
+ "loss": 0.3642,
840
+ "step": 56500
841
+ },
842
+ {
843
+ "epoch": 1.5157893337233577,
844
+ "grad_norm": 0.9960741996765137,
845
+ "learning_rate": 1.8484655887671524e-05,
846
+ "loss": 0.3543,
847
+ "step": 57000
848
+ },
849
+ {
850
+ "epoch": 1.5290856147535534,
851
+ "grad_norm": 1.5157722234725952,
852
+ "learning_rate": 1.847135942984789e-05,
853
+ "loss": 0.3684,
854
+ "step": 57500
855
+ },
856
+ {
857
+ "epoch": 1.5423818957837492,
858
+ "grad_norm": 1.2276800870895386,
859
+ "learning_rate": 1.8458062972024254e-05,
860
+ "loss": 0.3682,
861
+ "step": 58000
862
+ },
863
+ {
864
+ "epoch": 1.5556781768139452,
865
+ "grad_norm": 0.7182838916778564,
866
+ "learning_rate": 1.8444793107116265e-05,
867
+ "loss": 0.3571,
868
+ "step": 58500
869
+ },
870
+ {
871
+ "epoch": 1.5689744578441411,
872
+ "grad_norm": 1.4738359451293945,
873
+ "learning_rate": 1.8431496649292632e-05,
874
+ "loss": 0.3495,
875
+ "step": 59000
876
+ },
877
+ {
878
+ "epoch": 1.582270738874337,
879
+ "grad_norm": 1.1519092321395874,
880
+ "learning_rate": 1.8418200191468995e-05,
881
+ "loss": 0.3656,
882
+ "step": 59500
883
+ },
884
+ {
885
+ "epoch": 1.5955670199045326,
886
+ "grad_norm": 0.9778446555137634,
887
+ "learning_rate": 1.8404903733645358e-05,
888
+ "loss": 0.3528,
889
+ "step": 60000
890
+ },
891
+ {
892
+ "epoch": 1.5955670199045326,
893
+ "eval_loss": 0.33645519614219666,
894
+ "eval_runtime": 45.1628,
895
+ "eval_samples_per_second": 67.29,
896
+ "eval_steps_per_second": 16.828,
897
+ "step": 60000
898
+ },
899
+ {
900
+ "epoch": 1.6088633009347286,
901
+ "grad_norm": 1.5486047267913818,
902
+ "learning_rate": 1.8391607275821725e-05,
903
+ "loss": 0.3587,
904
+ "step": 60500
905
+ },
906
+ {
907
+ "epoch": 1.6221595819649244,
908
+ "grad_norm": 0.8840271234512329,
909
+ "learning_rate": 1.8378337410913732e-05,
910
+ "loss": 0.3584,
911
+ "step": 61000
912
+ },
913
+ {
914
+ "epoch": 1.6354558629951201,
915
+ "grad_norm": 0.9995741248130798,
916
+ "learning_rate": 1.83650409530901e-05,
917
+ "loss": 0.3737,
918
+ "step": 61500
919
+ },
920
+ {
921
+ "epoch": 1.6487521440253161,
922
+ "grad_norm": 1.7882977724075317,
923
+ "learning_rate": 1.835177108818211e-05,
924
+ "loss": 0.3655,
925
+ "step": 62000
926
+ },
927
+ {
928
+ "epoch": 1.662048425055512,
929
+ "grad_norm": 0.7783369421958923,
930
+ "learning_rate": 1.8338474630358473e-05,
931
+ "loss": 0.3673,
932
+ "step": 62500
933
+ },
934
+ {
935
+ "epoch": 1.6753447060857078,
936
+ "grad_norm": 1.585405945777893,
937
+ "learning_rate": 1.832517817253484e-05,
938
+ "loss": 0.3595,
939
+ "step": 63000
940
+ },
941
+ {
942
+ "epoch": 1.6886409871159036,
943
+ "grad_norm": 1.3467520475387573,
944
+ "learning_rate": 1.8311881714711203e-05,
945
+ "loss": 0.3627,
946
+ "step": 63500
947
+ },
948
+ {
949
+ "epoch": 1.7019372681460996,
950
+ "grad_norm": 1.1446176767349243,
951
+ "learning_rate": 1.8298585256887566e-05,
952
+ "loss": 0.3574,
953
+ "step": 64000
954
+ },
955
+ {
956
+ "epoch": 1.7152335491762953,
957
+ "grad_norm": 0.9003855586051941,
958
+ "learning_rate": 1.828531539197958e-05,
959
+ "loss": 0.3665,
960
+ "step": 64500
961
+ },
962
+ {
963
+ "epoch": 1.728529830206491,
964
+ "grad_norm": 0.8468024730682373,
965
+ "learning_rate": 1.827201893415594e-05,
966
+ "loss": 0.3574,
967
+ "step": 65000
968
+ },
969
+ {
970
+ "epoch": 1.741826111236687,
971
+ "grad_norm": 1.0654727220535278,
972
+ "learning_rate": 1.8258722476332307e-05,
973
+ "loss": 0.3651,
974
+ "step": 65500
975
+ },
976
+ {
977
+ "epoch": 1.755122392266883,
978
+ "grad_norm": 0.6998704671859741,
979
+ "learning_rate": 1.824542601850867e-05,
980
+ "loss": 0.36,
981
+ "step": 66000
982
+ },
983
+ {
984
+ "epoch": 1.7684186732970788,
985
+ "grad_norm": 0.8846042156219482,
986
+ "learning_rate": 1.8232129560685034e-05,
987
+ "loss": 0.362,
988
+ "step": 66500
989
+ },
990
+ {
991
+ "epoch": 1.7817149543272746,
992
+ "grad_norm": 1.2595306634902954,
993
+ "learning_rate": 1.82188331028614e-05,
994
+ "loss": 0.3607,
995
+ "step": 67000
996
+ },
997
+ {
998
+ "epoch": 1.7950112353574705,
999
+ "grad_norm": 1.0262569189071655,
1000
+ "learning_rate": 1.8205536645037764e-05,
1001
+ "loss": 0.3528,
1002
+ "step": 67500
1003
+ },
1004
+ {
1005
+ "epoch": 1.8083075163876665,
1006
+ "grad_norm": 1.034749150276184,
1007
+ "learning_rate": 1.8192240187214127e-05,
1008
+ "loss": 0.3659,
1009
+ "step": 68000
1010
+ },
1011
+ {
1012
+ "epoch": 1.8216037974178623,
1013
+ "grad_norm": 1.6457266807556152,
1014
+ "learning_rate": 1.817897032230614e-05,
1015
+ "loss": 0.3618,
1016
+ "step": 68500
1017
+ },
1018
+ {
1019
+ "epoch": 1.834900078448058,
1020
+ "grad_norm": 1.0784075260162354,
1021
+ "learning_rate": 1.81656738644825e-05,
1022
+ "loss": 0.3678,
1023
+ "step": 69000
1024
+ },
1025
+ {
1026
+ "epoch": 1.848196359478254,
1027
+ "grad_norm": 1.7093465328216553,
1028
+ "learning_rate": 1.8152377406658868e-05,
1029
+ "loss": 0.3586,
1030
+ "step": 69500
1031
+ },
1032
+ {
1033
+ "epoch": 1.8614926405084498,
1034
+ "grad_norm": 1.0873297452926636,
1035
+ "learning_rate": 1.8139080948835234e-05,
1036
+ "loss": 0.3486,
1037
+ "step": 70000
1038
+ },
1039
+ {
1040
+ "epoch": 1.8614926405084498,
1041
+ "eval_loss": 0.33657437562942505,
1042
+ "eval_runtime": 44.8327,
1043
+ "eval_samples_per_second": 67.785,
1044
+ "eval_steps_per_second": 16.952,
1045
+ "step": 70000
1046
+ },
1047
+ {
1048
+ "epoch": 1.8747889215386455,
1049
+ "grad_norm": 1.2218577861785889,
1050
+ "learning_rate": 1.8125784491011594e-05,
1051
+ "loss": 0.3538,
1052
+ "step": 70500
1053
+ },
1054
+ {
1055
+ "epoch": 1.8880852025688415,
1056
+ "grad_norm": 1.7675495147705078,
1057
+ "learning_rate": 1.8112541219019256e-05,
1058
+ "loss": 0.36,
1059
+ "step": 71000
1060
+ },
1061
+ {
1062
+ "epoch": 1.9013814835990375,
1063
+ "grad_norm": 1.546845555305481,
1064
+ "learning_rate": 1.809924476119562e-05,
1065
+ "loss": 0.3681,
1066
+ "step": 71500
1067
+ },
1068
+ {
1069
+ "epoch": 1.9146777646292332,
1070
+ "grad_norm": 1.0978270769119263,
1071
+ "learning_rate": 1.8085948303371982e-05,
1072
+ "loss": 0.349,
1073
+ "step": 72000
1074
+ },
1075
+ {
1076
+ "epoch": 1.927974045659429,
1077
+ "grad_norm": 1.0180225372314453,
1078
+ "learning_rate": 1.807265184554835e-05,
1079
+ "loss": 0.3607,
1080
+ "step": 72500
1081
+ },
1082
+ {
1083
+ "epoch": 1.941270326689625,
1084
+ "grad_norm": 1.5221401453018188,
1085
+ "learning_rate": 1.8059355387724712e-05,
1086
+ "loss": 0.3603,
1087
+ "step": 73000
1088
+ },
1089
+ {
1090
+ "epoch": 1.9545666077198207,
1091
+ "grad_norm": 1.8798828125,
1092
+ "learning_rate": 1.8046058929901076e-05,
1093
+ "loss": 0.3609,
1094
+ "step": 73500
1095
+ },
1096
+ {
1097
+ "epoch": 1.9678628887500165,
1098
+ "grad_norm": 1.5229233503341675,
1099
+ "learning_rate": 1.8032762472077442e-05,
1100
+ "loss": 0.3721,
1101
+ "step": 74000
1102
+ },
1103
+ {
1104
+ "epoch": 1.9811591697802124,
1105
+ "grad_norm": 0.7534819841384888,
1106
+ "learning_rate": 1.8019466014253802e-05,
1107
+ "loss": 0.3606,
1108
+ "step": 74500
1109
+ },
1110
+ {
1111
+ "epoch": 1.9944554508104084,
1112
+ "grad_norm": 1.0966280698776245,
1113
+ "learning_rate": 1.800616955643017e-05,
1114
+ "loss": 0.3581,
1115
+ "step": 75000
1116
+ },
1117
+ {
1118
+ "epoch": 2.0077517318406044,
1119
+ "grad_norm": 1.165394902229309,
1120
+ "learning_rate": 1.7992873098606532e-05,
1121
+ "loss": 0.3426,
1122
+ "step": 75500
1123
+ },
1124
+ {
1125
+ "epoch": 2.0210480128708,
1126
+ "grad_norm": 1.4377708435058594,
1127
+ "learning_rate": 1.7979576640782895e-05,
1128
+ "loss": 0.3315,
1129
+ "step": 76000
1130
+ },
1131
+ {
1132
+ "epoch": 2.034344293900996,
1133
+ "grad_norm": 1.065831184387207,
1134
+ "learning_rate": 1.7966280182959262e-05,
1135
+ "loss": 0.3348,
1136
+ "step": 76500
1137
+ },
1138
+ {
1139
+ "epoch": 2.047640574931192,
1140
+ "grad_norm": 1.0377391576766968,
1141
+ "learning_rate": 1.7953010318051273e-05,
1142
+ "loss": 0.3297,
1143
+ "step": 77000
1144
+ },
1145
+ {
1146
+ "epoch": 2.0609368559613874,
1147
+ "grad_norm": 1.28931725025177,
1148
+ "learning_rate": 1.7939713860227636e-05,
1149
+ "loss": 0.3353,
1150
+ "step": 77500
1151
+ },
1152
+ {
1153
+ "epoch": 2.0742331369915834,
1154
+ "grad_norm": 0.9061357378959656,
1155
+ "learning_rate": 1.7926417402404003e-05,
1156
+ "loss": 0.3373,
1157
+ "step": 78000
1158
+ },
1159
+ {
1160
+ "epoch": 2.0875294180217794,
1161
+ "grad_norm": 1.162048101425171,
1162
+ "learning_rate": 1.7913120944580366e-05,
1163
+ "loss": 0.3287,
1164
+ "step": 78500
1165
+ },
1166
+ {
1167
+ "epoch": 2.1008256990519754,
1168
+ "grad_norm": 1.1295771598815918,
1169
+ "learning_rate": 1.7899851079672377e-05,
1170
+ "loss": 0.3306,
1171
+ "step": 79000
1172
+ },
1173
+ {
1174
+ "epoch": 2.114121980082171,
1175
+ "grad_norm": 1.3394831418991089,
1176
+ "learning_rate": 1.788655462184874e-05,
1177
+ "loss": 0.3425,
1178
+ "step": 79500
1179
+ },
1180
+ {
1181
+ "epoch": 2.127418261112367,
1182
+ "grad_norm": 0.7900285124778748,
1183
+ "learning_rate": 1.7873258164025103e-05,
1184
+ "loss": 0.3339,
1185
+ "step": 80000
1186
+ },
1187
+ {
1188
+ "epoch": 2.127418261112367,
1189
+ "eval_loss": 0.33293381333351135,
1190
+ "eval_runtime": 45.161,
1191
+ "eval_samples_per_second": 67.293,
1192
+ "eval_steps_per_second": 16.829,
1193
+ "step": 80000
1194
+ },
1195
+ {
1196
+ "epoch": 2.140714542142563,
1197
+ "grad_norm": 1.2849562168121338,
1198
+ "learning_rate": 1.785996170620147e-05,
1199
+ "loss": 0.3299,
1200
+ "step": 80500
1201
+ },
1202
+ {
1203
+ "epoch": 2.1540108231727584,
1204
+ "grad_norm": 1.2442893981933594,
1205
+ "learning_rate": 1.7846665248377833e-05,
1206
+ "loss": 0.3234,
1207
+ "step": 81000
1208
+ },
1209
+ {
1210
+ "epoch": 2.1673071042029544,
1211
+ "grad_norm": 0.9078606963157654,
1212
+ "learning_rate": 1.7833368790554197e-05,
1213
+ "loss": 0.3405,
1214
+ "step": 81500
1215
+ },
1216
+ {
1217
+ "epoch": 2.1806033852331503,
1218
+ "grad_norm": 1.1512479782104492,
1219
+ "learning_rate": 1.782009892564621e-05,
1220
+ "loss": 0.332,
1221
+ "step": 82000
1222
+ },
1223
+ {
1224
+ "epoch": 2.1938996662633463,
1225
+ "grad_norm": 1.0047634840011597,
1226
+ "learning_rate": 1.7806802467822574e-05,
1227
+ "loss": 0.3294,
1228
+ "step": 82500
1229
+ },
1230
+ {
1231
+ "epoch": 2.207195947293542,
1232
+ "grad_norm": 0.8341502547264099,
1233
+ "learning_rate": 1.7793506009998937e-05,
1234
+ "loss": 0.3428,
1235
+ "step": 83000
1236
+ },
1237
+ {
1238
+ "epoch": 2.220492228323738,
1239
+ "grad_norm": 1.1106518507003784,
1240
+ "learning_rate": 1.7780209552175304e-05,
1241
+ "loss": 0.3312,
1242
+ "step": 83500
1243
+ },
1244
+ {
1245
+ "epoch": 2.233788509353934,
1246
+ "grad_norm": 1.0450632572174072,
1247
+ "learning_rate": 1.7766913094351664e-05,
1248
+ "loss": 0.3228,
1249
+ "step": 84000
1250
+ },
1251
+ {
1252
+ "epoch": 2.24708479038413,
1253
+ "grad_norm": 1.2051053047180176,
1254
+ "learning_rate": 1.775361663652803e-05,
1255
+ "loss": 0.3231,
1256
+ "step": 84500
1257
+ },
1258
+ {
1259
+ "epoch": 2.2603810714143253,
1260
+ "grad_norm": 1.2320146560668945,
1261
+ "learning_rate": 1.7740320178704394e-05,
1262
+ "loss": 0.3275,
1263
+ "step": 85000
1264
+ },
1265
+ {
1266
+ "epoch": 2.2736773524445213,
1267
+ "grad_norm": 1.878662347793579,
1268
+ "learning_rate": 1.7727023720880757e-05,
1269
+ "loss": 0.3394,
1270
+ "step": 85500
1271
+ },
1272
+ {
1273
+ "epoch": 2.2869736334747173,
1274
+ "grad_norm": 0.7427929043769836,
1275
+ "learning_rate": 1.771375385597277e-05,
1276
+ "loss": 0.3334,
1277
+ "step": 86000
1278
+ },
1279
+ {
1280
+ "epoch": 2.300269914504913,
1281
+ "grad_norm": 0.6553477644920349,
1282
+ "learning_rate": 1.7700483991064782e-05,
1283
+ "loss": 0.3426,
1284
+ "step": 86500
1285
+ },
1286
+ {
1287
+ "epoch": 2.313566195535109,
1288
+ "grad_norm": 1.1945719718933105,
1289
+ "learning_rate": 1.7687187533241145e-05,
1290
+ "loss": 0.3329,
1291
+ "step": 87000
1292
+ },
1293
+ {
1294
+ "epoch": 2.3268624765653048,
1295
+ "grad_norm": 1.410997986793518,
1296
+ "learning_rate": 1.7673891075417512e-05,
1297
+ "loss": 0.3309,
1298
+ "step": 87500
1299
+ },
1300
+ {
1301
+ "epoch": 2.3401587575955007,
1302
+ "grad_norm": 1.0178992748260498,
1303
+ "learning_rate": 1.7660594617593872e-05,
1304
+ "loss": 0.3247,
1305
+ "step": 88000
1306
+ },
1307
+ {
1308
+ "epoch": 2.3534550386256963,
1309
+ "grad_norm": 1.4672740697860718,
1310
+ "learning_rate": 1.764729815977024e-05,
1311
+ "loss": 0.3373,
1312
+ "step": 88500
1313
+ },
1314
+ {
1315
+ "epoch": 2.3667513196558922,
1316
+ "grad_norm": 1.1154576539993286,
1317
+ "learning_rate": 1.763402829486225e-05,
1318
+ "loss": 0.3331,
1319
+ "step": 89000
1320
+ },
1321
+ {
1322
+ "epoch": 2.3800476006860882,
1323
+ "grad_norm": 1.5529305934906006,
1324
+ "learning_rate": 1.7620731837038613e-05,
1325
+ "loss": 0.3325,
1326
+ "step": 89500
1327
+ },
1328
+ {
1329
+ "epoch": 2.3933438817162838,
1330
+ "grad_norm": 1.4075485467910767,
1331
+ "learning_rate": 1.760743537921498e-05,
1332
+ "loss": 0.3252,
1333
+ "step": 90000
1334
+ },
1335
+ {
1336
+ "epoch": 2.3933438817162838,
1337
+ "eval_loss": 0.3309895098209381,
1338
+ "eval_runtime": 45.3672,
1339
+ "eval_samples_per_second": 66.987,
1340
+ "eval_steps_per_second": 16.752,
1341
+ "step": 90000
1342
+ },
1343
+ {
1344
+ "epoch": 2.4066401627464797,
1345
+ "grad_norm": 1.5770002603530884,
1346
+ "learning_rate": 1.7594138921391343e-05,
1347
+ "loss": 0.3341,
1348
+ "step": 90500
1349
+ },
1350
+ {
1351
+ "epoch": 2.4199364437766757,
1352
+ "grad_norm": 1.350996494293213,
1353
+ "learning_rate": 1.7580842463567706e-05,
1354
+ "loss": 0.3316,
1355
+ "step": 91000
1356
+ },
1357
+ {
1358
+ "epoch": 2.4332327248068717,
1359
+ "grad_norm": 1.3606607913970947,
1360
+ "learning_rate": 1.7567599191575368e-05,
1361
+ "loss": 0.3437,
1362
+ "step": 91500
1363
+ },
1364
+ {
1365
+ "epoch": 2.4465290058370672,
1366
+ "grad_norm": 1.3645395040512085,
1367
+ "learning_rate": 1.7554302733751728e-05,
1368
+ "loss": 0.3287,
1369
+ "step": 92000
1370
+ },
1371
+ {
1372
+ "epoch": 2.459825286867263,
1373
+ "grad_norm": 1.6436042785644531,
1374
+ "learning_rate": 1.7541006275928094e-05,
1375
+ "loss": 0.3443,
1376
+ "step": 92500
1377
+ },
1378
+ {
1379
+ "epoch": 2.473121567897459,
1380
+ "grad_norm": 0.837202250957489,
1381
+ "learning_rate": 1.7527709818104458e-05,
1382
+ "loss": 0.3221,
1383
+ "step": 93000
1384
+ },
1385
+ {
1386
+ "epoch": 2.4864178489276547,
1387
+ "grad_norm": 1.0937107801437378,
1388
+ "learning_rate": 1.751441336028082e-05,
1389
+ "loss": 0.3406,
1390
+ "step": 93500
1391
+ },
1392
+ {
1393
+ "epoch": 2.4997141299578507,
1394
+ "grad_norm": 1.4137901067733765,
1395
+ "learning_rate": 1.7501116902457187e-05,
1396
+ "loss": 0.3243,
1397
+ "step": 94000
1398
+ },
1399
+ {
1400
+ "epoch": 2.5130104109880467,
1401
+ "grad_norm": 0.8641353249549866,
1402
+ "learning_rate": 1.748782044463355e-05,
1403
+ "loss": 0.328,
1404
+ "step": 94500
1405
+ },
1406
+ {
1407
+ "epoch": 2.5263066920182427,
1408
+ "grad_norm": 1.4282252788543701,
1409
+ "learning_rate": 1.7474523986809914e-05,
1410
+ "loss": 0.3362,
1411
+ "step": 95000
1412
+ },
1413
+ {
1414
+ "epoch": 2.5396029730484386,
1415
+ "grad_norm": 1.3759071826934814,
1416
+ "learning_rate": 1.746122752898628e-05,
1417
+ "loss": 0.3339,
1418
+ "step": 95500
1419
+ },
1420
+ {
1421
+ "epoch": 2.552899254078634,
1422
+ "grad_norm": 1.239725947380066,
1423
+ "learning_rate": 1.7447931071162644e-05,
1424
+ "loss": 0.3419,
1425
+ "step": 96000
1426
+ },
1427
+ {
1428
+ "epoch": 2.56619553510883,
1429
+ "grad_norm": 0.919045627117157,
1430
+ "learning_rate": 1.7434634613339007e-05,
1431
+ "loss": 0.3369,
1432
+ "step": 96500
1433
+ },
1434
+ {
1435
+ "epoch": 2.5794918161390257,
1436
+ "grad_norm": 1.8451613187789917,
1437
+ "learning_rate": 1.7421338155515374e-05,
1438
+ "loss": 0.3417,
1439
+ "step": 97000
1440
+ },
1441
+ {
1442
+ "epoch": 2.5927880971692217,
1443
+ "grad_norm": 1.0559651851654053,
1444
+ "learning_rate": 1.740806829060738e-05,
1445
+ "loss": 0.3291,
1446
+ "step": 97500
1447
+ },
1448
+ {
1449
+ "epoch": 2.6060843781994176,
1450
+ "grad_norm": 1.215611219406128,
1451
+ "learning_rate": 1.7394771832783748e-05,
1452
+ "loss": 0.3252,
1453
+ "step": 98000
1454
+ },
1455
+ {
1456
+ "epoch": 2.6193806592296136,
1457
+ "grad_norm": 0.8735470175743103,
1458
+ "learning_rate": 1.738147537496011e-05,
1459
+ "loss": 0.3338,
1460
+ "step": 98500
1461
+ },
1462
+ {
1463
+ "epoch": 2.6326769402598096,
1464
+ "grad_norm": 1.5552194118499756,
1465
+ "learning_rate": 1.7368178917136474e-05,
1466
+ "loss": 0.3336,
1467
+ "step": 99000
1468
+ },
1469
+ {
1470
+ "epoch": 2.645973221290005,
1471
+ "grad_norm": 1.0891499519348145,
1472
+ "learning_rate": 1.735490905222849e-05,
1473
+ "loss": 0.3302,
1474
+ "step": 99500
1475
+ },
1476
+ {
1477
+ "epoch": 2.659269502320201,
1478
+ "grad_norm": 1.1763145923614502,
1479
+ "learning_rate": 1.73416391873205e-05,
1480
+ "loss": 0.33,
1481
+ "step": 100000
1482
+ },
1483
+ {
1484
+ "epoch": 2.659269502320201,
1485
+ "eval_loss": 0.3290364444255829,
1486
+ "eval_runtime": 45.2861,
1487
+ "eval_samples_per_second": 67.107,
1488
+ "eval_steps_per_second": 16.782,
1489
+ "step": 100000
1490
+ },
1491
+ {
1492
+ "epoch": 2.6725657833503966,
1493
+ "grad_norm": 1.2883906364440918,
1494
+ "learning_rate": 1.7328342729496863e-05,
1495
+ "loss": 0.3306,
1496
+ "step": 100500
1497
+ },
1498
+ {
1499
+ "epoch": 2.6858620643805926,
1500
+ "grad_norm": 1.4006438255310059,
1501
+ "learning_rate": 1.7315072864588877e-05,
1502
+ "loss": 0.3342,
1503
+ "step": 101000
1504
+ },
1505
+ {
1506
+ "epoch": 2.6991583454107886,
1507
+ "grad_norm": 1.019209861755371,
1508
+ "learning_rate": 1.7301776406765237e-05,
1509
+ "loss": 0.3321,
1510
+ "step": 101500
1511
+ },
1512
+ {
1513
+ "epoch": 2.7124546264409846,
1514
+ "grad_norm": 1.1669857501983643,
1515
+ "learning_rate": 1.7288479948941604e-05,
1516
+ "loss": 0.334,
1517
+ "step": 102000
1518
+ },
1519
+ {
1520
+ "epoch": 2.7257509074711805,
1521
+ "grad_norm": 0.8057425022125244,
1522
+ "learning_rate": 1.7275183491117967e-05,
1523
+ "loss": 0.3347,
1524
+ "step": 102500
1525
+ },
1526
+ {
1527
+ "epoch": 2.739047188501376,
1528
+ "grad_norm": 0.6847462058067322,
1529
+ "learning_rate": 1.726188703329433e-05,
1530
+ "loss": 0.3299,
1531
+ "step": 103000
1532
+ },
1533
+ {
1534
+ "epoch": 2.752343469531572,
1535
+ "grad_norm": 1.1315343379974365,
1536
+ "learning_rate": 1.7248590575470697e-05,
1537
+ "loss": 0.3424,
1538
+ "step": 103500
1539
+ },
1540
+ {
1541
+ "epoch": 2.765639750561768,
1542
+ "grad_norm": 1.2325061559677124,
1543
+ "learning_rate": 1.723529411764706e-05,
1544
+ "loss": 0.3346,
1545
+ "step": 104000
1546
+ },
1547
+ {
1548
+ "epoch": 2.7789360315919636,
1549
+ "grad_norm": 1.2530683279037476,
1550
+ "learning_rate": 1.7221997659823423e-05,
1551
+ "loss": 0.3299,
1552
+ "step": 104500
1553
+ },
1554
+ {
1555
+ "epoch": 2.7922323126221595,
1556
+ "grad_norm": 1.4341391324996948,
1557
+ "learning_rate": 1.720870120199979e-05,
1558
+ "loss": 0.3331,
1559
+ "step": 105000
1560
+ },
1561
+ {
1562
+ "epoch": 2.8055285936523555,
1563
+ "grad_norm": 0.983672559261322,
1564
+ "learning_rate": 1.7195404744176153e-05,
1565
+ "loss": 0.3229,
1566
+ "step": 105500
1567
+ },
1568
+ {
1569
+ "epoch": 2.8188248746825515,
1570
+ "grad_norm": 1.336562156677246,
1571
+ "learning_rate": 1.7182108286352516e-05,
1572
+ "loss": 0.3336,
1573
+ "step": 106000
1574
+ },
1575
+ {
1576
+ "epoch": 2.832121155712747,
1577
+ "grad_norm": 1.2226226329803467,
1578
+ "learning_rate": 1.7168811828528883e-05,
1579
+ "loss": 0.3296,
1580
+ "step": 106500
1581
+ },
1582
+ {
1583
+ "epoch": 2.845417436742943,
1584
+ "grad_norm": 0.9698474407196045,
1585
+ "learning_rate": 1.7155515370705243e-05,
1586
+ "loss": 0.3362,
1587
+ "step": 107000
1588
+ },
1589
+ {
1590
+ "epoch": 2.858713717773139,
1591
+ "grad_norm": 1.006219744682312,
1592
+ "learning_rate": 1.7142245505797257e-05,
1593
+ "loss": 0.3435,
1594
+ "step": 107500
1595
+ },
1596
+ {
1597
+ "epoch": 2.8720099988033345,
1598
+ "grad_norm": 1.3786289691925049,
1599
+ "learning_rate": 1.712894904797362e-05,
1600
+ "loss": 0.335,
1601
+ "step": 108000
1602
+ },
1603
+ {
1604
+ "epoch": 2.8853062798335305,
1605
+ "grad_norm": 1.3363252878189087,
1606
+ "learning_rate": 1.7115652590149984e-05,
1607
+ "loss": 0.3295,
1608
+ "step": 108500
1609
+ },
1610
+ {
1611
+ "epoch": 2.8986025608637265,
1612
+ "grad_norm": 0.9451866149902344,
1613
+ "learning_rate": 1.710235613232635e-05,
1614
+ "loss": 0.3226,
1615
+ "step": 109000
1616
+ },
1617
+ {
1618
+ "epoch": 2.9118988418939225,
1619
+ "grad_norm": 0.9065133929252625,
1620
+ "learning_rate": 1.708908626741836e-05,
1621
+ "loss": 0.3356,
1622
+ "step": 109500
1623
+ },
1624
+ {
1625
+ "epoch": 2.925195122924118,
1626
+ "grad_norm": 2.04927396774292,
1627
+ "learning_rate": 1.7075789809594725e-05,
1628
+ "loss": 0.3325,
1629
+ "step": 110000
1630
+ },
1631
+ {
1632
+ "epoch": 2.925195122924118,
1633
+ "eval_loss": 0.32618749141693115,
1634
+ "eval_runtime": 44.7607,
1635
+ "eval_samples_per_second": 67.894,
1636
+ "eval_steps_per_second": 16.979,
1637
+ "step": 110000
1638
+ },
1639
+ {
1640
+ "epoch": 2.938491403954314,
1641
+ "grad_norm": 1.1163157224655151,
1642
+ "learning_rate": 1.706249335177109e-05,
1643
+ "loss": 0.3257,
1644
+ "step": 110500
1645
+ },
1646
+ {
1647
+ "epoch": 2.95178768498451,
1648
+ "grad_norm": 0.929104208946228,
1649
+ "learning_rate": 1.7049196893947454e-05,
1650
+ "loss": 0.3384,
1651
+ "step": 111000
1652
+ },
1653
+ {
1654
+ "epoch": 2.9650839660147055,
1655
+ "grad_norm": 1.0319048166275024,
1656
+ "learning_rate": 1.7035927029039465e-05,
1657
+ "loss": 0.3337,
1658
+ "step": 111500
1659
+ },
1660
+ {
1661
+ "epoch": 2.9783802470449015,
1662
+ "grad_norm": 1.5630576610565186,
1663
+ "learning_rate": 1.702263057121583e-05,
1664
+ "loss": 0.3372,
1665
+ "step": 112000
1666
+ },
1667
+ {
1668
+ "epoch": 2.9916765280750974,
1669
+ "grad_norm": 1.2456904649734497,
1670
+ "learning_rate": 1.7009334113392192e-05,
1671
+ "loss": 0.339,
1672
+ "step": 112500
1673
+ },
1674
+ {
1675
+ "epoch": 3.0049728091052934,
1676
+ "grad_norm": 1.1409209966659546,
1677
+ "learning_rate": 1.699603765556856e-05,
1678
+ "loss": 0.3236,
1679
+ "step": 113000
1680
+ },
1681
+ {
1682
+ "epoch": 3.018269090135489,
1683
+ "grad_norm": 1.1287899017333984,
1684
+ "learning_rate": 1.698276779066057e-05,
1685
+ "loss": 0.2909,
1686
+ "step": 113500
1687
+ },
1688
+ {
1689
+ "epoch": 3.031565371165685,
1690
+ "grad_norm": 1.0493087768554688,
1691
+ "learning_rate": 1.6969471332836933e-05,
1692
+ "loss": 0.2988,
1693
+ "step": 114000
1694
+ },
1695
+ {
1696
+ "epoch": 3.044861652195881,
1697
+ "grad_norm": 0.8281829357147217,
1698
+ "learning_rate": 1.69561748750133e-05,
1699
+ "loss": 0.3001,
1700
+ "step": 114500
1701
+ },
1702
+ {
1703
+ "epoch": 3.058157933226077,
1704
+ "grad_norm": 1.5901657342910767,
1705
+ "learning_rate": 1.6942878417189663e-05,
1706
+ "loss": 0.2954,
1707
+ "step": 115000
1708
+ },
1709
+ {
1710
+ "epoch": 3.0714542142562724,
1711
+ "grad_norm": 1.0881614685058594,
1712
+ "learning_rate": 1.6929581959366026e-05,
1713
+ "loss": 0.2985,
1714
+ "step": 115500
1715
+ },
1716
+ {
1717
+ "epoch": 3.0847504952864684,
1718
+ "grad_norm": 1.8229660987854004,
1719
+ "learning_rate": 1.6916285501542392e-05,
1720
+ "loss": 0.2959,
1721
+ "step": 116000
1722
+ },
1723
+ {
1724
+ "epoch": 3.0980467763166644,
1725
+ "grad_norm": 0.8429426550865173,
1726
+ "learning_rate": 1.6902989043718752e-05,
1727
+ "loss": 0.2977,
1728
+ "step": 116500
1729
+ },
1730
+ {
1731
+ "epoch": 3.11134305734686,
1732
+ "grad_norm": 0.7262697219848633,
1733
+ "learning_rate": 1.688969258589512e-05,
1734
+ "loss": 0.2984,
1735
+ "step": 117000
1736
+ },
1737
+ {
1738
+ "epoch": 3.124639338377056,
1739
+ "grad_norm": 1.0413541793823242,
1740
+ "learning_rate": 1.6876396128071482e-05,
1741
+ "loss": 0.2959,
1742
+ "step": 117500
1743
+ },
1744
+ {
1745
+ "epoch": 3.137935619407252,
1746
+ "grad_norm": 1.0889323949813843,
1747
+ "learning_rate": 1.686315285607914e-05,
1748
+ "loss": 0.3,
1749
+ "step": 118000
1750
+ },
1751
+ {
1752
+ "epoch": 3.151231900437448,
1753
+ "grad_norm": 1.1839392185211182,
1754
+ "learning_rate": 1.6849856398255507e-05,
1755
+ "loss": 0.2975,
1756
+ "step": 118500
1757
+ },
1758
+ {
1759
+ "epoch": 3.1645281814676434,
1760
+ "grad_norm": 0.779194712638855,
1761
+ "learning_rate": 1.683655994043187e-05,
1762
+ "loss": 0.3028,
1763
+ "step": 119000
1764
+ },
1765
+ {
1766
+ "epoch": 3.1778244624978393,
1767
+ "grad_norm": 1.3120828866958618,
1768
+ "learning_rate": 1.6823263482608234e-05,
1769
+ "loss": 0.2929,
1770
+ "step": 119500
1771
+ },
1772
+ {
1773
+ "epoch": 3.1911207435280353,
1774
+ "grad_norm": 1.7091648578643799,
1775
+ "learning_rate": 1.68099670247846e-05,
1776
+ "loss": 0.2948,
1777
+ "step": 120000
1778
+ },
1779
+ {
1780
+ "epoch": 3.1911207435280353,
1781
+ "eval_loss": 0.3274182677268982,
1782
+ "eval_runtime": 59.8113,
1783
+ "eval_samples_per_second": 50.81,
1784
+ "eval_steps_per_second": 12.707,
1785
+ "step": 120000
1786
+ },
1787
+ {
1788
+ "epoch": 3.2044436171202912,
1789
+ "grad_norm": 1.2384920120239258,
1790
+ "learning_rate": 1.679667056696096e-05,
1791
+ "loss": 0.2823,
1792
+ "step": 120500
1793
+ },
1794
+ {
1795
+ "epoch": 3.217739898150487,
1796
+ "grad_norm": 1.2225531339645386,
1797
+ "learning_rate": 1.6783400702052975e-05,
1798
+ "loss": 0.2925,
1799
+ "step": 121000
1800
+ },
1801
+ {
1802
+ "epoch": 3.231036179180683,
1803
+ "grad_norm": 1.3342504501342773,
1804
+ "learning_rate": 1.6770104244229338e-05,
1805
+ "loss": 0.2906,
1806
+ "step": 121500
1807
+ },
1808
+ {
1809
+ "epoch": 3.244332460210879,
1810
+ "grad_norm": 1.2133206129074097,
1811
+ "learning_rate": 1.67568077864057e-05,
1812
+ "loss": 0.2877,
1813
+ "step": 122000
1814
+ },
1815
+ {
1816
+ "epoch": 3.2576287412410747,
1817
+ "grad_norm": 1.1577247381210327,
1818
+ "learning_rate": 1.6743511328582068e-05,
1819
+ "loss": 0.2876,
1820
+ "step": 122500
1821
+ },
1822
+ {
1823
+ "epoch": 3.2709250222712707,
1824
+ "grad_norm": 1.1587375402450562,
1825
+ "learning_rate": 1.673021487075843e-05,
1826
+ "loss": 0.2937,
1827
+ "step": 123000
1828
+ },
1829
+ {
1830
+ "epoch": 3.2842213033014667,
1831
+ "grad_norm": 1.1512658596038818,
1832
+ "learning_rate": 1.6716918412934794e-05,
1833
+ "loss": 0.2971,
1834
+ "step": 123500
1835
+ },
1836
+ {
1837
+ "epoch": 3.297517584331662,
1838
+ "grad_norm": 0.8885172605514526,
1839
+ "learning_rate": 1.670362195511116e-05,
1840
+ "loss": 0.2825,
1841
+ "step": 124000
1842
+ },
1843
+ {
1844
+ "epoch": 3.310813865361858,
1845
+ "grad_norm": 1.435475468635559,
1846
+ "learning_rate": 1.6690325497287524e-05,
1847
+ "loss": 0.2984,
1848
+ "step": 124500
1849
+ },
1850
+ {
1851
+ "epoch": 3.324110146392054,
1852
+ "grad_norm": 1.7068617343902588,
1853
+ "learning_rate": 1.6677029039463887e-05,
1854
+ "loss": 0.2844,
1855
+ "step": 125000
1856
+ },
1857
+ {
1858
+ "epoch": 3.33740642742225,
1859
+ "grad_norm": 0.8843211531639099,
1860
+ "learning_rate": 1.6663732581640254e-05,
1861
+ "loss": 0.2961,
1862
+ "step": 125500
1863
+ },
1864
+ {
1865
+ "epoch": 3.3507027084524457,
1866
+ "grad_norm": 1.1262086629867554,
1867
+ "learning_rate": 1.6650436123816614e-05,
1868
+ "loss": 0.2896,
1869
+ "step": 126000
1870
+ },
1871
+ {
1872
+ "epoch": 3.3639989894826416,
1873
+ "grad_norm": 1.073869228363037,
1874
+ "learning_rate": 1.663713966599298e-05,
1875
+ "loss": 0.2923,
1876
+ "step": 126500
1877
+ },
1878
+ {
1879
+ "epoch": 3.3772952705128376,
1880
+ "grad_norm": 1.1707890033721924,
1881
+ "learning_rate": 1.6623843208169347e-05,
1882
+ "loss": 0.3003,
1883
+ "step": 127000
1884
+ },
1885
+ {
1886
+ "epoch": 3.3905915515430336,
1887
+ "grad_norm": 1.2858471870422363,
1888
+ "learning_rate": 1.6610546750345707e-05,
1889
+ "loss": 0.3041,
1890
+ "step": 127500
1891
+ },
1892
+ {
1893
+ "epoch": 3.403887832573229,
1894
+ "grad_norm": 0.9487111568450928,
1895
+ "learning_rate": 1.6597250292522074e-05,
1896
+ "loss": 0.2916,
1897
+ "step": 128000
1898
+ },
1899
+ {
1900
+ "epoch": 3.417184113603425,
1901
+ "grad_norm": 1.1401737928390503,
1902
+ "learning_rate": 1.6583953834698437e-05,
1903
+ "loss": 0.2905,
1904
+ "step": 128500
1905
+ },
1906
+ {
1907
+ "epoch": 3.430480394633621,
1908
+ "grad_norm": 0.862232506275177,
1909
+ "learning_rate": 1.65706573768748e-05,
1910
+ "loss": 0.292,
1911
+ "step": 129000
1912
+ },
1913
+ {
1914
+ "epoch": 3.443776675663817,
1915
+ "grad_norm": 2.11210560798645,
1916
+ "learning_rate": 1.6557360919051167e-05,
1917
+ "loss": 0.2873,
1918
+ "step": 129500
1919
+ },
1920
+ {
1921
+ "epoch": 3.4570729566940126,
1922
+ "grad_norm": 1.325928807258606,
1923
+ "learning_rate": 1.654406446122753e-05,
1924
+ "loss": 0.2874,
1925
+ "step": 130000
1926
+ },
1927
+ {
1928
+ "epoch": 3.4570729566940126,
1929
+ "eval_loss": 0.27055609226226807,
1930
+ "eval_runtime": 44.0558,
1931
+ "eval_samples_per_second": 68.981,
1932
+ "eval_steps_per_second": 17.251,
1933
+ "step": 130000
1934
+ },
1935
+ {
1936
+ "epoch": 3.4703692377242086,
1937
+ "grad_norm": 0.9289048314094543,
1938
+ "learning_rate": 1.6530768003403894e-05,
1939
+ "loss": 0.2955,
1940
+ "step": 130500
1941
+ },
1942
+ {
1943
+ "epoch": 3.4836655187544046,
1944
+ "grad_norm": 1.1079119443893433,
1945
+ "learning_rate": 1.651747154558026e-05,
1946
+ "loss": 0.3043,
1947
+ "step": 131000
1948
+ },
1949
+ {
1950
+ "epoch": 3.4969617997846,
1951
+ "grad_norm": 0.927392840385437,
1952
+ "learning_rate": 1.650420168067227e-05,
1953
+ "loss": 0.2993,
1954
+ "step": 131500
1955
+ },
1956
+ {
1957
+ "epoch": 3.510258080814796,
1958
+ "grad_norm": 1.21388578414917,
1959
+ "learning_rate": 1.6490931815764282e-05,
1960
+ "loss": 0.2922,
1961
+ "step": 132000
1962
+ },
1963
+ {
1964
+ "epoch": 3.523554361844992,
1965
+ "grad_norm": 0.8197039365768433,
1966
+ "learning_rate": 1.6477635357940645e-05,
1967
+ "loss": 0.2945,
1968
+ "step": 132500
1969
+ },
1970
+ {
1971
+ "epoch": 3.536850642875188,
1972
+ "grad_norm": 1.3974609375,
1973
+ "learning_rate": 1.646433890011701e-05,
1974
+ "loss": 0.2892,
1975
+ "step": 133000
1976
+ },
1977
+ {
1978
+ "epoch": 3.5501469239053836,
1979
+ "grad_norm": 1.4183751344680786,
1980
+ "learning_rate": 1.6451042442293375e-05,
1981
+ "loss": 0.2879,
1982
+ "step": 133500
1983
+ },
1984
+ {
1985
+ "epoch": 3.5634432049355795,
1986
+ "grad_norm": 1.218549132347107,
1987
+ "learning_rate": 1.6437772577385386e-05,
1988
+ "loss": 0.3097,
1989
+ "step": 134000
1990
+ },
1991
+ {
1992
+ "epoch": 3.5767394859657755,
1993
+ "grad_norm": 1.6601920127868652,
1994
+ "learning_rate": 1.642447611956175e-05,
1995
+ "loss": 0.2867,
1996
+ "step": 134500
1997
+ },
1998
+ {
1999
+ "epoch": 3.590035766995971,
2000
+ "grad_norm": 1.3657351732254028,
2001
+ "learning_rate": 1.6411179661738116e-05,
2002
+ "loss": 0.3013,
2003
+ "step": 135000
2004
+ },
2005
+ {
2006
+ "epoch": 3.603332048026167,
2007
+ "grad_norm": 1.6279337406158447,
2008
+ "learning_rate": 1.639788320391448e-05,
2009
+ "loss": 0.3024,
2010
+ "step": 135500
2011
+ },
2012
+ {
2013
+ "epoch": 3.616628329056363,
2014
+ "grad_norm": 1.4741065502166748,
2015
+ "learning_rate": 1.6384586746090842e-05,
2016
+ "loss": 0.2994,
2017
+ "step": 136000
2018
+ },
2019
+ {
2020
+ "epoch": 3.629924610086559,
2021
+ "grad_norm": 1.1416784524917603,
2022
+ "learning_rate": 1.637129028826721e-05,
2023
+ "loss": 0.2996,
2024
+ "step": 136500
2025
+ },
2026
+ {
2027
+ "epoch": 3.6432208911167545,
2028
+ "grad_norm": 1.3834564685821533,
2029
+ "learning_rate": 1.635799383044357e-05,
2030
+ "loss": 0.2884,
2031
+ "step": 137000
2032
+ },
2033
+ {
2034
+ "epoch": 3.6565171721469505,
2035
+ "grad_norm": 1.2296993732452393,
2036
+ "learning_rate": 1.6344697372619936e-05,
2037
+ "loss": 0.2908,
2038
+ "step": 137500
2039
+ },
2040
+ {
2041
+ "epoch": 3.6698134531771465,
2042
+ "grad_norm": 1.3037883043289185,
2043
+ "learning_rate": 1.63314009147963e-05,
2044
+ "loss": 0.2896,
2045
+ "step": 138000
2046
+ },
2047
+ {
2048
+ "epoch": 3.683109734207342,
2049
+ "grad_norm": 1.5054666996002197,
2050
+ "learning_rate": 1.6318104456972662e-05,
2051
+ "loss": 0.2949,
2052
+ "step": 138500
2053
+ },
2054
+ {
2055
+ "epoch": 3.696406015237538,
2056
+ "grad_norm": 1.01513671875,
2057
+ "learning_rate": 1.630480799914903e-05,
2058
+ "loss": 0.2964,
2059
+ "step": 139000
2060
+ },
2061
+ {
2062
+ "epoch": 3.709702296267734,
2063
+ "grad_norm": 1.0762807130813599,
2064
+ "learning_rate": 1.6291511541325392e-05,
2065
+ "loss": 0.2938,
2066
+ "step": 139500
2067
+ },
2068
+ {
2069
+ "epoch": 3.72299857729793,
2070
+ "grad_norm": 1.9730701446533203,
2071
+ "learning_rate": 1.6278241676417403e-05,
2072
+ "loss": 0.3019,
2073
+ "step": 140000
2074
+ },
2075
+ {
2076
+ "epoch": 3.72299857729793,
2077
+ "eval_loss": 0.2732769846916199,
2078
+ "eval_runtime": 44.8357,
2079
+ "eval_samples_per_second": 67.781,
2080
+ "eval_steps_per_second": 16.951,
2081
+ "step": 140000
2082
+ },
2083
+ {
2084
+ "epoch": 3.7362948583281255,
2085
+ "grad_norm": 1.622164011001587,
2086
+ "learning_rate": 1.626494521859377e-05,
2087
+ "loss": 0.272,
2088
+ "step": 140500
2089
+ },
2090
+ {
2091
+ "epoch": 3.7495911393583214,
2092
+ "grad_norm": 1.4417494535446167,
2093
+ "learning_rate": 1.6251648760770133e-05,
2094
+ "loss": 0.2696,
2095
+ "step": 141000
2096
+ },
2097
+ {
2098
+ "epoch": 3.7628874203885174,
2099
+ "grad_norm": 0.9713847637176514,
2100
+ "learning_rate": 1.6238352302946496e-05,
2101
+ "loss": 0.276,
2102
+ "step": 141500
2103
+ },
2104
+ {
2105
+ "epoch": 3.776183701418713,
2106
+ "grad_norm": 0.9661402106285095,
2107
+ "learning_rate": 1.6225055845122863e-05,
2108
+ "loss": 0.2773,
2109
+ "step": 142000
2110
+ },
2111
+ {
2112
+ "epoch": 3.789479982448909,
2113
+ "grad_norm": 1.307726502418518,
2114
+ "learning_rate": 1.6211759387299223e-05,
2115
+ "loss": 0.2763,
2116
+ "step": 142500
2117
+ },
2118
+ {
2119
+ "epoch": 3.802776263479105,
2120
+ "grad_norm": 1.0191502571105957,
2121
+ "learning_rate": 1.619846292947559e-05,
2122
+ "loss": 0.2872,
2123
+ "step": 143000
2124
+ },
2125
+ {
2126
+ "epoch": 3.816072544509301,
2127
+ "grad_norm": 1.083509922027588,
2128
+ "learning_rate": 1.6185166471651952e-05,
2129
+ "loss": 0.2737,
2130
+ "step": 143500
2131
+ },
2132
+ {
2133
+ "epoch": 3.829368825539497,
2134
+ "grad_norm": 1.0461057424545288,
2135
+ "learning_rate": 1.6171870013828316e-05,
2136
+ "loss": 0.2745,
2137
+ "step": 144000
2138
+ },
2139
+ {
2140
+ "epoch": 3.8426651065696924,
2141
+ "grad_norm": 1.0616631507873535,
2142
+ "learning_rate": 1.6158573556004682e-05,
2143
+ "loss": 0.2753,
2144
+ "step": 144500
2145
+ },
2146
+ {
2147
+ "epoch": 3.8559613875998884,
2148
+ "grad_norm": 1.418392539024353,
2149
+ "learning_rate": 1.6145277098181046e-05,
2150
+ "loss": 0.2815,
2151
+ "step": 145000
2152
+ },
2153
+ {
2154
+ "epoch": 3.869257668630084,
2155
+ "grad_norm": 1.022754430770874,
2156
+ "learning_rate": 1.613198064035741e-05,
2157
+ "loss": 0.2721,
2158
+ "step": 145500
2159
+ },
2160
+ {
2161
+ "epoch": 3.88255394966028,
2162
+ "grad_norm": 1.5018330812454224,
2163
+ "learning_rate": 1.6118684182533776e-05,
2164
+ "loss": 0.2804,
2165
+ "step": 146000
2166
+ },
2167
+ {
2168
+ "epoch": 3.895850230690476,
2169
+ "grad_norm": 1.0998148918151855,
2170
+ "learning_rate": 1.610538772471014e-05,
2171
+ "loss": 0.2801,
2172
+ "step": 146500
2173
+ },
2174
+ {
2175
+ "epoch": 3.909146511720672,
2176
+ "grad_norm": 1.5807846784591675,
2177
+ "learning_rate": 1.6092091266886502e-05,
2178
+ "loss": 0.2821,
2179
+ "step": 147000
2180
+ },
2181
+ {
2182
+ "epoch": 3.922442792750868,
2183
+ "grad_norm": 1.443129539489746,
2184
+ "learning_rate": 1.607879480906287e-05,
2185
+ "loss": 0.2734,
2186
+ "step": 147500
2187
+ },
2188
+ {
2189
+ "epoch": 3.9357390737810634,
2190
+ "grad_norm": 1.2961505651474,
2191
+ "learning_rate": 1.6065498351239232e-05,
2192
+ "loss": 0.2694,
2193
+ "step": 148000
2194
+ },
2195
+ {
2196
+ "epoch": 3.9490353548112593,
2197
+ "grad_norm": 1.235811471939087,
2198
+ "learning_rate": 1.6052228486331243e-05,
2199
+ "loss": 0.2811,
2200
+ "step": 148500
2201
+ },
2202
+ {
2203
+ "epoch": 3.962331635841455,
2204
+ "grad_norm": 1.3737879991531372,
2205
+ "learning_rate": 1.6038932028507606e-05,
2206
+ "loss": 0.2796,
2207
+ "step": 149000
2208
+ },
2209
+ {
2210
+ "epoch": 3.975627916871651,
2211
+ "grad_norm": 1.3180122375488281,
2212
+ "learning_rate": 1.602563557068397e-05,
2213
+ "loss": 0.2752,
2214
+ "step": 149500
2215
+ },
2216
+ {
2217
+ "epoch": 3.988924197901847,
2218
+ "grad_norm": 1.293617606163025,
2219
+ "learning_rate": 1.6012339112860336e-05,
2220
+ "loss": 0.2853,
2221
+ "step": 150000
2222
+ },
2223
+ {
2224
+ "epoch": 3.988924197901847,
2225
+ "eval_loss": 0.24187567830085754,
2226
+ "eval_runtime": 44.4954,
2227
+ "eval_samples_per_second": 68.299,
2228
+ "eval_steps_per_second": 17.08,
2229
+ "step": 150000
2230
+ },
2231
+ {
2232
+ "epoch": 4.002220478932043,
2233
+ "grad_norm": 1.8393179178237915,
2234
+ "learning_rate": 1.59990426550367e-05,
2235
+ "loss": 0.2718,
2236
+ "step": 150500
2237
+ },
2238
+ {
2239
+ "epoch": 4.015516759962239,
2240
+ "grad_norm": 1.5193239450454712,
2241
+ "learning_rate": 1.5985746197213063e-05,
2242
+ "loss": 0.2665,
2243
+ "step": 151000
2244
+ },
2245
+ {
2246
+ "epoch": 4.028813040992435,
2247
+ "grad_norm": 1.0958846807479858,
2248
+ "learning_rate": 1.597244973938943e-05,
2249
+ "loss": 0.2725,
2250
+ "step": 151500
2251
+ },
2252
+ {
2253
+ "epoch": 4.04210932202263,
2254
+ "grad_norm": 1.564980387687683,
2255
+ "learning_rate": 1.595917987448144e-05,
2256
+ "loss": 0.2745,
2257
+ "step": 152000
2258
+ },
2259
+ {
2260
+ "epoch": 4.055405603052826,
2261
+ "grad_norm": 1.3031373023986816,
2262
+ "learning_rate": 1.5945883416657803e-05,
2263
+ "loss": 0.2646,
2264
+ "step": 152500
2265
+ },
2266
+ {
2267
+ "epoch": 4.068701884083022,
2268
+ "grad_norm": 1.3589409589767456,
2269
+ "learning_rate": 1.593258695883417e-05,
2270
+ "loss": 0.2777,
2271
+ "step": 153000
2272
+ },
2273
+ {
2274
+ "epoch": 4.081998165113218,
2275
+ "grad_norm": 1.092926263809204,
2276
+ "learning_rate": 1.591929050101053e-05,
2277
+ "loss": 0.2787,
2278
+ "step": 153500
2279
+ },
2280
+ {
2281
+ "epoch": 4.095294446143414,
2282
+ "grad_norm": 1.362304925918579,
2283
+ "learning_rate": 1.5906020636102544e-05,
2284
+ "loss": 0.273,
2285
+ "step": 154000
2286
+ },
2287
+ {
2288
+ "epoch": 4.10859072717361,
2289
+ "grad_norm": 1.3661013841629028,
2290
+ "learning_rate": 1.5892724178278907e-05,
2291
+ "loss": 0.2624,
2292
+ "step": 154500
2293
+ },
2294
+ {
2295
+ "epoch": 4.121887008203806,
2296
+ "grad_norm": 0.923414945602417,
2297
+ "learning_rate": 1.587942772045527e-05,
2298
+ "loss": 0.2708,
2299
+ "step": 155000
2300
+ },
2301
+ {
2302
+ "epoch": 4.135183289234002,
2303
+ "grad_norm": 1.3246159553527832,
2304
+ "learning_rate": 1.5866131262631637e-05,
2305
+ "loss": 0.2661,
2306
+ "step": 155500
2307
+ },
2308
+ {
2309
+ "epoch": 4.148479570264197,
2310
+ "grad_norm": 1.3546644449234009,
2311
+ "learning_rate": 1.5852834804808e-05,
2312
+ "loss": 0.268,
2313
+ "step": 156000
2314
+ },
2315
+ {
2316
+ "epoch": 4.161775851294393,
2317
+ "grad_norm": 1.048305630683899,
2318
+ "learning_rate": 1.5839538346984364e-05,
2319
+ "loss": 0.2711,
2320
+ "step": 156500
2321
+ },
2322
+ {
2323
+ "epoch": 4.175072132324589,
2324
+ "grad_norm": 1.0217567682266235,
2325
+ "learning_rate": 1.5826268482076378e-05,
2326
+ "loss": 0.2689,
2327
+ "step": 157000
2328
+ },
2329
+ {
2330
+ "epoch": 4.188368413354785,
2331
+ "grad_norm": 1.253959059715271,
2332
+ "learning_rate": 1.5812972024252738e-05,
2333
+ "loss": 0.2709,
2334
+ "step": 157500
2335
+ },
2336
+ {
2337
+ "epoch": 4.201664694384981,
2338
+ "grad_norm": 1.3631120920181274,
2339
+ "learning_rate": 1.5799675566429105e-05,
2340
+ "loss": 0.2753,
2341
+ "step": 158000
2342
+ },
2343
+ {
2344
+ "epoch": 4.214960975415177,
2345
+ "grad_norm": 1.5026861429214478,
2346
+ "learning_rate": 1.578637910860547e-05,
2347
+ "loss": 0.2729,
2348
+ "step": 158500
2349
+ },
2350
+ {
2351
+ "epoch": 4.228257256445373,
2352
+ "grad_norm": 1.2928775548934937,
2353
+ "learning_rate": 1.577308265078183e-05,
2354
+ "loss": 0.2645,
2355
+ "step": 159000
2356
+ },
2357
+ {
2358
+ "epoch": 4.241553537475568,
2359
+ "grad_norm": 0.8542189598083496,
2360
+ "learning_rate": 1.5759812785873845e-05,
2361
+ "loss": 0.2779,
2362
+ "step": 159500
2363
+ },
2364
+ {
2365
+ "epoch": 4.254849818505764,
2366
+ "grad_norm": 1.4010766744613647,
2367
+ "learning_rate": 1.574651632805021e-05,
2368
+ "loss": 0.2709,
2369
+ "step": 160000
2370
+ },
2371
+ {
2372
+ "epoch": 4.254849818505764,
2373
+ "eval_loss": 0.24389813840389252,
2374
+ "eval_runtime": 44.8113,
2375
+ "eval_samples_per_second": 67.818,
2376
+ "eval_steps_per_second": 16.96,
2377
+ "step": 160000
2378
+ },
2379
+ {
2380
+ "epoch": 4.26814609953596,
2381
+ "grad_norm": 1.1164586544036865,
2382
+ "learning_rate": 1.5733219870226572e-05,
2383
+ "loss": 0.2675,
2384
+ "step": 160500
2385
+ },
2386
+ {
2387
+ "epoch": 4.281442380566156,
2388
+ "grad_norm": 1.361721158027649,
2389
+ "learning_rate": 1.571992341240294e-05,
2390
+ "loss": 0.2767,
2391
+ "step": 161000
2392
+ },
2393
+ {
2394
+ "epoch": 4.294738661596352,
2395
+ "grad_norm": 1.3751747608184814,
2396
+ "learning_rate": 1.5706626954579302e-05,
2397
+ "loss": 0.2672,
2398
+ "step": 161500
2399
+ },
2400
+ {
2401
+ "epoch": 4.308034942626548,
2402
+ "grad_norm": 1.032375693321228,
2403
+ "learning_rate": 1.5693357089671313e-05,
2404
+ "loss": 0.2786,
2405
+ "step": 162000
2406
+ },
2407
+ {
2408
+ "epoch": 4.321331223656744,
2409
+ "grad_norm": 1.288988471031189,
2410
+ "learning_rate": 1.568006063184768e-05,
2411
+ "loss": 0.2845,
2412
+ "step": 162500
2413
+ },
2414
+ {
2415
+ "epoch": 4.334627504686939,
2416
+ "grad_norm": 1.0330158472061157,
2417
+ "learning_rate": 1.566676417402404e-05,
2418
+ "loss": 0.2811,
2419
+ "step": 163000
2420
+ },
2421
+ {
2422
+ "epoch": 4.347923785717135,
2423
+ "grad_norm": 1.7785344123840332,
2424
+ "learning_rate": 1.5653467716200406e-05,
2425
+ "loss": 0.2791,
2426
+ "step": 163500
2427
+ },
2428
+ {
2429
+ "epoch": 4.361220066747331,
2430
+ "grad_norm": 1.7403333187103271,
2431
+ "learning_rate": 1.5640197851292417e-05,
2432
+ "loss": 0.2773,
2433
+ "step": 164000
2434
+ },
2435
+ {
2436
+ "epoch": 4.374516347777527,
2437
+ "grad_norm": 1.0584579706192017,
2438
+ "learning_rate": 1.562690139346878e-05,
2439
+ "loss": 0.2711,
2440
+ "step": 164500
2441
+ },
2442
+ {
2443
+ "epoch": 4.387812628807723,
2444
+ "grad_norm": 1.419774055480957,
2445
+ "learning_rate": 1.5613604935645147e-05,
2446
+ "loss": 0.2762,
2447
+ "step": 165000
2448
+ },
2449
+ {
2450
+ "epoch": 4.401108909837919,
2451
+ "grad_norm": 1.3535691499710083,
2452
+ "learning_rate": 1.560030847782151e-05,
2453
+ "loss": 0.271,
2454
+ "step": 165500
2455
+ },
2456
+ {
2457
+ "epoch": 4.414405190868115,
2458
+ "grad_norm": 1.4730337858200073,
2459
+ "learning_rate": 1.5587012019997873e-05,
2460
+ "loss": 0.2753,
2461
+ "step": 166000
2462
+ },
2463
+ {
2464
+ "epoch": 4.42770147189831,
2465
+ "grad_norm": 1.1973862648010254,
2466
+ "learning_rate": 1.557371556217424e-05,
2467
+ "loss": 0.2731,
2468
+ "step": 166500
2469
+ },
2470
+ {
2471
+ "epoch": 4.440997752928506,
2472
+ "grad_norm": 1.2967299222946167,
2473
+ "learning_rate": 1.5560472290181895e-05,
2474
+ "loss": 0.2815,
2475
+ "step": 167000
2476
+ },
2477
+ {
2478
+ "epoch": 4.454294033958702,
2479
+ "grad_norm": 1.1886343955993652,
2480
+ "learning_rate": 1.554717583235826e-05,
2481
+ "loss": 0.27,
2482
+ "step": 167500
2483
+ },
2484
+ {
2485
+ "epoch": 4.467590314988898,
2486
+ "grad_norm": 1.6000491380691528,
2487
+ "learning_rate": 1.5533879374534625e-05,
2488
+ "loss": 0.28,
2489
+ "step": 168000
2490
+ },
2491
+ {
2492
+ "epoch": 4.480886596019094,
2493
+ "grad_norm": 1.6364268064498901,
2494
+ "learning_rate": 1.5520582916710988e-05,
2495
+ "loss": 0.2741,
2496
+ "step": 168500
2497
+ },
2498
+ {
2499
+ "epoch": 4.4941828770492895,
2500
+ "grad_norm": 1.4403234720230103,
2501
+ "learning_rate": 1.5507286458887355e-05,
2502
+ "loss": 0.2768,
2503
+ "step": 169000
2504
+ },
2505
+ {
2506
+ "epoch": 4.5074791580794855,
2507
+ "grad_norm": 1.1721081733703613,
2508
+ "learning_rate": 1.5493990001063718e-05,
2509
+ "loss": 0.2812,
2510
+ "step": 169500
2511
+ },
2512
+ {
2513
+ "epoch": 4.520775439109681,
2514
+ "grad_norm": 1.213318943977356,
2515
+ "learning_rate": 1.548072013615573e-05,
2516
+ "loss": 0.2772,
2517
+ "step": 170000
2518
+ },
2519
+ {
2520
+ "epoch": 4.520775439109681,
2521
+ "eval_loss": 0.2475610226392746,
2522
+ "eval_runtime": 44.8886,
2523
+ "eval_samples_per_second": 67.701,
2524
+ "eval_steps_per_second": 16.931,
2525
+ "step": 170000
2526
+ },
2527
+ {
2528
+ "epoch": 4.534071720139877,
2529
+ "grad_norm": 1.3363765478134155,
2530
+ "learning_rate": 1.5467423678332095e-05,
2531
+ "loss": 0.2748,
2532
+ "step": 170500
2533
+ },
2534
+ {
2535
+ "epoch": 4.547368001170073,
2536
+ "grad_norm": 1.1976338624954224,
2537
+ "learning_rate": 1.5454127220508455e-05,
2538
+ "loss": 0.2731,
2539
+ "step": 171000
2540
+ },
2541
+ {
2542
+ "epoch": 4.5606642822002685,
2543
+ "grad_norm": 1.1618109941482544,
2544
+ "learning_rate": 1.5440830762684822e-05,
2545
+ "loss": 0.2772,
2546
+ "step": 171500
2547
+ },
2548
+ {
2549
+ "epoch": 4.5739605632304645,
2550
+ "grad_norm": 1.5081138610839844,
2551
+ "learning_rate": 1.542753430486119e-05,
2552
+ "loss": 0.2711,
2553
+ "step": 172000
2554
+ },
2555
+ {
2556
+ "epoch": 4.5872568442606605,
2557
+ "grad_norm": 1.2148953676223755,
2558
+ "learning_rate": 1.541423784703755e-05,
2559
+ "loss": 0.2708,
2560
+ "step": 172500
2561
+ },
2562
+ {
2563
+ "epoch": 4.6005531252908565,
2564
+ "grad_norm": 0.9179415702819824,
2565
+ "learning_rate": 1.5400941389213915e-05,
2566
+ "loss": 0.2674,
2567
+ "step": 173000
2568
+ },
2569
+ {
2570
+ "epoch": 4.613849406321052,
2571
+ "grad_norm": 1.1413265466690063,
2572
+ "learning_rate": 1.538764493139028e-05,
2573
+ "loss": 0.2747,
2574
+ "step": 173500
2575
+ },
2576
+ {
2577
+ "epoch": 4.6271456873512475,
2578
+ "grad_norm": 1.0427721738815308,
2579
+ "learning_rate": 1.537434847356664e-05,
2580
+ "loss": 0.2802,
2581
+ "step": 174000
2582
+ },
2583
+ {
2584
+ "epoch": 4.6404419683814435,
2585
+ "grad_norm": 1.8311312198638916,
2586
+ "learning_rate": 1.5361052015743008e-05,
2587
+ "loss": 0.2746,
2588
+ "step": 174500
2589
+ },
2590
+ {
2591
+ "epoch": 4.6537382494116395,
2592
+ "grad_norm": 1.482964038848877,
2593
+ "learning_rate": 1.534775555791937e-05,
2594
+ "loss": 0.2661,
2595
+ "step": 175000
2596
+ },
2597
+ {
2598
+ "epoch": 4.6670345304418355,
2599
+ "grad_norm": 1.2513574361801147,
2600
+ "learning_rate": 1.5334485693011382e-05,
2601
+ "loss": 0.2654,
2602
+ "step": 175500
2603
+ },
2604
+ {
2605
+ "epoch": 4.6803308114720314,
2606
+ "grad_norm": 1.0544030666351318,
2607
+ "learning_rate": 1.532118923518775e-05,
2608
+ "loss": 0.2783,
2609
+ "step": 176000
2610
+ },
2611
+ {
2612
+ "epoch": 4.693627092502227,
2613
+ "grad_norm": 1.350472331047058,
2614
+ "learning_rate": 1.530789277736411e-05,
2615
+ "loss": 0.2846,
2616
+ "step": 176500
2617
+ },
2618
+ {
2619
+ "epoch": 4.7069233735324225,
2620
+ "grad_norm": 1.0775673389434814,
2621
+ "learning_rate": 1.5294596319540476e-05,
2622
+ "loss": 0.2759,
2623
+ "step": 177000
2624
+ },
2625
+ {
2626
+ "epoch": 4.7202196545626185,
2627
+ "grad_norm": 0.8350367546081543,
2628
+ "learning_rate": 1.5281299861716842e-05,
2629
+ "loss": 0.2778,
2630
+ "step": 177500
2631
+ },
2632
+ {
2633
+ "epoch": 4.7335159355928145,
2634
+ "grad_norm": 1.2634960412979126,
2635
+ "learning_rate": 1.5268003403893202e-05,
2636
+ "loss": 0.2754,
2637
+ "step": 178000
2638
+ },
2639
+ {
2640
+ "epoch": 4.7468122166230104,
2641
+ "grad_norm": 1.2727707624435425,
2642
+ "learning_rate": 1.5254706946069569e-05,
2643
+ "loss": 0.2811,
2644
+ "step": 178500
2645
+ },
2646
+ {
2647
+ "epoch": 4.760108497653206,
2648
+ "grad_norm": 1.2721821069717407,
2649
+ "learning_rate": 1.5241410488245934e-05,
2650
+ "loss": 0.2765,
2651
+ "step": 179000
2652
+ },
2653
+ {
2654
+ "epoch": 4.773404778683402,
2655
+ "grad_norm": 1.0992257595062256,
2656
+ "learning_rate": 1.522816721625359e-05,
2657
+ "loss": 0.2767,
2658
+ "step": 179500
2659
+ },
2660
+ {
2661
+ "epoch": 4.786701059713598,
2662
+ "grad_norm": 0.9015975594520569,
2663
+ "learning_rate": 1.5214870758429955e-05,
2664
+ "loss": 0.2713,
2665
+ "step": 180000
2666
+ },
2667
+ {
2668
+ "epoch": 4.786701059713598,
2669
+ "eval_loss": 0.24829423427581787,
2670
+ "eval_runtime": 44.7845,
2671
+ "eval_samples_per_second": 67.858,
2672
+ "eval_steps_per_second": 16.97,
2673
+ "step": 180000
2674
+ },
2675
+ {
2676
+ "epoch": 4.7999973407437935,
2677
+ "grad_norm": 1.1640844345092773,
2678
+ "learning_rate": 1.5201600893521966e-05,
2679
+ "loss": 0.2836,
2680
+ "step": 180500
2681
+ },
2682
+ {
2683
+ "epoch": 4.8132936217739895,
2684
+ "grad_norm": 1.0341435670852661,
2685
+ "learning_rate": 1.5188304435698331e-05,
2686
+ "loss": 0.2761,
2687
+ "step": 181000
2688
+ },
2689
+ {
2690
+ "epoch": 4.826589902804185,
2691
+ "grad_norm": 1.327324628829956,
2692
+ "learning_rate": 1.5175007977874696e-05,
2693
+ "loss": 0.2791,
2694
+ "step": 181500
2695
+ },
2696
+ {
2697
+ "epoch": 4.839886183834381,
2698
+ "grad_norm": 0.8103187084197998,
2699
+ "learning_rate": 1.5161738112966707e-05,
2700
+ "loss": 0.2879,
2701
+ "step": 182000
2702
+ },
2703
+ {
2704
+ "epoch": 4.853182464864577,
2705
+ "grad_norm": 0.7801828384399414,
2706
+ "learning_rate": 1.5148441655143072e-05,
2707
+ "loss": 0.2756,
2708
+ "step": 182500
2709
+ },
2710
+ {
2711
+ "epoch": 4.866478745894773,
2712
+ "grad_norm": 1.30203378200531,
2713
+ "learning_rate": 1.5135145197319437e-05,
2714
+ "loss": 0.277,
2715
+ "step": 183000
2716
+ },
2717
+ {
2718
+ "epoch": 4.879775026924969,
2719
+ "grad_norm": 1.3664675951004028,
2720
+ "learning_rate": 1.5121848739495799e-05,
2721
+ "loss": 0.2731,
2722
+ "step": 183500
2723
+ },
2724
+ {
2725
+ "epoch": 4.893071307955165,
2726
+ "grad_norm": 1.314031720161438,
2727
+ "learning_rate": 1.5108552281672163e-05,
2728
+ "loss": 0.2729,
2729
+ "step": 184000
2730
+ },
2731
+ {
2732
+ "epoch": 4.90636758898536,
2733
+ "grad_norm": 0.8956036567687988,
2734
+ "learning_rate": 1.5095255823848527e-05,
2735
+ "loss": 0.2788,
2736
+ "step": 184500
2737
+ },
2738
+ {
2739
+ "epoch": 4.919663870015556,
2740
+ "grad_norm": 1.520376205444336,
2741
+ "learning_rate": 1.5081959366024892e-05,
2742
+ "loss": 0.2845,
2743
+ "step": 185000
2744
+ },
2745
+ {
2746
+ "epoch": 4.932960151045752,
2747
+ "grad_norm": 1.3505889177322388,
2748
+ "learning_rate": 1.5068662908201257e-05,
2749
+ "loss": 0.2818,
2750
+ "step": 185500
2751
+ },
2752
+ {
2753
+ "epoch": 4.946256432075948,
2754
+ "grad_norm": 0.6950279474258423,
2755
+ "learning_rate": 1.505536645037762e-05,
2756
+ "loss": 0.2737,
2757
+ "step": 186000
2758
+ },
2759
+ {
2760
+ "epoch": 4.959552713106144,
2761
+ "grad_norm": 1.2404699325561523,
2762
+ "learning_rate": 1.5042069992553985e-05,
2763
+ "loss": 0.2813,
2764
+ "step": 186500
2765
+ },
2766
+ {
2767
+ "epoch": 4.97284899413634,
2768
+ "grad_norm": 1.1977089643478394,
2769
+ "learning_rate": 1.5028800127645997e-05,
2770
+ "loss": 0.2758,
2771
+ "step": 187000
2772
+ },
2773
+ {
2774
+ "epoch": 4.986145275166536,
2775
+ "grad_norm": 1.2810465097427368,
2776
+ "learning_rate": 1.501550366982236e-05,
2777
+ "loss": 0.287,
2778
+ "step": 187500
2779
+ },
2780
+ {
2781
+ "epoch": 4.999441556196732,
2782
+ "grad_norm": 2.558598756790161,
2783
+ "learning_rate": 1.5002207211998726e-05,
2784
+ "loss": 0.2701,
2785
+ "step": 188000
2786
+ },
2787
+ {
2788
+ "epoch": 5.012737837226927,
2789
+ "grad_norm": 1.2791069746017456,
2790
+ "learning_rate": 1.498891075417509e-05,
2791
+ "loss": 0.2406,
2792
+ "step": 188500
2793
+ },
2794
+ {
2795
+ "epoch": 5.026034118257123,
2796
+ "grad_norm": 1.0913165807724,
2797
+ "learning_rate": 1.4975614296351452e-05,
2798
+ "loss": 0.2307,
2799
+ "step": 189000
2800
+ },
2801
+ {
2802
+ "epoch": 5.039330399287319,
2803
+ "grad_norm": 1.1861522197723389,
2804
+ "learning_rate": 1.4962317838527817e-05,
2805
+ "loss": 0.2429,
2806
+ "step": 189500
2807
+ },
2808
+ {
2809
+ "epoch": 5.052626680317515,
2810
+ "grad_norm": 1.2686249017715454,
2811
+ "learning_rate": 1.494902138070418e-05,
2812
+ "loss": 0.2395,
2813
+ "step": 190000
2814
+ },
2815
+ {
2816
+ "epoch": 5.052626680317515,
2817
+ "eval_loss": 0.2490723580121994,
2818
+ "eval_runtime": 43.0318,
2819
+ "eval_samples_per_second": 70.622,
2820
+ "eval_steps_per_second": 17.661,
2821
+ "step": 190000
2822
+ },
2823
+ {
2824
+ "epoch": 5.065922961347711,
2825
+ "grad_norm": 1.4228299856185913,
2826
+ "learning_rate": 1.4935724922880545e-05,
2827
+ "loss": 0.2493,
2828
+ "step": 190500
2829
+ },
2830
+ {
2831
+ "epoch": 5.079219242377907,
2832
+ "grad_norm": 1.0055017471313477,
2833
+ "learning_rate": 1.492242846505691e-05,
2834
+ "loss": 0.2522,
2835
+ "step": 191000
2836
+ },
2837
+ {
2838
+ "epoch": 5.092515523408102,
2839
+ "grad_norm": 0.8438519239425659,
2840
+ "learning_rate": 1.4909132007233274e-05,
2841
+ "loss": 0.2546,
2842
+ "step": 191500
2843
+ },
2844
+ {
2845
+ "epoch": 5.105811804438298,
2846
+ "grad_norm": 1.2010571956634521,
2847
+ "learning_rate": 1.4895862142325286e-05,
2848
+ "loss": 0.2352,
2849
+ "step": 192000
2850
+ },
2851
+ {
2852
+ "epoch": 5.119108085468494,
2853
+ "grad_norm": 1.275994062423706,
2854
+ "learning_rate": 1.4882565684501651e-05,
2855
+ "loss": 0.2372,
2856
+ "step": 192500
2857
+ },
2858
+ {
2859
+ "epoch": 5.13240436649869,
2860
+ "grad_norm": 0.7371090054512024,
2861
+ "learning_rate": 1.4869269226678014e-05,
2862
+ "loss": 0.2487,
2863
+ "step": 193000
2864
+ },
2865
+ {
2866
+ "epoch": 5.145700647528886,
2867
+ "grad_norm": 1.3083887100219727,
2868
+ "learning_rate": 1.485597276885438e-05,
2869
+ "loss": 0.2485,
2870
+ "step": 193500
2871
+ },
2872
+ {
2873
+ "epoch": 5.158996928559082,
2874
+ "grad_norm": 0.972006618976593,
2875
+ "learning_rate": 1.4842676311030744e-05,
2876
+ "loss": 0.2517,
2877
+ "step": 194000
2878
+ },
2879
+ {
2880
+ "epoch": 5.172293209589278,
2881
+ "grad_norm": 1.3523329496383667,
2882
+ "learning_rate": 1.4829379853207106e-05,
2883
+ "loss": 0.243,
2884
+ "step": 194500
2885
+ },
2886
+ {
2887
+ "epoch": 5.185589490619474,
2888
+ "grad_norm": 1.2454783916473389,
2889
+ "learning_rate": 1.4816109988299118e-05,
2890
+ "loss": 0.2456,
2891
+ "step": 195000
2892
+ },
2893
+ {
2894
+ "epoch": 5.198885771649669,
2895
+ "grad_norm": 1.2170870304107666,
2896
+ "learning_rate": 1.480284012339113e-05,
2897
+ "loss": 0.2524,
2898
+ "step": 195500
2899
+ },
2900
+ {
2901
+ "epoch": 5.212182052679865,
2902
+ "grad_norm": 1.7185641527175903,
2903
+ "learning_rate": 1.4789570258483142e-05,
2904
+ "loss": 0.2403,
2905
+ "step": 196000
2906
+ },
2907
+ {
2908
+ "epoch": 5.225478333710061,
2909
+ "grad_norm": 1.2505462169647217,
2910
+ "learning_rate": 1.4776273800659507e-05,
2911
+ "loss": 0.248,
2912
+ "step": 196500
2913
+ },
2914
+ {
2915
+ "epoch": 5.238774614740257,
2916
+ "grad_norm": 1.061320424079895,
2917
+ "learning_rate": 1.4762977342835868e-05,
2918
+ "loss": 0.2522,
2919
+ "step": 197000
2920
+ },
2921
+ {
2922
+ "epoch": 5.252070895770453,
2923
+ "grad_norm": 1.1784652471542358,
2924
+ "learning_rate": 1.4749680885012235e-05,
2925
+ "loss": 0.2471,
2926
+ "step": 197500
2927
+ },
2928
+ {
2929
+ "epoch": 5.265367176800649,
2930
+ "grad_norm": 1.6446605920791626,
2931
+ "learning_rate": 1.47363844271886e-05,
2932
+ "loss": 0.2459,
2933
+ "step": 198000
2934
+ },
2935
+ {
2936
+ "epoch": 5.278663457830845,
2937
+ "grad_norm": 1.7436884641647339,
2938
+ "learning_rate": 1.4723087969364962e-05,
2939
+ "loss": 0.2452,
2940
+ "step": 198500
2941
+ },
2942
+ {
2943
+ "epoch": 5.29195973886104,
2944
+ "grad_norm": 1.2458696365356445,
2945
+ "learning_rate": 1.4709791511541326e-05,
2946
+ "loss": 0.2535,
2947
+ "step": 199000
2948
+ },
2949
+ {
2950
+ "epoch": 5.305256019891236,
2951
+ "grad_norm": 1.1361490488052368,
2952
+ "learning_rate": 1.469649505371769e-05,
2953
+ "loss": 0.2444,
2954
+ "step": 199500
2955
+ },
2956
+ {
2957
+ "epoch": 5.318552300921432,
2958
+ "grad_norm": 1.2189501523971558,
2959
+ "learning_rate": 1.4683198595894055e-05,
2960
+ "loss": 0.2509,
2961
+ "step": 200000
2962
+ },
2963
+ {
2964
+ "epoch": 5.318552300921432,
2965
+ "eval_loss": 0.25082090497016907,
2966
+ "eval_runtime": 56.1856,
2967
+ "eval_samples_per_second": 54.089,
2968
+ "eval_steps_per_second": 13.527,
2969
+ "step": 200000
2970
+ }
2971
+ ],
2972
+ "logging_steps": 500,
2973
+ "max_steps": 752080,
2974
+ "num_input_tokens_seen": 0,
2975
+ "num_train_epochs": 20,
2976
+ "save_steps": 10000,
2977
+ "stateful_callbacks": {
2978
+ "TrainerControl": {
2979
+ "args": {
2980
+ "should_epoch_stop": false,
2981
+ "should_evaluate": false,
2982
+ "should_log": false,
2983
+ "should_save": true,
2984
+ "should_training_stop": false
2985
+ },
2986
+ "attributes": {}
2987
+ }
2988
+ },
2989
+ "total_flos": 4.334195659820237e+17,
2990
+ "train_batch_size": 4,
2991
+ "trial_name": null,
2992
+ "trial_params": null
2993
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2675020d4b231408657b4cd1dce6eb7893190230e81226562aa7d0c3f7d0d7dd
3
+ size 5112
vocab.json ADDED
The diff for this file is too large to render. See raw diff